UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Design of a Rocketsonde Buoy System for collecting weather sounding data Readyhough, Catherine Ene 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2003-0494.pdf [ 18.83MB ]
JSON: 831-1.0090961.json
JSON-LD: 831-1.0090961-ld.json
RDF/XML (Pretty): 831-1.0090961-rdf.xml
RDF/JSON: 831-1.0090961-rdf.json
Turtle: 831-1.0090961-turtle.txt
N-Triples: 831-1.0090961-rdf-ntriples.txt
Original Record: 831-1.0090961-source.json
Full Text

Full Text

DESIGN OF A ROCKETSONDE BUOY SYSTEM FOR COLLECTING WEATHER SOUNDING DATA by CATHERINE ENE  READYHOUGH  B.A. S c . , The U n i v e r s i t y  o f T o r o n t o , 2001  T H E S I S SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER  OF A P P L I E D SCIENCE in  THE  FACULTY  OF G R A D U A T E S T U D I E S  THE DEPARTMENT OF MECHANICAL ENGINEERING  We a c c e p t t h i s t h e s i s as c o n f o r m i n g to the required standard  THE UNIVERSITY OF B R I T I S H COLUMBIA May  2003  © C a t h e r i n e Ene R e a d y h o u g h ,  2003  In presenting this thesis in partial fulfilment  of the requirements for an advanced  degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department  or by his  or  her  representatives.  It  is  understood  that  copying  or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department The University of British Columbia Vancouver, Canada  DE-6 (2/88)  ABSTRACT A Rocketsonde-Buoy System is presented as a possible solution to the problem of gathering atmospheric data over large bodies o f water, such as oceans. This system consists of a series  o f deep-ocean  buoys  at fixed  locations, each  instrumentation rockets, called rocketsondes.  containing a battery  of  weather  Each buoy w i l l autonomously fire a rocketsonde  daily that w i l l parachute down while transmitting local atmospheric data.  The goal of the  research is to prove that this system is technologically feasible. The main focus is to design an appropriate rocket and guarantee that it w i l l launch vertically even during severe winter storm conditions.  Simulations and experimental tests indicate that it is feasible to design an  appropriate rocket for this system. The rocket w i l l use a fast-burning, K class solid propellant motor to reach the desired 6km altitude and cut through the wind conditions. The dual diameter body w i l l be made from aluminum, which is degradable i n sea water, and w i l l have 6 fixed fins. This rocket can be stored and launched from a 4 " diameter tube, which improves the ability for the system to be easily weatherproofed. However, due to weight issues, the system w i l l only be able to hold 200 o f these rockets, rather than the desired 400. A prototype launch control system was also developed, implementing a 3-axis orientation sensor to track buoy motion.  Tests  demonstrated that the system is capable o f launching the prototype rocket near-vertical in rocking conditions equivalent to the worst likely to be faced by the Rocketsonde Buoy System. The launch control system also identifies launch opportunities i n wave troughs, where there is less wind to affect the rocket's initial flight.  A n optimization program is also implemented to  change the maximum acceptable launch angle based on current wave conditions.  ii  TABLE OF CONTENTS  Abstract  "  Table of Contents  »'  List of Tables  v  List of Figures  v  List of Symbols  '  v i i  1.  INTRODUCTION  1  2.  DETAILED B A C K G R O U N D  6  2.1  W E A T H E R O B S E R V A T I O N IN T H EN O R T H PACIFIC  2.1.1 2.1.2  6  Past and Existing Systems THORpex  6 8  2.2  R O C K E T S O N D E - B U O Y S Y S T E M OBJECTIVES A N DR E Q U I R E M E N T S  11  2.3  ADDITIONAL ROCKETSONDE REQUIREMENTS  17  2.4  LAUNCH SYSTEM REQUIREMENTS  18  2.5  S U M M A R Y OF SYSTEM REQUIREMENTS  20  3.  R O C K E T R Y DESIGN 3.1  22  THEORETICAL B A C K G R O U N D  3.1.1 3.1.2 3.1.3 3.1.4 3.1.5 3.1.6  22  Basic Rocket Design. Centre of Pressure , Rocket Stability Circular Fins or "Ringtails " Drag Engine Design  ;  •  • •  ,  22 23 26 28 29 30  3.2  INITIAL D E S I G N D E C I S I O N S  32  3.3  SIMULATIONS  37  3.3.1 3.3.2 3.3.3 3.3.4 3.3.5 3.3.6 3.3.7 3.4  Introduction to RocSim Determining Motor Range Nose Cone and Body Tube Design Effects of Motor Burn Characteristics Fin Design Considerations Launch Tube Length Wind Effects on Recovery  37 40 42 44 46 50 51  EXPERIMENTAL RESULTS  3.4.1 3.4.2  3.4.3  3.4.4 3.5  51  Experimental Objectives Experimental Procedure Location Launching Apparatus Rocket Prototype Construction Launch History Fin Designs Results Conventional Fins Spring-Released Folding Fins Spring-Released Folding with Locking Mechanism Tangent Fold-Down Fins Circular Fins Tube Launch Results C O N C L U S I O N S A N DF I N A L R O C K E T S O N D E D E S I G N  iii  52 52 52 53 54 55 57 57 58 60 '.. 61 63 65 66  J  4.  LAUNCH CONTROL 4.1  69  BACKGROUND A N DTHEORY  4.1.1 4.1.2 4.2  •  70 71  SENSOR SELECTION  4.2.1 4.2.2 4.2.3 4.2.4 4.2.5 4.2.6 4.3  •  Sensor Requirements Orientation Sensor Technology. ; : MicroStrain 3DM-G Gyro-Enhanced Orientation Sensor Test Apparatus Results Sensor Conclusions ALGORITHM DESIGN  4.4  5.  •  74 75 76 78 80 83 85 85 90 91 93 94  EXPERIMENTAL VALIDATION  4.4.1 4.4.2 4.4.3  •  73  84  4.3.1 Determining Launch Angle Alpha 4.3.2 Determining Wave Position 4.3.3 Optimizing the Launch Angle 4.3.4 Algorithm Implementation 4.3.5 Filtering 4.3.6 Basic Laboratory Testing  4.5  70  Reference Frames System Interfaces  96  Experimental Set Up and Procedure Results and Discussion Experimental Conclusions  97 99 105  CONCLUSIONS A N DFINAL L A U N C H CONTROL DESIGN  CONCLUSIONS AND FUTURE W O R K  105  108  REFERENCES  112  APPENDIX A: R E F E R E N C E M A T E R I A L F R O M WEBPAGES  114  iv  LIST OF  TABLES  T A B L E 3.1:  M A X I M U M ACCELERATION FOR PROTOTYPE DESIGNS  T A B L E 3.2:  R E Q U I R E D L A U N C H T U B E L E N G T H S FOR C I R C U L A R FIN PROTOTYPE  50  T A B L E 3.3:  W I N D EFFECTS O N 4-FIN R O C K E T WITH K250 ENGINE  51  T A B L E 3.4:  COMPLETE ROCKET  56  LAUNCH RECORD  v  49  LIST OF FIGURES F I G U R E 1.1: R A O B S I T E S I N N O R T H A M E R I C A  2  F I G U R E 1.2: O B S E R V A T I O N " S A N D W I C H "  3  F I G U R E 2.1: D R I F T S O N D E G O N D O L A  9  F I G U R E 2.2: T H E A E R O S O N D E  NOMAD NOMAD B U O Y  F I G U R E 2.3: N O R T H F I G U R E 2.4:  10 E N V I R O N M E N T A L C O N D I T I O N S , F E B 2002  9  13 15  F I G U R E 3.1: C O N V E N T I O N A L R O C K E T D E S I G N  22  F I G U R E 3.2: E F F E C T S O F C P O N R O C K E T S T A B I L I T Y  27  F I G U R E 3.3: S O L I D P R O P E L L A N T M O T O R C R O S S - S E C T I O N  31  F I G U R E 3.4: E X A M P L E R O C S I M O U T P U T  39  F I G U R E 3.5: M O T O R I M P U L S E V S . R O C K E T A L T I T U D E A N D M A S S  41  F I G U R E 3.6: P R O T O T Y P E R O C K E T W I T H 4 - F I N C O N F I G U R A T I O N  43  F I G U R E 3.7: W I N D S P E E D V S . A L T I T U D E R E A C H E D B Y 4 - F I N D E S I G N  45  F I G U R E 3.8: P R O T O T Y P E R O C K E T W I T H 6 - F I N C O N F I G U R A T I O N  47  F I G U R E 3.9: P R O T O T Y P E R O C K E T W I T H C I R C U L A R F I N C O N F I G U R A T I O N  48  F I G U R E 3.10: E F F E C T S O F F I N D R A G O N R O C K E T A L T I T U D E  49  F I G U R E 3.11: R O C K E T L A U N C H O V E R H A R R I S O N L A K E , B C  53  F I G U R E 3.12: L A U N C H A P P A R A T U S  54  F I G U R E 3.13: P R O T O T Y P E D E V E L O P M E N T  55  F I G U R E 3.14: S P R I N G - H I N G E D  FINS  58  F I G U R E 3.15: T A N G E N T F O L D - D O W N F I N S  61  F I G U R E 3.16: C I R C U L A R F I N D E S I G N  63  F I G U R E 4.1: I M P O R T A N T R E F E R E N C E F R A M E S  70  F I G U R E 4.2: C O M P L E T E L A U N C H C O N T R O L I N T E R F A C E S  72  F I G U R E 4.3: I N T E R F A C E S U S E D F O R P R O O F - O F - C O N C E P T D E S I G N  73  F I G U R E 4.4: M L C R O S T R A I N 3 D M - G S E N S O R  76  F I G U R E 4.5: T E S T I N G A P P A R A T U S S E T U P  78  F I G U R E 4.6: L I N K A G E G E O M E T R Y  79  F I G U R E 4.7: C O M P U T E D A N D S E N S E D A N G L E S F O R B A S I C C A S E  81  F I G U R E 4.8: S E N S O R E R R O R F O R B A S I C C A S E  81  F I G U R E 4.9: S E N S O R E R R O R F O R 3 0 D E G  82  F I G U R E 4.10: R E F E R E N C E F R A M E S W I T H S E N S O R  87  F I G U R E 4.11: L A U N C H O P P O R T U N I T I E S  93  F I G U R E 4.12: L A U N C H O P P O R T U N I T I E S N o V E R T I C A L M O T I O N  95  F I G U R E 4.13: L A U N C H O P P O R T U N I T I E S M I N I M A L T I L T  96  F I G U R E 4.14: O U T D O O R E X P E R I M E N T A L A P P A R A T U S  97  F I G U R E 4.15: C L O S E U P O F L A U N C H P L A T F O R M  98  F I G U R E 4.16: D E T E R M I N I N G A C T U A L L A U N C H A N G L E  100  F I G U R E 4.17: E X P E R I M E N T A L R E S U L T S N O O P T I M I Z A T I O N  101  F I G U R E 4 . 1 8 : L A U N C H A N G L E P R O B A B I L I T Y G R A P H ( P =60%)  103  F I G U R E 4.19: E X P E R I M E N T A L L A U N C H U S I N G O P T I M I Z A T I O N  104  vi  LIST OF S Y M B O L S Reference surface area Relative vertical acceleration of sensor  Aref ^relative  A et W  ^ z,accelerometer  Friction drag coefficient Coefficient o f skin friction Pressure coefficient o f the rocket F i n root chord F i n tip chord Diameter o f rocket at base of nose Earth to centre o f gravity displacement vector Earth to sensor displacement vector Diameter o f rocket at front of transition Diameter o f rocket at base of transition Sensor to centre o f gravity displacement vector Inertial earth reference frame Centre o f gravity reference frame Launch tube reference frame •. " Sensor reference frame Earth to centre o f gravity transformation matrix Earth to sensor transformation matrix Sensor to centre o f gravity transformation matrix Length o f fin at mid-chord line Length o f nose cone Length o f transition Rotation matrix from earth to launch tube reference frame  Cot c  f  (CN)R  CR  C  T  d dEG d£S  d  F  d  R  dsG F  E  F  G  F  L  F  S  HEG  H  E  H L  S  S G  F  LN  L  T  ^3*3  N R REG RES RSG  S  X X  B  X  F  X E , YE, Z E X L , YL, Z  X N XP X R XT  Wetted surface area Z-axis accelerometer reading from sensor  L  Number o f fins Radius o f body at aft end Earth to centre o f gravity rotation matrix Earth to sensor rotation matrix Sensor to centre o f gravity rotation matrix F i n semi-span Location o f rocket centre o f pressure from nose tip Distance from nose tip to fin root chord leading edge Location o f fin centre o f pressure from nose tip Axes o f inertial earth reference frame Axes o f launch tube reference frame Location o f nose cone centre o f pressure from nose tip Distance from tip o f nose to front o f transition Distance between fin root leading edge and fin tip leading edge Location o f transition centre o f pressure from nose tip  vii  a j a 8 0i 82 83 \\f m  o p t  n  Angle between vertical in launch tube reference frame and vertical axis in inertial reference frame; "launch angle" Safety limit launch angle Optimized launch angle R o l l angle o f the buoy relative to the earth Angle o f linkage bar #1 relative to the vertical Angle o f linkage bar #2 relative to the horizontal Angle o f linkage bar #3 relative to the vertical Pitch angle o f the buoy relative to the earth  viii  1. Introduction There is a common problem for weather forecasting - it is very difficult to make accurate weather forecasts for any region on the eastern coast o f an ocean. The affected areas include the west coasts o f North and South America, Europe, Africa, and regions in the Arctic. The source of this problem stems from both how weather behaves, and how forecasts are generated.  In general, atmospheric air circulation moves around the world from west to east at midlatitudes. Weather patterns that originate i n one place are convected by the prevailing winds and subsequently influence areas to the east.  Thus, when making a weather forecast, the most  influential data w i l l be obtained from the area to the west o f the target location. For example, to forecast tomorrow's weather in Toronto, the starting point would be data for today's weather at the Ontario-Manitoba border, one "weather day" to the west. These data would be entered into a computational fluid-dynamics weather-model, and the model advanced the desired time. This forecasting method is known as Numerical Weather Prediction ( N W P ) .  The atmosphere is a  highly non-linear system that is very sensitive to initial conditions such as weather conditions upwind.  The key to accurate forecasting is plentiful, reliable, weather data for the region west o f the desired forecast location. In North America, these data are gathered by various weather centres, which send up radiosonde balloons at regular 3-12 hour intervals.  A weathersonde  consists o f a sensor package, known as a sonde, to measure a combination o f temperature, pressure, humidity, and wind speed/direction at various elevations. The package includes some means o f transporting that sonde through the atmosphere.  1  For example, "driftsondes" are  dropped from stratospheric balloons, and "dropsondes" are dropped by aircraft.  Most  commonly, "radiosondes" are sent up on weather balloons, and equipped with a transmitter to relay the data.  Figure 1.1 shows all the North American locations where daily radiosonde observations ( R A O B s ) are taken. Particularly in the United States, the sounding locations are evenly spread throughout the country.  This results in a bounty o f atmospheric data available for weather  forecasting in the eastern U . S . Data are readily available for weather patterns 2-3 days away o f the east coast, providing a good starting point for 3-day forecasts. However, the western coast o f North America is not so fortunate. The only fixed radiosonde sounding location is in Hawaii. There are no other fixed radiosonde sounding locations over the expanse o f the ocean 1-3 weather days west o f the west coast o f North America. This dataless region has been named the Pacific Data V o i d . Studies have shown that 1-day forecasting capability for western Canada is 1  very sensitive to the impact o f the Pacific Data V o i d .  In addition, forecasts for central and  eastern Canada are also affected on a 2-3 day scale. 12CTW  180'W  90 N. P  Figure 1.1: RAOB sites in North America  2  Various methods have been employed to gather data within the Pacific Data V o i d . Environment Canada currently maintains seventeen weather buoys off the western coast o f Canada that gather meteorological data at approximately l - 3 m above sea level. However, the majority o f these buoys are close to land, with the farthest offshore being approximately 250 nautical miles west o f Vancouver Island. This corresponds to only about a half of a forecast day away from Vancouver.  Commercial aircraft are equipped with automatic sensors to gather  atmospheric data at elevations between 9-15 k m above sea level. W h i l e the buoys and aircraft succeed in reducing part o f the Pacific Data V o i d , Figure 1.2 shows that there is still a large dataless region i n the mid-troposphere. This missing cross-section o f elevations contains some of the most pertinent weather data for forecasting.  J>) Observation Sandwich" n  15  Aircraft obs. (AIREP, A C A R S , A M D A R ) (order of 250 / anal.)  1  P (I <Pa)  1  _  30  ii  Pacific Oi ta Void  | Sic. Obs.-Stilp S Buoy (order of 100-1 S W a n a ^  I 180'W  I  I  I  I  I  I  I  I  pffl  I 1*7 1 120'W  I  150°W  Figure 1.2: Observation "Sandwich"  2  Ships w i l l occasionally agree to send up radiosondes as they cross the North Pacific, and these sondes do gather data in the missing 1-9 k m range o f elevations.  However, it is not  practicable for the ships to occupy the same location permanently so that regular data may be taken. The ships also try to avoid areas with storm conditions. This practice is good for marine safety, but bad for weather forecasting because foul-weather data are the most useful. This  3  situation is known as "sympathetic data denial".  Environment Canada used to operate  permanent "weather-ships" in key locations in the North Pacific.  However, these were  decommissioned during the 1980s due to high maintenance and operation costs.  Clearly, a practical method is needed to eliminate the Pacific Data V o i d . The crux of the problem is the current inability to gather daily atmospheric data i n known locations. A method is required to provide "in-situ" weather instruments in the north Pacific.  This method should  deploy these instruments to gather regular atmospheric data within the dataless 1-9 k m range. In addition, given the value o f atmospheric data in storm-conditions, it is important that the system should reliably operate during the worst winter weather.  The Atmospheric Science Program at the University o f British Columbia has proposed a possible solution to this problem in the form o f a Rocketsonde-Buoy System. The proposed system consists o f a series o f deep-ocean buoys at fixed locations, each containing a battery o f weather rockets, called rocketsondes.  A rocketsonde consists o f a standard sonde package  contained within a rocket capable o f lofting the package to a height o f 6-8km. Each buoy w i l l autonomously fire one rocketsonde once a day. Each sonde w i l l separate from the rocket at its peak height and parachute back to the ocean, transmitting local atmospheric data back to the buoy.  The data w i l l i n turn be relayed to shore, and be entered into appropriate computer  forecast models. Conceptually, this system could eventually be adapted for other oceanic, datasparse regions around the world, such as the Canadian Arctic.  The proposed Rocketsonde-Buoy system, though rooted i n existing technology, poses several significant new engineering challenges.  The initial main challenge is to design an  appropriate rocket and guarantee that it w i l l launch vertically even during the most severe winter  4  storm conditions. Therefore, the purpose o f this research work is to evaluate the feasibility o f the Rocketsonde-Buoy System, focusing specifically on the aforementioned challenges.  The study w i l l examine whether the required technology exists, whether the system is practical so that all the desired system requirements can be met, and what compromises may be necessary in the final design. The chapters o f this thesis address these issues. Following this introductory chapter, Chapter 2 reviews the technical challenges and the existing data collection systems. Chapter 3 describes the design, testing and development o f high-altitude rockets for the proposed rocketsonde system.  Chapter 4 describes the design and mathematical basis of a  control system for ensuring vertical launching o f the rocketsondes, even in rough seas. Finally, Chapter  5 summarizes  the  conclusions drawn from  recommendations for future work.  5  the  work done,  and  gives  some  2.  Detailed Background  2.1  Weather Observation in the North Pacific  A s outlined in Chapter 1, the existence o f the "Pacific Data V o i d " seriously hinders accurate short-term forecasting for western North America, and long-term forecasting for the rest o f the continent.  To overcome this problem, accurate, "in-situ" weather data is required that is  gathered: a)  on a daily basis  b)  at the same known locations  c)  at a cross section o f altitudes above sea level  d)  during storm conditions as well as fair weather  2.1.1  Past and Existing Systems The Pacific Data V o i d is not a new problem. Methods have been sought to solve the  atmospheric data collection issues in the North Pacific for over thirty years. The most successful solution so far was Environment Canada's introduction o f a Weather Ship, known as "Station P A P A " , during the 1970s. The purpose o f Station P A P A was to be a fixed data source about one weather day offshore o f Northern British Columbia.  A s such, it took not only daily ocean  surface weather data, but could release daily radiosondes to gather regular upper atmospheric data. In this way, it fulfilled the first three requirements for a solution to the Pacific Data V o i d : daily data in fixed known locations for a variety o f altitudes.  It also had limited capacity to  gather data during storm conditions, although radiosondes are difficult to release during high winds.  6  The major complication with Station P A P A was the very high cost o f maintaining a permanent crew and ship at a fixed point in the ocean. In 1981, Environment Canada made the decision to terminate all weather ships due to the high cost. A t this time the International Union of Geodesy and Geophysics states that resulting from increasing operation costs that special weather ships cannot be relied on to provide continuous fixed-point observations.  The Union  also calls the termination o f Station P A P A "a serious loss to operational weather forecasting" . 3  Since the termination o f Station P A P A , both Environment Canada and the National Oceanic and Atmospheric Administration ( N O A A ) i n the United States have attempted to compensate through networks o f buoys. Environment Canada currently maintains 17 buoys off the western coast o f Canada.  Fourteen o f these buoys are shallow-water 3m discus buoys,  located in coastal waters, while three are deep-sea " N O M A D " buoys located approximately one weather day offshore . 4  Similarly, N O A A maintains the Tsunami network o f deep sea buoys  along the southern coast o f Alaska.  A l l these buoys are equipped to measure a variety o f  atmospheric and oceanic data at sea level on a daily basis, are in known locations, and are operational during storm conditions. However, they do not have a means to gather upper air data that would provide weather information at a cross-section o f altitudes.  The existing buoy  networks therefore contribute to a solution to the Pacific Data V o i d , but they do not eliminate it.  Ron  McTaggart-Cowan  5  in the Atmospheric Science Program o f U B C began  a  preliminary examination o f two new proposed solutions to the Pacific Data V o i d problem. The first was the Tethered Guided Balloon System, in which a weather balloon outfitted with a sonde  7  is permanently tethered above a fixed ocean buoy. A control system would be used to allow the balloon to continually climb to 2km and descend, over a range o f wind speeds. This system was limited in that it only gathered weather sounding data over a limited range o f altitudes. A more significant concern was that the weather balloon would be easily lost during storm conditions o f freezing rain and ice build-up, and no further development was taken on this proposed solution. The second solution examined by McTaggart-Cowan was the Rocketsonde-Buoy System, which is the basis o f this thesis. A feasibility study of the Rocketsonde-Buoy System has been funded by the Canadian Foundation for Climatology and Atmospheric Sciences ( C F C A S ) .  2.1.2  THORpex The Observing-system  Research  and predictability experiment  initiative o f the W o r l d Meteorological Organization ( W M O ) .  (THORpex) is an  A m o n g its Mission statements,  one o f the goals o f T H O R p e x is to "accelerate improvements in the prediction of high-impact weather on time-scales out to two weeks through international collaboration between  the  operational and research communities" . Within T H O R p e x a number o f research projects were 6  formed to look into improving 0-2 day predictability through developing in-situ observation systems for regions including the Pacific Data V o i d . The Rocketsonde-Buoy System (at U B C ) is one research project that has been included in T H O R p e x . Other T H O R p e x projects currently attempting to solve the Pacific Data V o i d problem include the Driftsonde Gondola ( N C A R ) , the Aerosonde and the deployment o f dropsondes from a Gulfstream G-4 aircraft ( N O A A ) .  The Driftsonde Gondola is being jointly developed by the National Centre  for  Atmospheric Research ( N C A R ) and the Naval Research Labs ( N R L ) . The Gondola, shown in  8  Figure 2.1, is released from Japan and outfitted with 24 dropsondes, which are standard devices equipped with parachutes.  Weather systems would push the Driftsonde east, and it would  autonomously drop one sonde every 6 hours. The use of dropsondes allows the system to gather atmospheric data at a cross section of elevations within the Data Void, and the system is a more affordable alternative to manned permanent stations like the weather ships. The Driftsonde Gondola developers, however, do acknowledge one severe limitation: while it can drop a sonde a prescribed times, it cannot do so at prescribed locations. This means that the system cannot be used to gather data once a day in the exact same location, because it is unknown exactly where in the Data Void the Gondola will drift. Also, it is unclear how the system will behave in storm conditions. The Gondola may either be pushed away from storm area, or be damaged if caught in high winds . Also, the Driftsonde developers have been denied permission to drop the sondes 7  through the commercial airlanes.  Figure 2.1: Driftsonde Gondola  9  Another related project that aims to gather data within the Pacific Data Void is the Aerosonde, developed by the Aerosonde Robotic Aircraft Company in Boulder Colorado. The Aerosonde is a small aircraft, with a wingspan in the order of 2.5m and equipped with meteorological instrumentation, that can be flown remotely over an ocean (see Figure 2.2) . The 8  development team proposes that an aerosonde may be sent daily from Hawaii through the Pacific Data Void to gather data in the 5-6km altitude range. The Aerosonde has already completed successful test flights across both the Pacific and Atlantic oceans. It is capable of following the same route repeatedly, so can take the same data daily at known locations. The Aerosonde, however, has not proven very successful during ocean storm conditions and tends to crash during high winds or icing conditions. In addition, while it provides some of the missing atmospheric data in the Pacific Data Void, namely the 5-6km altitude range, it misses equally important data below 5km and above 6km.  10  The third ongoing project, the use o f Gulfstream G-4 aircraft to deploy dropsondes, best meets the in-situ data requirements.  This solution was proposed by the National Oceanic and  Atmospheric Administration ( N O A A ) .  N O A A suggests a G-4 aircraft, equipped with a system  capable of monitoring up to 8 dropsondes simultaneously, cross the Pacific Data V o i d on a daily basis and deploy the dropsondes . 5  The G-4 aircraft can fly at a ceiling o f 12km, so that the  dropsondes would both cover a wide cross-section o f altitudes and could be dropped into lowerlevel storms.  This proposed solution meets all the desired weather data criteria to solve the  Pacific Data V o i d problem. The only significant limitation is that this system, like the weather ships, is very expensive to maintain. A G-4 aircraft required a crew o f 7-10 people, plus high fuel and maintenance costs involved in daily cross-Pacific flights. While technically an excellent solution, it may prove to be logistically unrealistic, except  for special-event  targeted  observations.  2.2  Rocketsonde-Buoy System Objectives and Requirements The main objective o f the Rocketsonde-Buoy System is to provide in-situ data i n the Pacific  Data V o i d that, unlike the Driftsonde Gondola and Aerosonde, fulfills  all four desired  characteristics. A secondary goal is to meet this objective i n a cost-effective manner.  •  First, the system should be capable of gathering daily data.  This means that the buoy  should be equipped with one rocketsonde for every day between servicing, plus allowance for spares. Essentially, i f the buoy is to be serviced once every 100 days, it should contain more than 100 rockets.  11  The Rocketsonde-Buoy System meets the requirement that the data be gathered in a known location. The buoys must be anchored, for large drifting buoys pose a major marine safety concern. The deep-sea N O M A D buoys operated by Environment Canada have a drift radius o f only about 3 k m so can be considered to be i n a fixed location. 3  G P S equipment on the buoy verifies the fixed location.  The third data requirement is that the system gathers data at a cross-section o f elevations. Rocketsondes are designed to hit a certain altitude, and then release a sonde that drifts down on a parachute or streamer, thus collecting the desired cross-section o f data. It is desirable that as many altitudes as possible within the Pacific Data V o i d be covered by the system, so that the rocketsonde for this system must go as high as realistically possible before deploying the sonde.  Aircraft traffic over the Pacific Ocean starts at  approximately 9 k m above sea level and it is obviously then not desirable for the rockets to exceed this altitude. Buoy System is 6-8km.  Therefore, the desired altitude o f rocket for the RocketsondeThis altitude range exceeds the capability o f the Aerosonde,  which only gathers data below 6km, but does not cause safety concerns for aircraft in the region.  The final requirement is that the system be capable o f gathering weather data during North Pacific storm conditions. This results in very specific wind, wave and temperature operation conditions for the Rocketsonde Buoy.  The Canadian Marine Environmental  Data Services ( M E D S ) records weather data taken from each o f the three deep-sea  12  N O M A D buoys off the west coast o f British Columbia. These buoys are in at latitudes similar to where the Rocketsonde-Buoys would be deployed, and therefore the M E D S data can be used as a guideline typical winter weather conditions, when storms are most likely to occur.  Figure 2.3 shows an example set o f data for the North N O M A D in  February 2 0 0 2 . 10  Examining December to March data for all three N O M A D buoys in 2002 and 2003, generalizations can be made regarding winter storms.  G i v e n average bad weather in  winter, the Rocketsonde-Buoy System should be operational during storms i f it can be designed to operate i n wind conditions up to 22m/s ( W S P D 1 ) with occasional gusts ranging from 25-28m/s, and a wave period as short as 5 seconds ( V T P K ) .  This wave  period appears only a couple o f days a month, and more often the wave period ranges from 8-17 seconds. A l s o , all equipment should be able to withstand an average temperature o f 5 Celsius (SSTD1).  A l s o , the system may be subjected to wave heights up to 10m ( V C A R ) ,  although it is still unknown how that w i l l affect the buoy rocking motion.  C46184  01  06  11  16  21  26  03  Date VCAR  VTPK  GSPD1  F i g u r e 2.3: N o r t h N O M A D E n v i r o n m e n t a l C o n d i t i o n s , F e b 2002  13  9  The above requirements allow the Rocketsonde-Buoy to meet its primary objective regarding the desired characteristics o f the meteorological data it collects. However, the issue o f cost still needs to be examined. The biggest cost consideration comes from the expense of using a ship to implement and maintain this system. The only Canadian Coast Guard ship currently capable o f deep sea buoy deployment in the Pacific, the C . C . G . S . Laurier, costs $13k per day . 11  Equivalent American Coast Guard and military ships cost up to $25k per day.  Therefore to  make this system financially feasibly, requirements have to be added that minimize ship time. There are several that w i l l do so.  First o f all, the Rocketsonde-Buoy system must be an autonomous, stand-alone system. It must operate safely and effectively without daily supervision from a nearby ship.  Secondly,  the system should be designed so that it only needs to be serviced once per year.  This  requirement then means that the system must be outfitted with approximately 400 rockets to allow for daily launches, plus spares. Finally, the system should be o f a size small enough that it can be lifted onto a ship by a derrick rather than requiring to be towed by the serving ship to be deployed. It generally takes three times as long to tow a buoy into position than to travel with it on-board. Deploying a 12-m buoy one weather day offshore, which takes approximately one day for a N O M A D , would take three, and thus increase the deployment cost by at least $26k.  14  Figure 2.4: N O M A D B u o y The largest buoy commonly manufactured that can be brought onboard the Laurier is the 6m N O M A D buoy, shown in Figure 2.4.  N O M A D buoys are currently used in deep sea  applications, so are an intelligent choice for the base o f this system. However, they have a safety margin that only allows approximately 1600kg extra weight on the b u o y . M a k i n g allowances 12  for other equipment and safety margins, the entire rocket and launch system battery on the Rocketsonde-Buoy System would therefore have to weigh less than 800kg. It then becomes a major requirement that the rocketsondes designed have minimal mass.  Similarly, the buoy has  limited deck space it is initially assumed that the rockets must have as small a diameter as possible.  Finally, separate from all data and cost concerns, the Rocketsonde-Buoy System must also fulfill reasonable environmental and safety requirements.  The system must operate in a  manner that has a minimal impact on the ocean environment.  This is particularly important  15  given that the rocketsondes are disposable and w i l l drop into the ocean.  A s for safety  requirements, the system must operate in a manner that poses no danger to surrounding marine traffic, aircraft or the deployment and servicing ship and crew.  Observations were made during the 2002 Canadian Coast Guard/ Environment Canada Annual O D A S buoy servicing trip regarding safety, environmental, and logistical concerns for the Rocketsonde B u o y System: •  First, it became clear that the Rocketsonde B u o y system must be serviced by an experienced buoy crew, such as the Canadian Coast Guard, even i f the system is located in international waters.  Deploying a buoy requires precision and caution as other  international buoy programs average 2-3 deaths each year. •  Second, even when using an experienced buoy servicing crew, the buoy should be a conventional design with which the crew is familiar. This increases the safety for the crew i f they have to react to off-normal situations.  A standard N O M A D buoy would  therefore be a good choice. •  Third, the system components located on top o f the buoy, namely the rockets and launch apparatus, need to be very durable and easy to replace i f damaged. Buoys w i l l frequently hit the side o f the ship hull during deployment i n high winds. A cartridge-type system, where the entire battery o f rockets is a single replaceable piece, may be preferable.  •  Fourth, the buoy construction/adaptation  should be done by an experienced buoy  designer familiar with the environmental conditions and deployment issues. •  Fifth, a data buoy system, like a network o f Rocketsonde buoys, depends on having a dedicated person to keep it running at a high standard as it needs constant attention.  16  •  Finally, the system should include a diagnostic program which allows the servicing vessel to check operations for 4-6 hours after deployment. This provides a way to check functionality without having to retrieve the buoy, which is a time consuming and costly process.  2.3  Additional Rocketsonde Requirements The first successful weather rocket tests were performed i n 1963 using a 5 inch diameter,  solid-fuel rocket made by Bristol Aerojet. The first rocketsondes were developed in 1965 as a means to gather upper atmospheric data i n the 25-60km range unreachable by conventional weather balloons . The original sondes were capable o f recording only temperature data. 13  The most recent rocketsonde development has been performed by Vaisala in Boulder Colorado.  They have developed a low altitude rocketsonde for the United States N a v y  " M O R I A H " project.  The goal o f M O R I A H is to monitor near-sea-level weather conditions  around N a v y ships which may interfere with their radar systems. interested in high resolution data 0 - l k m above the ocean.  The N a v y is particularly  The current Vaisala sondes weigh  approximately 200g and are equipped with temperature, pressure and dual humidity sensors, as well as a short-range transmitter, all powered by a 9 V Lithium battery . 14  Each sonde costs  approximately $300 U S D .  The  current Vaisala rocketsonde cannot be used  in its current  design for  the  Rocketsonde-Buoy system because it is designed to reach a maximum altitude o f only 1km. The requirements o f the previous section set the altitude goal o f the system to be 6-8km. Therefore,  17  significant redesign w i l l have to be done regarding the rocket component o f the rocketsonde. The Vaisala Instrumentation group is currently a world leader i n meteorological instrumentation and has developed a congenial relationship with the U B C Rocketsonde-Buoy System team. Therefore, it is currently assumed that the sonde package used i n the Rocketsonde-Buoy system w i l l be very similar to their current instrument package.  This sonde then determines several  requirements for the new rocket design.  First, the Vaisala sonde sets the rocket payload mass to 200g.  This is quite a small  payload for a high-powered rocket, so should not really be a design issue. More importantly, the Vaisala sonde has been designed to withstand forces only up to 50 G . This means that the new rocketsonde must never exceed 50 G o f acceleration at any time during its flight i f the current sonde is used.  This places restrictions on both the specific motor used in the rocket, and the  rocket's aerodynamic design.  Safety concerns place additional requirements on the rocketsonde design. The sonde w i l l release from the rocket at quite a high altitude, and, i f not properly deployed, both the sonde and rocket body could be quite dangerous i f they should free-fall. dependable recovery system.  The rocketsonde requires a  Another important safety concern involves the stability o f  propellant used i n the rocket motors, particularly since they w i l l sit onboard the buoy for a year.  2.4  Launch System Requirements The Rocketsonde-Buoy System is not the first project to require a rocket to be launched  from an ocean platform.  The most high-profile system currently i n operation is the Sea Launch  18  project undertaken i n 1993 by the Boeing C o . and several European partners . 15  Sea Launch  consists o f a former North Sea drilling platform that has been placed on the equator and adapted into a launch pad for rockets carrying geosynchronous satellites.  The project bears many  resemblances to the Rocketsonde-Buoy System, but it is on a much larger scale, so little of the technology is adaptable.  The platform for Sea Launch is 436 ft long, so is not as sensitive to  wave conditions as a 6m N O M A D buoy. A l s o , Sea Launch has the ability to indefinitely delay a launch during poor weather conditions. Its launch system does not have to be as versatile as the one required for the Rocketsonde-Buoy System.  O n a smaller scale, missiles and rockets have been fired from ships for decades. M O R I A H is probably the most closely-related ongoing project.  However most ship-board  rocket systems, like the Vaisala rocketsondes used i n the M O R I A H project, use mechanical systems and human control to position them shortly before launch.  Mechanical means o f  positioning the rockets i n the Rocketsonde-Buoy system are extremely difficult given the effects of salt water corrosion on any moving parts over the course o f a year. The Rocketsonde-Buoy system w i l l not receive daily maintenance like the launching systems on naval vessels.  The Rocketsonde-Buoy System is unique in that it must autonomously ensure close to vertical rocket launches i n a variety o f weather conditions from a rather small ocean platform. It therefore requires a launch control system capable o f monitoring the pitch and roll of a N O M A D buoy during 28 m/s winds and 10 m seas, and then be able to use that information to achieve reliable vertical launches. The system must also not require servicing more than once a year, so should not include any mechanical components. There is currently no such system available.  19  Gilhousen  16  performed a series o f tests using a buoy hull/mooring dynamics model to  examine the pitch o f a N O M A D buoy during particular wind-wave conditions. His results show that a N O M A D subject to 10m waves should have an average maximum pitch angle o f approximately 12 degrees. Information gathered from M E D S , and discussed in Section 2.2, gives the smallest period to occur in winter storm conditions to be 5 seconds. Given that the 12 degree pitch is an average value, the Rocketsonde-Buoy system should be able to successfully launch when the buoy is rocking with a maximum pitch angle o f 15 degrees and a 5 second period. This corresponds i n Gilhousen's work to 16 m wave conditions, and therefore is a good upper limit for the worst conditions the buoy should encounter while being expected to operate.  2.5  Summary of System Requirements This chapter has examined existing systems and the technological challenges in the design o f  the Rocketsonde-Buoy System. In the process, it developed a set o f design requirements for the system. •  The Rocketsonde-Buoy System ideally w i l l contain a battery o f 400 rockets.  •  The system w i l l use a conventional N O M A D buoy.  •  The system placed on the buoys must then weigh less than 800 kg.  •  The system is to be autonomous.  •  The buoy should require servicing only once a year.  •  It should remain operational i n 28 m/s winds and temperatures below 5 degrees C .  •  It also should be designed to have minimal environmental impact.  20  The added requirements for the rocketsonde design are: •  The rocketsonde w i l l use a Vaisala sonde as its payload, which weighs 200g.  •  The rocket should reach an altitude o f 6-8km.  •  The sonde w i l l be safely deployed at apogee.  •  The rocket must stay under 50 G o f acceleration at all times.  •  The rocket should be o f minimal mass and size to allow 400 rockets to be stored on a N O M A D buoy.  •  The propulsions system must be stable enough that the rockets can be safely left on an ocean buoy over the course o f a year.  The system also requires a launch control system that w i l l ensure near vertical launch. Its requirements are as follows: •  This system should be operational for buoy motion with a maximum amplitude o f 15 degrees and 5 second period.  •  The launch control system should include no mechanical parts.  The above requirements w i l l govern the design decisions made during the research process discussed in the following chapters.  21  3.  Rocketry Design  3.1  Theoretical Background  3.1.1  Basic Rocket Design  There are six main components to any basic rocket design, such as that shown i n Figure 3.1  •  The nose cone. Nose cones are generally made from plastic, and in some cases, wood. They tend to be hollow to minimize weight, but can be filled to change the rocket's centre o f gravity.  The most common rocket nose cone shapes are conical or ogive  (shown in Figure 3.1) because they are the most aerodynamic for subsonic and transonic flight.  Other designs include parabolic or hemispheric, which are used to optimize  supersonic flight.  The nose cone also contains a small shoulder o f a small diameter  which fits snugly into the rocket body. The nose cone separates when the rocket reaches apogee.  •  The body tube. Body tubes are usually just hollow cylinders made o f an appropriate material.  Features inside the body tube may include blocks to secure the motor, lugs for  the launch device, vent holes to allow pressure equalization, fasteners for the parachute chords and electronics such as altimeters.  Figure 3.1: Conventional Rocket Design 22  •  Thefins.These are possibly the most important component. The rocket's fins ensure its stability during flight. This feature w i l l be covered i n more detail i n Sections 3.1.2 and 3.1.3.  Figure 3.1 shows a very conventional fin design. Conventional rockets tend to  have 3-6 trapezoidal fins with either sharp o f rounded corners.  Unconventional fin  designs w i l l be dealt with later i n this chapter.  •  A parachute or recovery device. The parachute is usually attached to a "shock cord" made from an elastic material, which connects the nose cone to the main rocket body after separation. The shock cord must be long enough that it can absorb all the shock o f separation and not snap. The parachute, which is designed based on the rocket's weight, keeps the rocket from descending too fast.  •  A payload (optional). A n example payload is the Vaisala sonde package used for this project. Other examples would include altimeters and tracking transmitters. The payload needs to be either attached securely within the rocket body, attached securely to the rocket shock cord, or have its own parachute after deployment.  •  The engine. This is fitted in the main body, (not shown in Figure 3.1) Engine design w i l l be discussed i n more detail later in the chapter.  3.1.2  Centre of Pressure The centre o f pressure (CP) o f a rocket is the point at which, as the rocket moves through  the air, all forces due to air pressure can be considered to act.  Its location depends on the  distribution o f lift on the rocket. In most designs, the rocket fins generate the major portion o f the rocket's lift, so their design has the greatest influence on the location o f the C P .  23  The centre o f pressure for a rocket can be written as the sum o f the pressure coefficients for all the individual rocket components . These w i l l be defined using the Barrowman method. 17  (Cj^Rocket  =  (CN)Nosecone + (Ci^Transition + (Cj^Fins  (3"1)  Since the body tube is symmetric, it does not influence the centre o f pressure, unless it changes in diameter. The diameter change is represented by the transition term.  The location o f the centre o f pressure, measured from the nose o f the rocket, then becomes  where X N , X T , and X F are the distances between the nose cone tip and respectively the centre of pressure o f the nose cone, transition and fins.  The centre o f pressure o f a rocket depends on the airflow, so its exact location can only be determined experimentally. However, given the time it takes to build a model and perform wind tunnel tests, it is usually more practical to use a mathematical model to approximate the location o f the centre o f pressure.  The most commonly used method used to determine the C P location was designed in 1967 by James Barrowman, o f the N A S A sounding rocket branch.  The Barrowman method  contains a set o f algebraic equations capable o f determining the C P o f a rocket flying sub sonically at a small angle o f attack (0-10 degrees) to a high order o f accuracy. This Barrowman  24  method, although designed for subsonic flight, can also be accurately applied to rockets with a maximum velocity o f M a c h 2, given that their fins have a minimal thickness to length ratio.  According to the Barrowman method, the individual centre o f pressure locations and values are determined as follows  Nose Cone:  (CN)N  (3-3)  2  _  For Cone: X For Ogive:  X  = 0.466L  N  (3-4)  0.666LN  N =  (3-5)  N  Transitions d*  f  (C ) =2 N  T  Vd j  (3-6)  V « )  d  F  X  —X H  T  P  7  P  L  T  d  a  1+-  3  (3-7)  d  f  1-  V  __£  J  Fins:  1+ -  (C ) N  4N\  R  S_  \dj  S+R  F  1 + J1 +  1U yC +C R  Y X  _ y " ~  X  "  , X {C +2C ) R  +  3  R  1  T  6  T  where L N = length o f nose d = Diameter at base o f nose  25  j  T  {C C )R+  (C C ) R+  (3-8)  T  (C C ) f i  r  (c,+c ). r  (3-9)  dV dR LT Xp CR CT S Lp R X XB N R  = Diameter at front o f transition = Diameter at rear o f transition = length o f transition = Distance from tip o f nose to front o f transition = fin root chord = fin tip chord = fin semi span = length o f fin mid-chord line = radius o f body at aft end = distance between fin root leading edge and fin tip leading edge parallel to body = Distance from nose tip to fin root chord leading edge = number o f fins  3.1.3  Rocket Stability There are two points on a rocket that determine whether a rocket w i l l fly in a stable  manner.  These are the rocket's centre of pressure ( C P ) , which has been explained in the  previous section and its centre o f gravity ( C G ) . The fastest way to find the C G is by finding the point at which the rocket balances.  The stability o f a rocket depends on the location o f the centre o f pressure in relation to the centre o f gravity. A stable rocket is able to correct for perturbations i n its direction without tumbling uncontrollably.  Figure 3.2(a) shows a rocket that has been slightly perturbed during  flight. The engine still causes thrust i n a direction parallel to the body tube, acting on the centre of gravity.  The black arrows show how the thrust and gravity forces on the rocket can be split  into a horizontal and a vertical component.  A t the same time, the perturbation causes the rocket to alter its flight path, it also causes lift and drag forces on the rocket, located at the centre o f pressure (Figure 3.2(b)).  The drag  forces are generally minimal compared to the lift, so are not shown. The lift force, as shown in  26  the Figure, acts to pull the rocket body below the centre o f gravity back towards vertical, thus correcting its flight. If the centre of pressure were located above the centre o f gravity, this same lift force would push the rocket nose farther away from vertical, and eventually cause it to tumble.  (a)  (b) Figure 3.2: Effects of CP on Rocket Stability  Therefore, the stability condition for a rocket is that the centre of gravity must be a safe distance forward of the centre of pressure. A general design rule is that the two points must be at least one body diameter apart. Given that the location o f the centre o f pressure is a key feature o f a rocket's stability, and that the rocket's fins essentially determine the location o f the C P , fin configuration is one o f the main focuses i n a rocket's design.  27  3.1.4  Circular Fins or "Ringtails" The previous section demonstrated how crucial it is to move the centre o f pressure o f a  rocket as far back as possible.  A s discussed in 3.1.3, the components that have the largest  influence on the location o f the C P are the fins.  B y increasing the magnitude o f the pressure  coefficient o f the fins, the C P location of the entire rocket can be brought farther back, thus making the rocket more stable.  Examining the Barrowman equations, one notes that that pressure coefficient o f the conventional rocket can be increased by lengthening the root chord o f the fins, by increasing their number, or by increasing the distance the fins protrude from the rocket's body.  Each  change, however, comes with a cost. Lengthening the rocket fins is a good solution, but there is a limit to how long the fins can extend along the body before they start to move the C P forward, rather than back.  Increasing the number o f fins increases the drag, and thus decreases the  altitude the rocket can reach on a given engine. Extending the fins further out does not affect the drag as much, but results i n a rocket that is a much larger diameter. This becomes an issue for the Rocketsonde-Buoy system, because one o f the requirements is that the rocket takes up minimal space on the buoy.  Circular fins, or "ringtails", are a fin design that attempts to increase the rocket stability, by moving the C P farther aft, while not significantly increasing the rocket's diameter.  Circular  fins consist o f a short piece o f tube, larger in diameter than the body tube, fixed around the end of the rocket. Small conventional fins are used to hold the circular fin i n place. The circular fin  28  design increases the fin area away from the rocket body, thus increasing the fin pressure coefficient, without extending the fins away from the rocket body.  The U S Department o f Defense has performed a series o f wind tunnel tests on ringtail fin designs . 19  They conclude that a ringtail w i l l produce, at both subsonic and supersonic speeds,  approximately twice the restoring force of a set o f 4 conventional fins with equal span and root chord. However, ringtails greatly increase the base drag o f the rocket. This is less problematic at subsonic speeds where interference between the ringtail and jet plumes decreases some o f the drag, but it is a major consideration at supersonic speeds.  3.1.5 Drag Five different types o f drag affect a rocket in flight. Each contributes to the drag coefficient C D . Increasing C D decreases the altitude a rocket can reach on a given engine .  •  Friction Drag is caused by the interaction between the rough surface o f the rocket body, including fins, and the passing air. It can be approximated as  16  CDf Cf(A /A ef) =  wet  r  where Cf = the coefficient o f skin-friction  A t = surface wetted area we  Friction drag can be minimized by giving the rocket a smooth finish.  •  Pressure Drag o f a rocket is mainly influenced by the shape o f its nose cone. W i n d tunnel testing at various speeds has shown the pressure drag coefficient of a ogive or conical nosecone at transonic speeds to be on the order o f 0.2  29  •  Interference Drag on a rocket is generated where two surfaces are intersected at an angle, and the flow patterns thus interfere with each of. For rocket design, this drag is generally caused by the interference between the flow over the fins and that over the rocket body. This drag component increases greatly on a ring-tailed rocket as the number o f angles double.  Interference drag can be slightly minimized by keeping the  fillets  attaching the fins small, to minimize the joint surface area.  •  Parasitic Drag is caused by launch lugs and other objects being attached to the outside surface o f the rocket.  It can be minimized by eliminating all attachments, or keeping  them as small as possible.  •  Drag due to Lift occurs when the rocket fins are not at zero degrees angle of attack. First, it is important to properly align the fins during construction so that they are at zero degrees i n relation to the rocket's vertical direction. Secondly, this drag is minimal when the rocket is flying straight, so can be minimized by ensuring the rocket's stability.  3.1.6 Engine Design Rocket type is classified by the total impulse o f its engine.  "High-powered" rockets are  rockets that are launched using a motor with a total impulse motor larger than 160 N s . Generally, for rockets with altitude goals higher than l - 2 k m , such as the proposed rocketsonde, a high powered rocket motor is required. Three types o f rocket engines are commonly used in high powered rocketry: solid propellant, liquid propellant and hybrid motor systems.  30  Solid Propellant Solid propellant rocket motors are based on the classical "black powder" mixture used in cannons in the 19  th  century.  In fact, most current model class (small) motors (0-80 N s total  impulse) use a similar mixture o f charcoal, sulfur and potassium nitrate.  High-powered solid  rocket motors usually use a composite mixture, which include 75% propellant and 25% additives and binders. Binders include materials like polystyrene to allow the propellant to be pressed into hard structures and cast into molds. High powered propellant mixtures are generally proprietary and are often similar to plastic explosives. For example, the Pro-38 motor line manufactured by Cesaroni Technologies is made from a "proprietary thermoplastic propellant" (TPP). Common propellants used include ammonium perchlorate and ammonium nitrate  21  Figure 3.3 shows the cross section of a typical, off-the-shelf, solid propellant rocket engine.  The propellant is cast into "grains" with an interior circular cavity.  A ceramic or  graphite nozzle is then fitted on the end o f the grains. The rocket igniter is placed against the end of the cavity. When a large current is passed through the igniter, the propellant ignites and flames spread throughout the cavity. The rocket motor then burns outward from the centre, with the exhaust ejecting through the nozzle.  ejection charge  nozzle  Figure 3.3: Solid Propellant Motor Cross-Section  31  Liquid Propellant The idea o f using a liquid rocket engine was first proposed by the Russian mathematician Ziolkousky i n 1903. L i q u i d motors consist o f two separate tanks for the fuel and oxidizer. A turbine drives a pump for each tank to inject the liquids into a separate thrust chamber. The mixture is then ignited within the chamber and the flow goes through a nozzle.  Liquid  propellant motors are generally more complicated than solid propellant, as they require a separate combustion chamber and a system to feed the propellant into the chamber .  Hybrid Propellant Hybrid motor systems combine solid and liquid propellant motors by using a solid fuel and a liquid oxidizer.  The igniter is placed in the solid propellant, which acts as the combustion 99  chamber, and just prior to ignition, an oxidizer such as N2O is injected into the system . Hybrid motors keep the fuel and oxidizer separate until just prior to launch, so are generally considered to be safer than solid or liquid propellants. A l s o , the tank containing the oxidizer can be located on the launch pad, rather than within the rocket, saving mass.  3.2  Initial Design Decisions Before doing any simulation or experimental work, several key decisions were made  regarding the rocketry design. These were done early i n the process to eliminate some o f the variables and thereby simplify the design process.  32  Construction M a t e r i a l The first decision made was the choice o f material to use for the rocket body.  High  powered rockets require a material that is strong enough to withstand the accelerations and vibrations o f launch. This eliminates model rocketry materials such as cardboard and wood. Common construction materials for high powered rockets include carbon composites, fiberglass, Kevlar and metals such as copper, aluminum and steel.  A second consideration is that the rocket must have minimal mass to prevent the buoy from sinking when carrying a full load o f rockets. Therefore, the lighter composites, and light metals such as aluminum, are a preferred option. A lighter construction material also enables a rocket to attain a higher altitude on the same motor.  The final factor that determined the construction material was the desire that these rockets have minimal environmental impact.  The rocketsondes are designed to deploy the  payload at altitude, but then all components w i l l fall into the ocean.  Carbon composites and  fiberglass degrade very slowly, while metals break down quite quickly i n salt water. A l u m i n u m is the lightest o f the metals commonly used in rocketry, and it also has a moderate cost. Therefore it was selected as the construction material. U s i n g aluminum rocketsondes decreases possible electrolysis occurring i n the salt water environment between the rocketsonde and the aluminum buoy superstructure.  Using aluminum is also more cost-effective.  A n aluminum  high-powered rocket w i l l cost in the order of $50-100, while a similar rocket using carbon composites can be several hundred dollars or more.  33  Engine T y p e Solid propellant rocket motors appear to be the best choice for the given system.  Liquid  propellant systems are unnecessarily complex for smaller scale high powered rockets, such as the rocketsondes. L i q u i d propulsion systems are designed for reusable rockets, so are a much more expensive option than single-use solid propellant rocket motors. There are also possible issues with safely and effectively storing fuel and oxidizer tanks for up to a year.  Similarly, hybrid  motor systems are an expensive choice for non-reusable rockets, and it would be difficult to store sufficient N2O on the buoy for 400 launches.  Solid propellant engines have the advantage that they are self-contained and easy to install in the rocket. H i g h powered rocket engines have an ignition point above 450 °F, so are relatively stable, and safe to store on an ocean platform. They do have the major disadvantage that they are ineffective i f they get wet.  The rockets must therefore be protected from  environmental conditions which could damage the propellant grains.  Again, using a solid-propellant motor is also the most cost-effective choice. H i g h powered solid motors range from $60-250 each. Liquid and Hybrid systems are designed to be reusable and cost over $300 per launch, plus up to $1500 in additional equipment on the buoy.  Deployment M e t h o d A s stated in Chapter 2, it is necessary that the rocketsonde reliably deploys the sonde and rocket components so that they do not free-fall from the maximum elevation reached.  To  achieve this, the parachute should be deployed close to the rocket's apogee, when its vertical  34  velocity is close to zero.  Deployment significantly before or after apogee may cause the  parachute lines to break or the parachute to tangle because the rocket has excessive velocity.  Parachute deployment is triggered by a recovery charge being ignited within the rocket body. This charge causes an increase o f internal pressure i n the rocket, which pops off the nose cone and pushes out the parachute.  There are two common methods used to ignite the  deployment charge. The first uses a small ignition system in tandem with an altimeter. altimeter monitors the decrease in atmospheric pressure as the rocket ascends.  The  When the  pressure begins to increase, which occurs when the rocket passes apogee, the altimeter ignites the recovery charge. This is a reliable method for deployment to always occur close to apogee. However suitable altimeters cost $150-$200 C A N , and are more appropriate for reusable applications.  The second common way to ignite the deployment charge is by including a delay i n the solid propellant motor. A s shown i n Figure 3.3, after the solid propellant burnout, a delay charge is ignited. The burn time o f this delay is set to approximate the amount o f time the rocket coasts before attaining apogee.  A s the rocket reaches apogee, the delay should finish burning, and  ignite the recovery charge. This method requires that a good estimation be made o f the rocket behavior, particularly its time to apogee. Despite the uncertainties associated with estimations o f required time delay, the method is usually effective. triggering parachute deployment a popular one.  35  The low cost makes this method o f  The decision was made, given the disposable nature o f the project rocketsonde, to use a motor delay charge for recovery.  This decision w i l l mildly affect the motor selection in the  design process, as only motors that include delays w i l l be considered. However, the decision may be easily altered with no major final design changes i f it is later decided that the cost o f an altimeter is worth the reliability o f a safe deployment.  Launch Apparatus When a rocket is launched, it starts from zero velocity and rapidly accelerates.  During  the initial moments, before the velocity has sufficiently increased, it is possible that the stabilizing effect o f the fins may not be enough to correct any sideways perturbations to the rocket motions. The purpose o f a launch apparatus is to ensure that the rocket travels vertically until it reaches a velocity at which it becomes stable.  This means that when it leaves the  launching apparatus, it should be traveling fast enough that outside perturbations should not substantially disturb the rocket's flight. A typical desirable speed is 18-20 m/s.  Generally, rockets have lugs or pins along one side that allow them to be launched from either rods or rails. environment.  The problem with such arrangements is that they are open to the  This means that on an ocean platform with 10 metre waves, the rocket w i l l be  frequently exposed to salt water.  This w i l l cause the aluminum body to corrode, the sonde  package to be damaged and the solid propellant motors to be destroyed.  The logical alternative is to contain each rocket within a sealed tube on the buoy, and to use this tube as the launching apparatus. These tubes can either be circular, or honeycombed to  36  take up less area.  The tube cover w i l l either consist o f a membrane that the rocket w i l l be able  to punch through, or be removed with a small explosive prior to launch. The tube bottom w i l l similarly be removed so that the launch tubes w i l l not fill with water once empty. Logistically, tube launching appears to be a good solution to the environmental problems.  However, it is  unknown h o w the enclosed environment may affect engine performance, or i f the tube w i l l interfere with the rocket's flight.  3.3  Simulations  3.3.1 Introduction to RocSim A l l project simulations were performed using R o c S i m 5.021: a rocket simulation program designed by Apogee Components Inc. This program is the standard among the amateur rocket community as it is very user-friendly and accurate for small-scale projects with velocities less than M a c h 2.  R o c S i m consists o f four elements: •  the main program to run launch simulations for a given rocket design,  •  a rocket design program that allows most conventional designs to be modeled,  •  an engine editor to create thrust files for new rocket motors, and  •  an engine database compiler.  R o c S i m contains a database o f commonly available commercial rocket motors to use i n the launch simulations. It also comes equipped with an engine editor to enter new engine designs, or to design unique prototypes.  The engine editor requires information on the motor dimensions,  37  mass, and a thrust curve. The thrust curve shows the variation o f motor thrust over the course o f its burn time.  A l l centre o f pressure calculations are performed by R o c S i m using the Barrowman method outlined i n Section 3.1.2. This means that the stability analysis is only valid for rockets with a maximum velocity less than M a c h 2.  The drag coefficient is calculated as four separate components: •  nose a n d body: includes the pressure drag for the nose cone and friction drag for both components  •  base: is the pressure drag caused by l o w pressures at the rear o f the rocket  •  fin:  includes the interference drag for the fin joints as well as the pressure drag caused  by the fins, the friction drag on the fins and the drag due to lift •  l a u n c h lug: is the parasitic drag due to the lugs. For most o f the simulations, lugs are not used, so this coefficient w i l l be zero.  These components are all recalculated for new velocities at each step of the simulation. A l s o , for transonic and supersonic speeds (greater than M a c h 0.8) the drag Coefficient is adjusted to simulate the sharp increase i n drags at those speeds.  R o c S i m allows launch conditions to be set, including the launch elevation, launch angle, wind speed, the altitude at which the wind begins, and environmental conditions such as temperature, pressure and humidity.  38  The simulation software in RocSim plots the vertical position, velocity, and acceleration during the rocket flight, as shown in Figure 3.4. From this one can predict the apogee achievable with a given rocket using a given engine. It also indicates the maximum G-forces and velocity the rocket will experience, and at what time during flight these will occur. Launched with [J985TF-16 ] 50 a  2  25 0 -25  E  -  500 250 0  10000 % 7500 LL  0  5000  "D  3  2500  —  -~~  -  E  Burndi t  Ejection l  0 0.0  2.5  5.0  i i i  l  da  i i l  7.5  10.0  12.5  15.0  17.5  Time F i g u r e 3.4: E x a m p l e R o c S i m O u t p u t  Other program outputs are the optimal delay time for ejection, and the altitude and speed of the rocket when ejection occurs for the delay submitted to the program.  This is important to  determine whether the parachute safely deployed (i.e. the rocket did not have too high a velocity at ejection) and what the optimal delay would be to get the maximum altitude.  39  3.3.2  D e t e r m i n i n g M o t o r Range The first simulations were created to determine what approximate size o f rocket motor  would be required to loft the payload to the required 6-8km altitude.  The purpose was to  calculate the general range o f impulse which would be required, and to provide a starting point for optimizing the rocket design. These simulations were performed with a very conventional, 3 fin rocket design, shown i n Figure 3.1. This rocket was not meant to reflect the actual unknown final design, but to be approximately the right size and weight.  Simulations were performed with ten different commercially available rocket motors for this same rocket design. The motors had total impulse values ranging from 120 N s ( " G " class motor) to 7200 N s ( " M class" motor). Figure 3.5 shows the resulting altitudes for the rocket motors, and the approximate total rocket masses.  Figure 3.5 shows that an approximately 2500 N s impulse rocket engine is required to loft the payload to 6km. This corresponds to a " K " class rocket engine. K engines are commercially available, and are generally manufactured with a 54mm diameter.  T o reach 8km, the rocket  would require more than 3600 N s o f impulse, corresponding to a high L , or l o w M class motor. These motors are much larger, and have 72-98mm diameters. The other major consideration in selecting an impulse class for the motor is its total mass. Recall that one o f the system requirements is to have 400 rockets on a buoy, which only has 800kg available buoyancy.  G i v e n the mass o f the other system components, each rocket,  complete with launch tube should weigh no more than 2kg each.  40  120  436 2000 2677 3613 Motor Impulse f N-s) Rocket Mass (kg) - « - Altitude (km)  Figure 3.5: Motor Impulse vs. Rocket altitude and mass Examining Figure 3.5, a rocket capable of reaching 8km would weigh approximately 4kg, while a rocket capable of reaching 6km would weight around 3kg. Assuming another 1kg per rocket for its launch tube, these values limit the number of rockets on each buoy to 160 for 8km altitude rockets and 200 for 6km rockets. These values fall short of the desired 400 rockets per buoy.  Figure 3.5 indicates that to have 400 rockets on a buoy, the rockets would only be  able to reach an altitude of 3-4km, rather than the desired 6-8km.  At this point of the simulation work, the decision was made to give priority to the altitude requirement, over the desired number of rockets on the buoy. However, given the desire to have as many rockets as possible, plus the space restrictions on the buoy, it was also decided to revise the altitude goal to 6km. Therefore, all future simulation and experimental work assumes that  41  the final rocketsonde w i l l use a K class rocket engine, and the system w i l l be designed to launch 200 rockets.  The decision to use only 200 rockets on a buoy does not meet the requirement for daily launches. T w o possibilities to resolve this problem are either to service the buoy twice a year, or to locate two buoys i n a similar location, and have them alternate daily launches. Both these propositions are very costly however. A servicing trip w i l l take approximately 3 days per buoy, including travel time.  Therefore two maintenance cruises each year w i l l increase the yearly  maintenance cost per buoy by $40K.  Locating two buoys in the same location w i l l also be  costly. Beside the approximate $300K initial cost of a second buoy, each location w i l l require an additional 2 servicing days. This increases the yearly maintenance cost per site by $26K.  The cheapest solution is to still service a single buoy once a year, but to be more selective about when to take weather soundings. The data of most interest is storm data, which occurs mainly in the Fall-Winter-Spring season. A rocket could be fired daily during this half of the year only, and the Rocketsonde-Buoy System would still provide significant improvement to the Data V o i d problem.  3.3.3  Nose C o n e a n d B o d y T u b e Design The previous set o f simulations determined that the rocket prototype would use a K class  rocket engine.  This conclusion allows a more realistic rocket design to be made for future  simulations. While the specifications vary between manufacturers, K engines are almost always 54mm in diameter, and are 45-66cm in length. Given the desire that the rocket be as streamlined  42  as possible to hit the altitude goal, the rocket engine should fit exactly i n the body tube. This sets the outer diameter o f the rocket body tube to about 56mm. S c a l e : 1/8  '  Rocket length:1214.670  m m , diameter: 56.00D  m m , s p a n diameter: 169.880  mm  Rocket m a s s 539.630 g , Selected stage m a s s 539.630 g S h o w n w/o E n g i n e s .  (Ml  Method  CG m m  CP mm  CNa  Static margin  Analysis  Barrowman  529.981  1017.493  22.429  12.83  T h e rocket is o v e r stable.  Figure 3.6: Prototype Rocket with 4-fin configuration In addition, the rocket is assumed to carry a standard Vaisala rocketsonde payload. These sondes are currently manufactured in 35mm tubes, approximately 30 cm in length. They come complete with a conical plastic nosecone which contains the transmitter.  Rather than  trying to secure this smaller cylinder in the larger 56mm tube, it is simpler to create a dualdiameter rocket, as shown i n Figure 3.6. The design consists o f a 30cm long 35mm piece of aluminum tubing going through a short fiberglass transition piece to a 70cm long piece o f 58mm aluminum tube. This rocket design allows the 54mm engine to be slipped into one end, and secured with a set o f screws. The Vaisala sonde, complete with nosecone, is then fitted into the other end.  43  Therefore, it is both easy to manufacture and assemble i n large number, and reduces labor costs to mass-produce. A l s o , the Barrowman equations imply that having a larger diameter body tube at the fin end o f the rocket helps to move the centre o f pressure aft.  Thus, the dual diameter  rocket design increases the rocket's stability. This cost o f this design compared to a singlediameter body tube is negligible.  3.3.4  Effects of Motor Burn Characteristics A t this point, the basic rocket design and the required class o f motor have been  determined. However, a motor class corresponds only to the total impulse o f the motor, not how that impulse is distributed over the engine's burn time. There are many different types o f motors under the K classification. A motor like the Aerotech K 2 5 0 is known as a "long burn" motor because it has an average thrust o f 250 N , and burns for around 9 seconds. motor, like the C T I K 1 2 0 0 , has a burn time of 1-2 seconds.  A "short burn"  Both motors w i l l loft a rocket to  essentially the same altitude, given ideal conditions, but provide very different flight behavior. A short burn motor w i l l accelerate the rocket very quickly out o f the launch tube, and then w i l l have a long coast period. A long burn motor has a lower velocity and shorter coast period, but w i l l subject the rocket to smaller accelerations.  The ability for a rocket to reach its desired altitude in high winds is a desired characteristic. The Rocketsonde Buoy System is required to be operational in storm conditions that include 20-30m/s winds. Comparative simulations were performed using the rocket design shown in Figure 3.6 under wind conditions ranging from 0-30 m/s. Simulations were first run  44  using an Aerotech K250 long-burn motor, then a CTI K1200. The CTI K1200 is a prototype motor designed specifically for this project by Cesaroni Technologies Incorporated.  120 100 +• •o  80  I n  60  |  40 20 0 10  15  20  25  30  35  Wind speed (m/s) K1200 fast bum  K250 slow bum  F i g u r e 3.7: W i n d speed vs. altitude reached by 4-fin design  The resulting changes in the maximum altitude attained are shown in Figure 3.7. Simulations for other fin designs show comparable results.  The simulations show that the rocket with a fast burning motor will reach 93% of its maximum design altitude even during 30 m/s winds. This is because the rocket accelerates so rapidly from its tube that the cross winds have little effect on its trajectory. The slow burning K250 motor, however, will only propel the rocket it 75% of its design altitude in similar wind conditions.  These results suggest that a fast burning motor is more appropriate for the  Rocketsonde-Buoy system, given the environmental conditions.  45  3.3.5 Fin Design Considerations U p to this point, all simulations were performed using the 4 finned prototype design model. However, given the limited space on the buoy, a conventional 4-fin rocket may not be the optimal design.  A s previously discussed, there are several ways to decrease the rocket  diameter without moving the centre o f pressure forward.  One method is to increase the number,  of fins. More fins mean that the fins do not have to extend as far away from the rocket body, thus decreasing the diameter. outlined i n section 3.1.4.  Another proposed method is to use a circular fin design, as  Circular fins generate twice the restoring moment o f 4 conventional  fins o f the same width and length, so a much smaller diameter is possible without changing the centre o f pressure location. Both methods, however, mean more drag on the rocket and logically should come with a loss in altitude.  A series o f simulations were run to determine just how severe the altitude loss is compared to the area gain on the buoy surface. The simulations used the dual-diameter rocket prototype with 4 different fin configurations: 3 conventional fins, 4 conventional fins, 6 conventional fins and a circular fin. In all simulations, a C T I K 1 2 0 0 was used as the rocket engine.  First, the rocket was optimally designed with 3 conventional fins to be stable.  Subsequently, the fin design was changed and optimized so that the diameter was decreased, but the centre o f pressure remained i n the same location as the 3 fin design. respectively show the 4 and 6 finned designs.  46  Figures 3.6 and 3.8  S c a l e : 1/8 R o c k e t length: 1 3 0 3 . 9 2 0  m m . d i a m e t e r : 56;000  mm , s p a n diameter: 184.880 m m  R o c k e t m a s s 6 3 4 . 5 8 1 g / S e l e c t e d s t a g e m a s s 634.581 g S h o w n w/o E n g i n e s .  Method  CGmm  C Pmm  CNa  Static margin  Analysis  Barrowman  630.551  1075.237  37.245  11.70  T h e rocket i s o v e r s t a b l e .  Figure 3 . 8 : Prototype rocket with 6-fin configuration Simulating the circular fin design was more complicated, as RocSim is intended only for conventional fin designs. The circular fin was simulated as 4 square fins with equivalent area as the circular fin to place the centre of pressure in the appropriate equation, as shown in Figure 3.9. For this simulation, the coefficient of drag was not calculated by RocSim, but set manually. Wind tunnel testing of "ringtail" rockets determines the coefficient of drag to be approximately 0.68 for speed from Mach 0.5 to Mach 3 . 16  This value was used for the circularfinmodel.  47  Scale: 1/8 Rocket length: 1210.000 mm .diameter: 56.111)0 mm , span diameter: 234.880 Rocket m a s s 1738.941 g . Selected stage m a s s 1738.941 g Engines: |K 111(111 2b ]  Figure 3.9: Prototype rocket with circular fin configuration The results o f the simulations are shown in Figure 3.10.  The simulations indicate that  there is not much o f an altitude difference between the 3 and 4 fin designs, but the 4-finned rocket w i l l take around 2 5 % less space on the buoy. It is clear that the 4 fin design is a better option than the 3 fin. However, it is more difficult to chose an "optimal" design from the 4, 6 and circular fin models. The circular fin design is by far the most space-efficient, taking only half o f the surface area as the 4-fin design, but it only reached two-thirds o f the altitude. Thus there exists a very clear trade-off between the altitude and the area.  However, even with the  largest rocket diameter required for 3-fin design, the 200 rocket battery w i l l still easily fit on the surface on the N O M A D Buoy. In the end, the space constraint o f the system becomes a nonissue. However, a larger diameter launch tubes weigh more, so there does exists another tradeoff between the system mass and the rocket altitude.  48  6 4-fin  • 3-fin  5H A 6-fin  -a 4 3  <  X Circular 3  E =- 2  E 2  0.01  0.02  0.03  0.04  0.05  0.06  L a u n c h T u b e A r e a (m2)  Figure 3.10: Effects of fin drag on rocket altitude The fin design also has an impact on the acceleration forces on the rocket, as does the burn time o f the motor. Recall that the current Vaisala sonde is only capable of experiencing accelerations up to 50 G before damage may occur to the electronics. Table 3.1 shows various simulated accelerations experienced by rockets with the different fin designs and motors. It is apparent that the rocket w i l l only stay under the 50 G limit i f a slow burn motor, the K250, is used. However, earlier simulations demonstrated that this is not an ideal motor to use, as it does not perform well under windy conditions, and requires excessively long launch tubes.  Fin Design 4 fin  6 fins  circular fin  Motor  Acceleration (G)  K250 K1000 K1200 K250 K1000 K1200 K250  14 54 70 14 54 65 15  K1000  64  K1200  77  Table 3.1: Maximum Acceleration for prototype designs  49  The K 1 0 0 0 and K1200, both C T I fast burn prototypes designed specifically for this project, exceed the 50 G limit. However, they exceed it by a maximum o f only 50%.  It may be  possible for Vaisala to strengthen their electronics slightly to handle up to 70 G o f acceleration. Alternatively, a motor may be required that has a slightly longer burn time than the K1000, but only enough to keep the rocket under the 50 G limit.  3.3.6  Launch Tube Length The R o c S i m simulations also provided a means to determine the required length o f the  launch tubes for the Rocketsonde-Buoy system.  A s mentioned previously, the purpose o f a  launch device, like a tube, is to protect the rocket's vertical course until it reaches approximately 20m/s. The required launch tube length for a given rocket can be found from the simulation data by examining the altitude o f the rocket when it reaches 20m/s. The resulting launch tube lengths for the various proposed K designs are in Table 3.2.  Motor  Time (v=20m/s)  Tube Length (above rocket)  K250 K1000 K1200  0.2s 0.035s 0.03s  2m Less than l m Less than 0.8m  Table 3.2: Required launch tube lengths for circular fin prototype Table 3.2 demonstrates another reason why it is preferable to use a short burn motor for the proposed rocket. The long burn K 2 5 0 motor requires a 2m long launch tube in addition to the 1.2m rocket length, while the K1000 and K1200 motors need less than half that additional length. L o n g launch tubes are very hard to balance on a N O M A D buoy, and would provide a larger surface area for the wind. This increases the chance o f the buoy flipping during very bad storm conditions. Larger launch tubes also mean more weight, which decreases.the number o f rockets on each buoy.  50  3.3.7  Wind Effects on Recovery A final issue which was examined using the simulations was the estimated drift of the  sonde package as it descends during various wind speed. Recall that the system is supposed to gather daily data i n the same location. Given that the wind w i l l move the sonde package, it is important to know how much drift w i l l occur from the initial location. The forecasters w i l l then have to determine whether it can be treated as one known location i n their models.  Table 3.3 shows the effect the wind w i l l have on the drift o f a sonde deployed from the dual-diameter proposed design using a K 2 5 0 motor. In 30 m/s wind conditions, the package w i l l drift about 5.3 k m from the buoy as it descends. This is relatively small compared to the drift o f radiosonde balloons; hence the sounding can be interpreted as a point observation.  Wind speed (m/s) 0 10 20 30  Drift Range (km)  Altitude (km) 7.6 7 6.3 5.8  0 3.8 5 5.3  Table 3.3: Wind effects on 4-fin rocket with K250 engine  3.4  Experimental Results There are several limitations to the simulation software.  Most importantly, R o c S i m  simulates aerodynamic characteristics o f the various designs, not the mechanical characteristics. It gives no indication o f whether the rocket materials and construction techniques w i l l withstand the flight velocities and accelerations. A l s o , the simulation does not consider that the rockets are  51  going to be tube launched. There is not much literature available regarding tube launching, so it is unclear whether the tube w i l l significantly affect the rocket's flight, or whether the back pressure within the tube w i l l affect the motor.  While the simulation work provides a good  theoretical basis on which to design the rocketsonde, experimental prototypes are still required to test the various designs i n practice.  3.4.1  Experimental Objectives The main objective o f carrying out experimental launches is to test designs and design  features that cannot be examined using simulation software.  The majority o f experimental  launches focused on testing unconventional fin designs, particularly folding fins. Conventional and circular fin designs were also tested to examine construction techniques.  It was also  unknown how tube launching may affect motor behavior and rocket flight, so this was examined concurrently during fin testing.  3.4.2  Experimental Procedure Location Due  to the difficulty in obtaining airspace waivers to launch high powered rockets,  testing locations had to be chosen opportunistically throughout western Canada and the United States.  Initially, testing was done at launches organized by various amateur rocketry  organizations i n M i s s i o n B C , Lethbridge A B and Brothers O R .  In summer 2002, the project  team was granted a weekly waiver to 20000' i n the middle o f Harrison Lake, B C . This launch site is shown in Figure 3.11.  52  F i g u r e 3.11: Rocket launch over H a r r i s o n L a k e , B C For experience and certification purposes, undocumented rockets were also launched in Vancouver B C , Boulder C O , and Belleville O N . L a u n c h i n g A p p a r a t u s Rocket launching requires a guidance device, two leads, a long length o f cable allowing a safe distance between the launch officer and the rocket, and a power supply to provide a large current. Initial testing was done at land-based rocketry meets, where all equipment except the guidance device was provided by the organizers.  The earliest tests were performed using a  conventional V" or 14" rail, as shown in Figure 3.12(a). Progression was soon made to a P V C launch tube as the guidance device. A s shown in Figure 3.12(b), this tube had an elbow piece at its base to allow the rocket exhaust to escape from the bottom, rather than filling the tube.  53  Figure 3.12: Launch Apparatus (a) rail launch  (b) land-based tube launch  (c) water tube launch  After the Harrison Lake site became available, a floating platform was constructed to support a P V C tube during water launches. (Figure 3.12(c))  The platform was deployed using  100 feet o f extension cord, supported by foam floats every 0.5m, to keep it a safe distance away from the boat with the project team. A large portable battery was used for ignition. Rocket Prototype Construction A l l prototype rockets were constructed at U B C by M a r k Stull, an experienced mechanic who joined the project group in the summer of 2002.  Generally, the rockets used one o f two  standard body tube designs: (a)  small testing rocket: 38mm diameter tube, 75 cm long  (b)  expected final prototype design: dual diameter with 38mm transitioning to 54mm (approximately 1.2m total length)  Construction materials, fins and nose cones varied on these designs.  Construction materials  included cardboard and wood, for small, lightweight, prototypes, and fiberglass and aluminum,  54  for more realistic prototype rockets. Fins were attached using J B W e l d , high-temperature epoxy fillets, and fiberglass reinforcing on K-class rockets. Launch History A timeline o f the first prototype tests for various motor classes, and the altitude reached, is shown in Figure 3.13. Due to flight safety requirements, the majority o f testing was done during the summer months, when there were fewer cloudy days.  6 K  5  •o  <  4 J  3  ( p r o t o t y p e )  <>>.  2  H  g  a a <  J  I G  1 0  A - D  28-Feb-01  16-Sep-01 . 4-Apr-02  21-Oct-02  9-May-03  Figure 3.13: Prototype Development  Most prototypes, particularly those with unconventional fin designs, were tested first using lower-range motors, then later with high-powered motors. A full table o f all rocket prototypes launched, including launch dates and locations, may be found i n Table 3.4  55  T a b l e 3.4: Complete Rocket L a u n c h R e c o r d  56  3.4.3  Fin Designs Results Conventional Fins Experimental Launches: Early in the project a series o f experimental launches were made to gain experience i n conventional rocketry.  These included a series o f launches to gain three  certification levels that allow the purchase and use o f high powered rocket engines in Canada. These launches were not formally documented.  Conventional fin rockets continued to be used  during the experimental process to check launch apparatus, wind direction, and to practice recovery chase methods.  It was concluded early that the simulation work done i n the previous  chapter provides reliable information about conventional designs.  Three significant experimental launches using conventional fin designs were done in summer 2002.  The main goal behind these launches was to test construction and launch  techniques for very high powered rockets. First, a 38mm rocket with five conventional fins was tube launched using a C T I J300 motor at the Lethbridge A B meet in June 2002. 2002, two identical,  final-prototype  In October  dual-diameter design with 4 conventional fins, were  launched from 6" P V C tubes using prototype C T I K 1 2 0 0 engines at the Harrison Lake site. These rockets were equipped with activated Vaisala sondes, and a receiver was set up in the boat.  Results and Discussion: A l l three rockets launched successfully. The J300 rocket hit an altitude of approximately 2.7km and was recovered with very minimal damage.  T w o fins detached at  impact, but the flight was straight and unaffected by the tube launch. The rocket is simulated to have reached over 5 5 - G and M a c h 1.5, so it was a good test o f the construction techniques. The  57  first K 1 2 0 0 rocket exited the launch tube at a slight angle and only reached 3.6km, while the second K1200 rocket passed 4.8km. Unfortunately, due to the launch site being over water, the K-class rockets were not recovered.  However, it appeared no components were lost during  flight. The straight and stable flights indicated that the fins remained intact. concluded that the construction techniques used are appropriate  Overall, it can be  for conventional  rocket  prototypes for this project. Spring-Released F o l d i n g Fins Design Concept: The goal behind the spring-released folding fin design is to allow large, conventional fins to fit into a small launch tube. The design incorporated four fins, each on a hinge held in an open position by a circular spring (see Figure 3.14). The fins are folded inward when the rocket is placed in the launch tube. They then spring open during launch when the rocket clears the tube.  F i g u r e 3.14: S p r i n g - H i n g e d F i n s E x p e r i m e n t a l L a u n c h e s : There were several potential concerns with this design. First, there was a concern that the fins would catch within the launch tube and affect the rocket's flight. It  58  was also unknown how the fins deploying would affect the rocket flight.  Six experimental  launches were made with this fin design using 2 different rockets. A small, 38mm, cardboard and balsa rocket was fired with Estes D and E engines off a launch rail, and then layer from a 3" launch tube with a C T I G60 engine at Hatzic Prairie, B C .  A final design, dual-diameter rocket  with this design was then constructed from aluminum and tested i n Brothers, O R . This rocket was tube-launched first using an Aerotech G80 engine, then twice more with an Aerotech H I 2 3 engine.  Results a n d Discussion: The first two launches, using a lightweight cardboard model rocket from a %" launch rail flew really nice and straight, with no problems. Unfortunately there was an engine failure, due to manufacturing problems, when this same rocket was fired from a launch tube. The prototype was destroyed before it left the tube. However, it appeared that the design worked at the low-powered range.  The first tube launch o f the final-design prototype rocket with spring-hinged fins also worked very well. The launch tube did not interfere with the rocket's fin deployment, and the rocket flew very straight. Given that it behaved well using a G80 motor, it was expected to yield equal results using the H123. However, the following two launches were not as successful. During both successive launches, the rocket left the tube straight, without the fins catching in any ways. However, within 100 ft of launch, one o f the fins would fold back in due to the air currents around the rocket. This would cause the rocket to cartwheel i n an unstable manner, and never recover. It was clear that this design would not work for high-powered rockets.  59 Spring-Released Folding with Locking Mechanism Design Concept: The spring-released folding design deployed well from the tube, and is capable o f saving space on the buoy. However, its major flaw is that the fins are able to re-fold during flight.  A proposed solution was to have the fins not only spring out, but also use the thrust  forces to slide the fins down the hinge 1cm after deployment.  Stoppers are placed at the rocket  ends so that once deployed, and slide into place, the fins would not be able to refold during flight.  Experimental Launch: A small-scale, aluminum rocket with this revised design was tube launched using a C T I G 6 0 motor i n Lethbridge. It was requested that no one touch the rocket besides the project team after launch so it would be easy to check whether all four fins slide into the locked position.  Results and Discussion: The launch was very successful, flying straight out o f the tube and deploying well.  U p o n recovering the rocket, three o f the fins were securely locked, and the  fourth looked like it had only come unlocked at impact. Further testing with higher powered motors needed to be done to be sure, but it appeared that the new locking device solved the previous problems with spring-hinged fins. However, it was subsequently decided that the fin mechanism was too complicated to mass-produce cheaply and accurately, and was therefore inappropriate for the final project design.  This ended all testing o f the spring-hinged and spring-  hinged locked designs.  60 Tangent F o l d - D o w n Fins Design Concept:  Four conventionally shaped fins are mounted tangentially, rather than  perpendicularly, to the rocket body tube. They are attached at a pivot point, as shown in Figure 3.15, so that they swing up to fit in a smaller launch tube. U p o n leaving the tube, aerodynamic forces push the fins down and hold them there during flight. There are no locking devices.  F i g u r e 3.15: Tangent F o l d - D o w n F i n s  E x p e r i m e n t a l L a u n c h e s : Again, there was a concern regarding this fin design catching within the launch tube, as well as whether the fins would move during flight. Three prototypes were built using this fin design. The design was first tested using a small, cardboard and wood, rocket from a launch rail, using Estes D and E engines. (Two identical rockets were built.) This same rocket design was then launch from a 3" launch tube, once on an Estes D engine, and twice using Aerotech G40 motors.  Later, this rocket was tube launched three more times at subsequent  meets, using an Aerotech G 8 0 , and two C T I G60 motors.  61  Second, an identical aluminum version o f this small rocket was tube launched using an Aerotech H I 12 engine.  The same rocket, after some reinforcements to the fins, was tube  launched during a C T I G60. Last, a final design prototype was tube launched with a C T I H I 10 engine.  Results a n d Discussion: Early tests showed great promise. The original cardboard prototypes performed well on low and mid-powered engines, and there were no deployment problems from the tube. The fins appeared to stay in position after deployment, and seven out of eight o f the launches using this first design were completely straight and successful. The only exception was when one o f the fins caught on a launch lug during deployment, causing only three to deploy successfully. This was easily corrected.  The two launches with the small-scale aluminum design were less successful. The first flight, using an H I 12, the rocket went cleanly out o f the tube, but the fins all ripped off at the pivot point. A n attempt was made to correct this by increasing the fin thickness, and putting larger washers around the pivot screws so a larger area would have to tear. The second launch was more successful, with the fins staying attached, although slightly warped. Also, one o f the pivot points was damaged.  The launch o f the final design prototype on an H I 10 also failed. T w o fins, on opposite sides o f the rocket, completely sheared off. A t this point it was decided that it would not be possible to adapt the tangent folding-fin design to a K-class rocket, given the difficulty at the H motor level.  62 C i r c u l a r F i n s Design Concept: The idea behind circular fins was outline in Section 3.1.4. Circular fins greatly increase the restoring moment on the rocket in flight, without increasing the rocket's outer diameter.  Several rocket prototypes were built using circular fins, combined with partial  conventional fins which extend from the ring section. A n example is shown in Figure 3.16. In all cases, the circular section was made from fiberglass, and the extended conventional fins that hold it are made from either aluminum or fiberglass,  Figure 3.16: C i r c u l a r F i n Design  E x p e r i m e n t a l L a u n c h e s : The use o f circular fins was introduced late in the experimental process, after it was determined that all folding-fin designs were unfeasible. Four experimental launches were made using three different rockets. First, a small, single diameter, rocket with a 4" diameter was tube launched using a C T I G60 engine.  Second, a dual-diameter, "final prototype", rocket with a 4" diameter circular fin was tube launched with first a C T I G60 engine. Given the success o f the previous launch, the same  63  prototype rocket was then tube-launched using a C T I H I 10 engine. A l l first three launches occurred at the RocLake meet in Lethbridge, A B .  Finally, a 38mm rocket, designed to fit inside a 3" diameter tube, was tube launched at the Harrison Lake site.  Results a n d Discussion: The first launch at Lethbridge failed because the motor delay time was accidentally set too long. damaged at impact.  The rocket separated from its parachute and was catastrophically  Prior to apogee, however, the rocket flew very straight and stable.  other two launches at Lethbridge were equally straight, and had safe recoveries.  The  The only  noticeable problem was that the high-temperature epoxy fillet would crack at the point where the conventional fins touched the circular ring. supports and dislodged about 0.5cm.  In one case the circular fin sheared off all four  It was determined that these problems were due to a  combination o f the conventional fins not being aligned carefully enough to give zero lift, and failure o f the epoxy.  The last circular fin launch, fired at Harrison Lake with a 3" diameter was severely unstable.  It was determined that the simulations were correct when they indicated that the  prototype rocket needed a minimum 4" circular fin.  The experiments demonstrated that the 4 " circular fin design is quite feasible for the final rocket prototype.  It is recommended, however, that the circular fin component be cast as a  single piece, rather than constructed i n 5 pieces and attached at the joints. This should solve  64  both the problem with finding an appropriate epoxy, and ensure the straightness o f the conventional fin components. However, even with a circular fin, the final design rocket cannot fit in a 3" launch tube. It is assumed that launch tubes w i l l be made from commonly available tubing, available in one inch increments.  Therefore, the circular fin w i l l fit in a minimum 4"  launch tube, since a 3" design is not stable. However, the 6 conventional fin design simulated in section 3.3 w i l l also fit in the 4" tube. This conventional design has less drag, and w i l l reach a higher altitude, so is a better option for the final design i f restricted to circular tubes in 1" increments. However, it is likely that the rocket tubes could be custom-made in any shape. In this case, it makes sense to pick the design (4-fin, 6-fin or circular) which reaches the highest altitude while still having a total mass, including launch tube, under 4kg.  3.4.4  T u b e L a u n c h Results Over the course o f six months of testing, 20 launches were made from 3", 4" and 6"  circular tubes. N o results ever showed that the tube interfered with the rocket fins, or that the backpressure affected the motor performance. However, a problem was encountered during tube launch, namely that the rockets tend to lean at a slight angle within the larger launch tube. This causes the rockets to leave the tube at a slight angle, which does not always self-correct. This problem can be solved in one o f several ways. Guide wires can be attached to the upper part o f the rocket body, to keep it centred in the tube. However, the wires add a lot o f drag once the rocket has left the tube. Instead, the wires can be attached to the inner surface o f the tube, rather than to the rocket. These would not interfere with the rocket flight.  65  3.5  Conclusions and Final Rocketsonde Design Prior to any practical work, several initial decisions were made regarding the rocket  design.  The rocket is to be constructed primarily from aluminum, because it is light, strong,  inexpensive and degradable in salt water. The rocket engine w i l l be solid propellant, and include a timed delay charge for sonde and parachute deployment at apogee.  A l s o , it was determined  that launching from a tube is the most feasible way to seal the rocket from the environment until launch. These initial decisions provided a good starting point to continue designing using simulation software.  The R o c S i m simulations provided a great deal o f important information for the rocketsonde design.  They indicated that the final design should use a short burning, K class  motor and a dual-diameter body tube design. Trade-offs must also be made between the desired altitude and the number o f rockets on the buoy, rocket mass and rocket diameter.  Rocketsondes  can be designed to reach a 6km altitude, but only 200 can be stored on a N O M A D buoy. In this scenario, it would be very expensive to still have daily launches. However, it may be sufficient to only have daily launches during the Fall-Winter-Spring storm season.  I f 400 rockets are  desired, the rocketsondes w i l l only be able to reach 3-4km.  Results were given for the maximum acceleration experienced by several designs and whether they are under the 50 G sonde limit. These results show that the currently available short-burn motors cause the rocket to exceed the G limit.  However, a medium to fast burn  motor, for example a K 8 0 0 or K 9 0 0 , should still have the desirable fast burn, while keeping the  66  rocket acceleration lower. It is expected that the rocketsonde's motor w i l l be custom made for the project, so requesting a motor with these burn characteristics is possible.  A great deal was learnt regarding fin design over the course o f the experimental launches. Tests with conventional fins show that the aluminum construction methods were sound, and that a combination o f high temperature epoxy and fiberglass securely fastened then fins to the body. However, methods were sought to fit the rockets in smaller launch tube than possible with conventional fin designs. T w o folding fin designs were tested: spring-hinged and folding fins. It was determined that neither could be redesigned to successfully work with a K-class rocket. Locked spring-hinged fins proved more successful, but they were too complicated to mass produce. It was concluded that the rocketsonde prototype must have a fixed-fin design.  Experimental results show that circular fins are a viable alternative to conventional fin designs. Circular fins, however, should be constructed from one single piece. Also, the circular fin needs to have a 4" diameter circle to be stable, assuming a dual-diameter, 1.2m rocket body. However, a conventional 6-fin design w i l l also fit in a 4" tube, and reach a higher altitude on the same rocket and motor. If the launch tubes are custom designed such that they are not limited to one inch increments, the fin design should be selected which reaches the highest altitude while staying under a 4 k g for total rocket and launch tube mass.  The fins w i l l likely be made as a  single, molded-plastic canister, which can be attached on the aluminum body tube.  Cost is  therefore not a fact, as the difference between casting a circular or conventional fin canister is negligible.  67  Experimental launches also show that tube-launching is a promising solution to the environmental issues facing the system. The tube does not impair the rocket's flight or affect the motor. Problems with the rocket slanting in the tube can be solved by including guide wires. Therefore, the final rocketsonde design w i l l be tube launched.  68  4.  Launch Control  The main purpose o f the launch control system is to achieve a near-vertical rocket launch. This feature maximizes the possible altitude o f the rocket, reduces lateral drift relative to the buoy location, and increases the system's safety.  A potential solution to the problem o f  launch control is to put the entire launch system on a gimbaled mechanism that would keep the launch tube frame vertical relative to the reference frame. The requirements developed in Chapter 2 discuss how the use o f moving parts on a buoy is undesirable because they would likely be unreliable when exposed to the harsh open-ocean conditions. Such mechanical systems would wear out or corrode long before a scheduled yearly maintenance.  Thus, the chosen strategy is to monitor the orientation o f the buoy, and to initiate rocket launch at an instant when the buoy happens to be "close" to vertical. Ideally, the launch should be exactly vertical. However, depending on the sea and wind conditions, the buoy may not come exactly vertical within an acceptable period.  Thus, the chosen strategy must also include a  criterion for determining the practical "closeness" to vertical for rocket launching. This chapter describes the chosen launch decision strategy, and the results o f a practical demonstration using a commercially available sensor connected to computerized data acquisition and launch control system.  69  4.1  Background and Theory  4.1.1  Reference Frames Two reference frames are needed to describe the orientation o f a buoy floating on the  ocean surface.  First, there is the inertial reference frame attached to the earth's surface at the  buoy location, F , as shown i n Figure 4.1. For the system, "vertical" is defined as parallel to the e  z-axis o f the earth's reference frame.  Second, there is moving reference frame attached to the  buoy surface such that the z-axis is parallel to the buoy launch tubes, F|. A vertical launch opportunity occurs when the angle between z and z\ is close to zero. e  ^ Fe: earth  > x reference frame  Figure 4.1: Important Reference Frames Translational displacements between the two frames are irrelevant. the reference frame axial directions are then related by equation (4.1)  X  L  yL  = M  Where M is a 3 x 3 rotation matrix.  70  y  E  The relative orientations o f  4.1.2  System Interfaces Launch control is central to the operation o f a Rocketsonde-Buoy System as it links all  the important subsystems and the outside environment. Prior to sensor selection and algorithm design, the R B S was examined to determine all possible inputs.and outputs the launch control algorithm is required to take into account. Figure 4.2 schematically shows these relationships. The interfaces can be placed i n 3 categories: data collection, launch implementation, and communications.  The launch control system takes into account several data sources before deciding when to initiate a launch. M o s t important are buoy tilting and vertical motion. These w i l l be dealt with in more detail in the next section. It also may be desirable to receive the meteorological data that is being collected by the buoy, such as wind speed and temperature. This would allow launches to be quickly cancelled i f the weather is more severe than worst-case design scenarios. It is also desirable to know whether the buoy deck is awash, which might indicate a flipped or partially submerged buoy.  The second set o f interfaces concern the actual process o f launching a rocket. First, there are outputs to turn on the sonde receiver and a warning lamp that launch is imminent. Once a launch opportunity is detected, the program outputs a signal to the ignition circuit, and the ignition system returns a signal i f the launch was successful. Before shutting down to wait for the next scheduled launch time, the program puts the ignition system into a safe mode, and shuts down the sonde receiver and warning lamp.  71  ARGOS  I  satellite  Circuitry  Transmitter s a f e t y on/off signal  from  Receiver:  shjp CLOCK  Ignit o r / L a u n c h  safety off: or  LAUNCH  7  successful, aunch  LAUNCH CONTROL PROGRAM  indicator  REQUIRED SENSORS ( b u o y tilt, wave data)  METEORLOGICAL SENSORS :(T, -P, a n e m o m e t e r ) D e c k Awash; Indicator  Figure 4.2: Complete Launch Control Interfaces  The third set o f interfaces involves communication with the buoy. There is a receiver so a boat approaching for maintenance can remotely set the system into a "safety o f f mode from a distance.  There is also a transmitter to send for the A R G O S satellite relay.  This is not to  transmit the sonde data, which does not concern the launch control program, but to report failed systems o f problems which require maintenance, such as a fired rocket stuck in a tube.  The above system has been simplified to facilitate designing and testing the important program features in an indoor laboratory setting. The simplified interface scheme for the launch control system used i n this research is shown in Figure 4.3.  72  The only data being considered is from the required sensors to determine a vertical launch, as this is the main challenge to the algorithm design. Meteorological and deck awash inputs are easy to add later i n commands such as " i f the wind is more than 30 m/s, turn o f f or " i f the deck awash signal is triggered, do not fire". The launch circuitry is simplified to an L E D to signal a rocket launch, or in latter stages, a single rocket firing circuit. Communication interfaces have been replaced by a P C interface, where a visual interface allows the program to be put into a safe mode and also records program variables.  LED Ignitor/Launch  Inputs from t h e C o m p u t e r  o r single  rocket  circuit  Circuitry safety off or  safety  on/off  signal  from  LAUNCH  REQUIRED SENSORS  Receiver  LAUNCH CONTROL PROGRAM  ship CLOCK  ( b u o y tilt,  wave  data)  on/off  | Warning  Lamp j  1  Beeping s i g n a l from computer  Figure 4.3: Interfaces Used for Proof-of-Concept Design  4.2  Sensor Selection One o f the principal components o f the proposed launch control system, as detailed in the  previous section, is one or more sensors that monitor the buoy motion and provide buoy tilt and orientation data to the launch control program. Acquiring and testing an appropriate sensor package therefore becomes an important element of the system design.  73  4.2.1  Sensor Requirements The principal objective o f the buoy sensor(s) is to provide enough o f the orientation  matrix M (equation 4-1) to determine the angle o f the launch tubes relative to vertical at any instant.  Other requirements developed i n Chapter 2 are that the system must be able to launch i n waves that tilt the buoy a maximum amplitude of 12 degrees. Therefore, the sensor should be able to determine at least ± 1 5 degrees roll and pitch. The maximum wave frequency is 0.2 H z , so it is not a significant concern to find a sensor with a sufficiently large sampling frequency. The sensor w i l l have to provide accurate readings even when subjected to high angular velocities and accelerations, so the settling time must be small. In addition, the sensor must be operational at 5°C and above.  Another concern for the sensor is that it must have a reference to gravity. It cannot just monitor its orientation with respect to its previous orientation, but must relate back to the earth's reference frame.  This reference must also be regularly corrected for drift, as most sensors use  accelerometers to find a gravity reference. With a typical accelerometer, i f the buoy lists off axis for an extended period, the accelerometer zero mark would drift so that there is an error in its gravity direction.  Although the rotation matrix M is the minimal data that the sensors must provide, it is desirable to have additional information to create a more sophisticated launch program. Vertical acceleration or velocity and angular velocities are possible useful quantities.  74  4.2.2  O r i e n t a t i o n Sensor Technology Orientation or tilt sensor technology is currently utilized on both the Environment  Canada Pacific B u o y network, and N O A A ' s Buoy networks.  N O A A ' s Pacific Marine  Environmental Laboratory ( P M E L ) uses electrolytic bi-axial inclinometers, such as those manufactured by A p p l i e d Geomechanics . These sensors consist o f 5 prongs in a small bubble 23  of electrolytic fluid. A s the fluid level changes with tilt, the capacitance o f the different pairs o f prongs changes, which indicate the pitch and roll angles o f the buoys. These sensors are a very inexpensive (approximately $200 U S D ) and simple method o f finding the buoy tilt. However, electrolytic sensors have a settling time o f approximately 0.45 - 0.9 seconds , which is much 24  too large for a launch control system application.  In addition, this sensor provides no  information other than tilt, so additional sensors would have to be included to find angular velocity and vertical accelerations. where their only purpose  However, the sensor is sufficient for the N O A A buoys,  is to provide general approximations o f wave conditions for  troubleshooting purposes.  While N O A A ' s buoys use a simple method to examine buoy tilt, Environment Canada's 25  buoys utilize a very complex package, developed by A x y s Technologies . A x y s ' T R I A X Y S Wave Sensor package uses 3 angular rate sensors and 3 orthogonal accelerometers.  These data  are then fed into an onboard computer that uses custom software to determine spectral statistics including wave height, direction, period and power spectra . This package is very sophisticated, 26  but is also not appropriate for the launch control system as it does not include an absolute vertical reference.  In order to determine wave data, the sensor package only needs to examine  75  the relative motion o f the buoy, not the absolute, so it cannot provide information such as the roll and tilt o f the buoy with reference to the earth.  The T R I A X Y S wave sensor is also very  expensive (approximately $25K C A N ) because it provides such sophisticated wave data, none 1 1  of which is required by the launch control system.  4.2.3  MicroStrain 3DM-G Gyro-Enhanced Orientation Sensor The  sensor chosen here for the launch control system is a MicroStrain  Orientation sensor.  It is a dynamic sensor package, shown in Figure 4.4.  3DM-G  It consists o f 3  orthogonal accelerometers, 3 orthogonal magnetometers and 3 angular rate sensors. It is capable of sensing orientation with a 360° full scale in all three directions, as well as providing angular rate data and linear acceleration along its three axes.  Figure 4.4: MicroStrain 3DM-G Sensor The sets o f accelerometer and magnetometers are each used to determine one axis o f the sensor reference frame, relative to the earth's reference frame. gravity direction, thus positioning the z-axis of the earth frame.  The accelerometers find the The three magnetometers  identify magnetic north, thereby locating the x and y axes around the z-axis o f the earth frame. The angular rate sensors are used to eliminate accelerations read by the accelerometer that are  76  actually due to the sensor motion. A n onboard processor calculates all this information. Thus, the sensor outputs the desired rotation matrix M , as described in equation 4-1.  This sensor meets all the requirements outlined i n system 4.2.1. It can find the rotation matrix M relative to the earth because it uses absolute references.  The sensor can compensate  for dynamic motion up to 50 H z , so there are no problems with settling time and is accurate to +/- 0.5 degrees, which should be sufficient for the system  .  A l s o , according to the  manufacturers, the internal algorithm corrects for any drift i n the accelerometer readings due to the sensor being on an angle for an extended time.  Another attractive feature about this package is that all additional information it provides is hypothetically useful in the launch system algorithm. R a w data can be taken from the angular rate sensors and accelerometers to provide the desired velocities and accelerations. However, the package includes no added features, such wave spectral analysis, which would greatly increase its cost without increasing its usefulness for this project. A s a result it costs only $3K, compared to packages like those used by Environment Canada which range from $12-25K. It is also very small, weighs only 40g, and comes in a safety case that could be easily adapted or copied to survive i n a marine environment.  The MicroStrain 3 D M - G appeared to be the ideal sensor for the launch control system. However, the documentation available for the sensor provides only very general descriptions of how its internal algorithm works. Specifically, no clear answer was available to how it corrected for long-term drift, or how the magnetometers would be affected by coming into contact with  77  metal objects. Therefore, prior to using the device in the actual launch control system, a set o f tests was done to examine its response to various conditions.  4.2.4  Test Apparatus In order to examine the behavior o f the 3 D M - G sensor, Figure 4.5 shows a test assembly  consisting o f a 4-bar linkage driven by a stepper motor. The sensor is mounted on the centre bar and its readings are polled by a computer. The computer also drives the stepper motor to rotate the first linkage, causing both angular velocity and vertical accelerations to the system.  Figure 4.5: Testing Apparatus Setup A n encoder is attached to the motor shaft and its signal inputted i n the computer so that 0 i , the actual angle o f the first linkage with respect to vertical, is known.  The linkage is designed so that two o f the bars have identical length L . This simplifies the linkage geometry, shown in Figure 4.6.  In this figure, point C is attached to the stepper  78  motor, so 0i is known. F r o m the figure, we see that the angle o f the linkage supporting the 3 D M G relative to the horizontal, 02, is:  6 =90-a-p  (4-2)  2  The easier angle to find is a, because it is part o f a right-angled triangle with the midpoint o f O B . A s such  tana  ^ a -1 sin 6 ^ X  yb + lcos&  1  (4-3)  j  B (a - Isin 0! , b+lcos 0 ) 2  C (a,b)  O (0,0)  F i g u r e 4.6: L i n k a g e Geometry Examining the geometry further, P can be determined using the trigonometric cosine formula such that (a-/sin6 ) X  2  +(b + lcos0 )  2  X  = L + L -2L 2  2  cos(l80-2/5)  2  (4-4)  A n d hence ( a - / s i n 6 9 , ) +(/3 + / c o s # ) - 2 L 2  R  1  2  1  p = — a cos 2  2L  2  79  2  (4-5)  The 3 D M - G sensor is also reading 0 , so the actual, computed angle of the middle bar 2  can be compared to the angle determined by the 3 D M - G sensor.  The computer uses a  F O R T R A N code to process both the sensor and the encoder's signal, and to plot the absolute difference between the two.  4.4.5  Results Three sets o f tests were performed to verify the functionality o f the 3 D M - G .  basic test case was performed to check the accuracy o f the  sensor  First, a  compared to  the  manufacturer's specifications, and also to provide a base case for other tests. Second, the sensor was left to sit at a 30° angle for over 24 hours, and then the apparatus was run. This was to ensure that, as stated by the manufacturer, the sensor would not pick up a constant error i f the buoy were to list for a day or more.  Finally, a few short tests were performed to check the  sensor's behavior in the presence o f pieces o f metal.  Basic Test Case For this case, the sensor was kept level for over 24 hours before the test. The linkage bar attached to the motor was then placed as close to vertical as the motor steps allowed, using a water level. The system was run for 4 complete revolutions.  Figure 4.7 shows the computed inclination o f the middle bar, and the received sensor data. They start nearly identical, indicating that the method o f leveling the first bar is effective.  80  The two plots follow very closely, with the largest difference being at the maximum amplitude. A t these points, the sensor tends occasionally to either overshoot or undershoot the actual angle.  of Oi  c  "o  0.  500.  :  1000,  Rotation Angle, deg  Figure 4.7: Computed and Sensed Angles for Basic Case Figure 4 . 8 shows more closely the absolute difference between the two signals. The error ranges up to just over 2 degrees, with the peaks at the maximum amplitude.  o»  % <D O C  S:  Q QJ  O) C O  500.  1000.  Rotation Angle, deg  Figure 4.8: Sensor Error for Basic Case The figure shows the high level o f vibrations to which the sensor is subjected by the stepper motor. It appears that the problems at the peaks are due to the motor vibration, as they coincide  81  with times where the first linkage is vertical and therefore subjects the system to the highest angular accelerations.  The buoy w i l l not be subject to such accelerations, so this is not a  concern. W i t h the exception o f the peaks and troughs, which have been discussed, the majority of the error falls between 0.5 and 1 degree. While the vibrations are likely the main cause for the error exceeding 0.5 degrees specified by the manufacturer, it is prudent to assume the upper bound, +/-1 degree, as the sensor accuracy for the future algorithm design.  DC Reference Drift The second test was run i n the same manner as the first, except that the sensor was left to sit on a 30 degree angle for 24 hours prior to the test. This was to test whether the accelerometer readings were affected. Again, the test apparatus was run for 4 rotations.  15h  O  C  o or 500,  1000,  Rotation Angle, deg  Figure 4.9: Sensor Error for 30 deg The difference between the angles for this case is shown i n Figure 4.9. The figure shows a maximum error o f 1.6 degrees, which is actually better than the normal case. Again, the most significant error occurs at the maximum amplitudes and the average error is within +/- 1 degree.  82  From this it can be assumed that the sensor w i l l function properly and provide equally accurate data i f the buoy were to list for an extended period.  Magnetometer Response A final feature o f interest that was examined briefly i n the testing process was the sensor's  response  when near pieces o f metal, such as steel, which would affect  the  magnetometers. This test was done using a computer interface supplied by the manufacturer that provided roll, pitch and yaw angles. The sensor was placed in a given orientation and the angles noted. A piece o f steel was then passed near to it, and the angles examined for any changes.  It was observed that the metal has no effect whatsoever on the pitch and roll angle readings, proving, as suspected, that the magnetometers are not used i n these calculations. The - yaw reading, however, did move as the metal interfered with the magnetic north signal. When the metal was removed, the yaw signal quickly corrected itself. This is not a significant problem because only pitch and roll data is required for the launch control system. In addition, the buoy w i l l be made completely from aluminum, so the magnetometers should not have this problem. Regardless, it is beneficial to know that i f any other metal objects are placed within the buoy near the 3 D M - G , they w i l l not interfere with the launch control system.  4.2.6  Sensor Conclusions The proposed launch control system requires a sensor that w i l l provide the rotation  matrix between the earth and launch tube reference frames. Therefore, it must be dynamic and have an absolute reference to the earth. It must also have a small settling time, and ideally also  83  provide angular rate and linear acceleration information. The MicroStrain 3 D M - G sensor fits all these requirements and is a cost-effective solution compared to other options.  Additional sensor tests show that the 3 D M - G accuracy may be slightly less than specified, but not so seriously that this is a problem for the system.  Tests also show that the  sensor is not subject to drift due to buoy angles, and the pitch and roll are not affected by metal interfering with the magnetometers. In conclusion, the MicroStrain 3 D M - G is a suitable choice for the required buoy motion sensor for the launch control system.  4.3  Algorithm design The purpose o f the launch control algorithm is to determine the optimal moment at which  to launch a rocket from a rocking buoy. Ideally, the buoy should be vertical when the rocket is launched. However, a closely vertical launch may not be possible. Depending on the weather conditions, the buoy may not reach close to a vertical position within an allowable period. Under such conditions, it may be preferable to launch at a small angle rather than not launch at all. The launch control algorithm needs to be sufficiently flexible to make appropriate launch decisions over a wide range o f sea conditions. A s per the requirements developed in Chapter 2, the worst conditions for which the system must operate are a tilt amplitude o f 15 degrees, with a 5 second period.  There are three main sections o f the launch control algorithm. They are: determining the launch angle, determining the position o f the buoy in a wave, and creating a launch opportunity  84  history to optimize the launch angle.  Practical implementation o f the algorithm w i l l also be  discussed.  4.3.1  Determining Launch Angle Alpha Equation (4-6) can be inverted to provide the earth coordinates corresponding to the  launch tube reference frame. The result is: M  E  X  y  E  E_  Z  = M  2 1  M  3 1  22  M  3 2  M  3 3  1 2  M  1 3  M  _M  23  u  x  (4-6)  yL .  .  where the inverse o f the M matrix equals its transpose. The angle between the "vertical" axes in the earth and launch tube reference frames, a, can be found from the cross product o f the two corresponding unit vectors.  L  31  sin a = M  32  M  33  x[0  0  l] = M  3 2  z-M  3 l  j  (4-7)  from which: a = a sin(-^M  4.3.2  2 32  +M  2 3 1  )  (4-8)  Determining Wave Position Although it may be sufficient to launch at any opportunity where the launch tube is near  vertical, it is necessary also to consider how the position o f the buoy in a wave may affect the rocket flight. The Rocketsonde-Buoy System w i l l be subjected to very high winds during storm conditions, and these winds w i l l affect how straight the rocketsondes w i l l travel. speed in a wave trough is much lower than that at a wave peak.  85  The wind  Therefore, it may help the  system operation, when i n large wave conditions, to launch only at opportunities where the buoy is also in a wave trough, as well as the launch tube being near vertical.  A wave trough can be identified by examining the buoy's vertical acceleration.  This  vertical acceleration is zero when the buoy is going into either a wave peak or trough. When the acceleration slope is positive prior to passing zero, that indicates the buoy is traveling downwards into a wave trough. Therefore, i f the vertical acceleration is known from the 3 D M - G readings, it is simple to identify when the buoy is going into a trough.  The 3 D M - G sensor allows the raw accelerometer data to be accessed. The simplest way to find vertical acceleration then is to take the z-axis accelerometer reading from the sensor. This reading is then corrected to remove the acceleration due to gravity, thereby finding relative vertical acceleration. The relative acceleration is simply relative  S  = z 5  .accelerometer ~  (g™Vity  _ Calibration)  COS CC  (4-9)  where the gravity calibration constant is ideally 9.81. (The 3 D M - G sensor reported 9.6). This constant can be found, and recalibrated, by placing the device motionless, set on an angle, and then adjusting the calibration value while monitoring the accelerometer output until the relative acceleration is zero.  Effect o f Sensor L o c a t i o n A n additional consideration in determining the position o f the buoy i n the waves is that i f the 3 D M - G sensor is.not located at the buoy's centre o f rotation, the accelerometer readings w i l l be moderately contaminated by the rocking motion. This may result in the program waiting for a  86  wave trough when the waves are very small. Since the contaminated readings are out o f phase o f the rocking motion, an opportunity w i l l never occur when both angle and velocity sensor readings meet all conditions, and the program w i l l miss possible launch windows.  A n y solution o f this problem requires that the centre o f gravity o f the buoy be known, which w i l l be at a location within the hull the buoy.  Since this position w i l l have to be  determined, and it is a place protected from the elements, then the simplest and most logical solution is to make sure the sensor is located there.  However, i f this is not possible, the  algorithm can be adapted to correct this error.  Figure 4.10 shows the original reference frame problem, but now the sensor does not have the same reference frame as the launch tube.  Fg: centre of  Fe: earth reference frame  F i g u r e 4.10: Reference Frames w i t h Sensor  87  If H G is the 9x9 matrix representing F g in terms Fe, H s is the 9x9 matrix representing Fs in E  E  terms o f Fe, and H G is the 9x9 matrix representing Fs i n terms o f Fg, then S  H =H H EG  ES  (4-10)  SG  Expanding the matrices provides the individual rotation matrices and translation vectors: [ EG \X3 R  l EG 1x3  [ ES 1x3  i  . [oL  d  L [oL  i ES ]lx3  R  d  [ SG\X3 R  i Sc] d  . [oL  1  1x3  i  (4-11)  The variable o f most interest is the translation o f the moment centre reference frame from the origin o f the earth reference frame: d  (4-12)  = Es sG+ Es R  EG  d  d  where d^G is the vertical position o f the centre o f gravity in the earth's reference frame.  Differentiating equation 4-12  EG  ~ ES SG  d  R  d  +  BS SG  R  d  +  (4-13)  BS  d  The sensor reference frame never changes place with respect to the centre o f gravity reference frame, so the second term on the right hand side is zero.  Expanding equation 4-13 into the three axial components solves for the vertical linear velocity o f the centre o f gravity, without contamination.  i  £  G  =^(-  s  i  n  #)*sG  +^-(  88  C 0 S  ^sin^  S G  +z  ES  (4-14)  where 0 is the roll angle o f the buoy relative to the earth, \j/ is the pitch angle o f the buoy relative to the earth, and the last term is the vertical velocity.  Expanding this equation gives z  FG  = (-cosd)x  + ( - c o s # s i n ^ + sin#cosi// )>> ,  SG  SG  +z  ES  (4-15)  The required data are available to solve this equation. R o l l and pitch angles are easily determined using the M rotation matrix. The vertical velocity can be found by integrating the accelerometer readings. The angular velocities can also be determined from the raw angular rate sensor data provide by the 3 D M - G sensor. The program can then find wave trough by looking at vertical velocity, rather than acceleration, and searching for periods where both the vertical velocity o f the buoy is zero, and the velocity readings are negative prior to approaching zero.  The primary disadvantage to correcting the algorithm this way is that it is unnecessarily complicated, and introduces four more sets o f data to the program that must be filtered. Each additional data set w i l l increase any errors already present i n the vertical velocity calculations. In addition, there w i l l be error accumulated in the integration process to get vertical velocity. It is really much simpler to just place the sensor in the appropriate location.  In both scenarios, it is required that the centre o f gravity does not move significantly as the rockets are launched.  Given the system weighs in the order o f 10 tons, the centre o f gravity  should not move substantially regardless o f how the rockets are launched. However, it would be beneficial to alternate launches on opposite side o f the buoy.  89  4.3.3  Optimizing the Launch Angle According to Transport Canada regulations, a rocket must be fired within 10 degrees o f  vertical to be considered a safe launch. Therefore, the minimum launch angle, a j is 10 degrees m  ni  for the launch program. However, it is also desirable to fire as close to vertical as possible. It is not effective for the program to fire a rocket at 10 degrees i f a significantly smaller angle is a practicable choice.  The launch program therefore needs a method for determining the smallest practicable minimum angle for launch, a , appropriate for the given wave conditions. This is done using a op  history algorithm that is implemented 20 minutes before the desired firing window. During this period, the normal launch algorithm operates, searching for minimum angle opportunities during wave troughs. However, during this period, no ignition signals are given, and the alpha for each launch opportunity is stored i n memory instead. A t the end o f the 20 minutes, these alphas are sorted to create a probability distribution o f the minimum launch angle possible in current wind and wave conditions.  If no launch opportunities are found, the system defaults to the maximum permitted alpha angle of 10 degrees. Otherwise, the program uses a predetermined acceptable probability (e.g. 3 launch opportunities i n the next 10 minutes). distribution to find the corresponding a  o p  for P. This value is then fed into the launch program as  the new a i for launch. m  The algorithm then uses the probability  n  90  The probability-based algorithm results in launches closer to vertical than would be achieved by simply looking for the first safe angle. It also guarantees that no launches w i l l occur at more than 10 degrees inclination.  4.3.4  Algorithm Implementation The launch-control algorithm was implemented using F O R T R A N . This language was  selected due to the availability o f existing interface subroutines.  The basic strategy o f the complete launch-control software is quite simple. First, a short time prior to the desired launch period, the program implements the Launch Angle Optimization Program for 2-20 minutes. This program determines a  opt  , the optimal launch angle possible, for a  reasonable launch probability in the given wave conditions,. This optimized angle is always smaller than the largest angle considered safe for launch, 10 degrees. a  o p t  is submitted to the  basic launch angle selection algorithm, which begins searching for the next possible launch opportunity. The complete steps to the launch control software, in pseudo-code, are as follows: •  C o n t i n u o u s l y  •  When  time  i s  i n i t i a l i z e •  Run  Check  o  P o l l  o  I f  the  2-20  minutes  OPTIMIZATION the  a  before  s e n s o r  to  asin(V(M  =  the  get  3DM-G  2 3 2  l a u n c h  I f  o  Update  the  1 0 - r e a d i n g  o  Update  the  4 - r e a d i n g  o  I f  10  s u c c e s s f u l a  r e  < I f  iative  (a  M  m a t r i x  to  2  3  get  raw  a c c e l e r o m e t e r  d a t a  then =  a  s e n  sor  degrees >  s t a b i l i z e d  Mi) )  +  P o l l  a  d e s i r e d  then  o  •  the  SUBROUTINE  o  '  c l o c k  c l o c k  the  s u c c e s s f u l •  i n t e r n a l  program  LAUNCH  o  check  a  {indicating  a v e r a g e  M  +  *9.66  r u n n i n g  (the )  3 3  r u n n i n g  safe  AND a  a launch  91  has  average average  launch p a s t  a a  a v e r a g e  a v e  rage  angle) a  MINIMUM  opportunity)  time,  OR •  I f  a  i s  a v e r a g e  •  IF  v e r y  (a iative < a ( indicating r e  zero  {indicating  s m a l l a v e r  no  I  tilting)  a erage I i s C l o s e trough)  age) A N D a wave  to  aV  OR I f a age i s v e r y s m a l l vertical motion) o This is a launch o N = N + 1  •  o • • END  o  I f  o  Resort  o  P  =  the  a  than  20  i n t o  i n c r e a s i n g  =  o p t  o f  launch  o p p o r t u n i t i e s  Check  o  P o l l  o  I f  window  (e.g.  where  launch  the  P  x  i s  a  the  P o l l  o  I f  sensor  launch  =  to  asin(V(M  the  3DM-G  get +  2 3 2  to  get  angle  p r o b a b i l i t y  r e  s e n  of  Update  the  10-reading  Update  the  4-reading  o  I f  a  o p t  I f  o p t  a ) opt  (a  m a t r i x  2  3  accelerometer  > a e age) a V  r  data  running running  AND a  very  ave  IF  *9.66 average  has  r e  launch past  a v e r a g e  a V  angle)  a • MINIMUM  opportunity)  small  (a iative < a ( indicating  a  a erage  average  a launch  OR • I f a rage i s zero  3 3  (the optimal  degrees  {indicating  •  M  M i ) )  sor + M  o  <  a  (input  s t a b i l i z e d  raw  o  •  5)  =  g i v e n  then  a iative = a  a  x  the  then  s u c c e s s f u l  "  i n  c l o c k  the a  o  where  o p p o r t u n i t y  o p t i m a l  s u c c e s s f u l •  window)  d e s i r e d  A N G L E (K)  Output  o  REPEAT order  minutes/time_intervaliaunch  launch a  minutes,  LAUNCH OPPORTUNITY SUBROUTINE  Run  a  l e s s  i n t e g e r ( N / P )  f i n d i n g o  i s  number  upcoming  o  =  IF  N ANGLEs  (x/N)*(20  K =  opportunity  IF  time  the  o  ANGLE(N)  no  IF  the  is  RECORD  END  END  o  (indicating  aver  a v e r a  (indicating  no  g e ) A N D I average I i s a wave trough)  tilting) close  to  OR •  I f a verage i s v e r y s m a l l vertical motion) o This is a launch a  o o  Send  a  s i g n a l  to  (indicating opportunity the  p a r a l l e l  complete  the  i g n i t i o n  c i r c u i t  Turn  the  i g n i t i o n  c i r c u i t  o f f  92  no  port  to  • • o o  END  END  Put  o  COMPLETELY  END  a l l  systems EXIT  i n  safe  mode  T H E PROGRAM  IF  IF  IF  Repeat  S u b r o u t i n e  programmed •  o  l a u n c h  u n t i l  time  i s  g r e a t e r  than  the  window  END  Figure 4.11 shows sample data for the algorithm. The possible launch times are marked with an "x". Launch opportunities are only marked for just past minimum alpha readings that also occur when the accelerometer reading shows the buoy going into a trough.  6.0  TIME  (s)  Figure 4.11: Launch Opportunities 4.3.5  Filtering Implementing the launch control program using the MicroStrain 3 D M - G was relatively  simple, as the signal contained a minimal amount o f noise. The vertical acceleration readings were minimally affected, and noise in the signal rarely caused false launch opportunities. Noise  93  effects on the alpha measurements were a different issue. Noise would cause false minimums i n the angle readings, which the program would then see as launch opportunities, and thus miss better launch opportunities in the future.  Several types o f filtering was tried to remove this problem, including L o w Pass Filtering, and running averages. The best solution was found to be using a 10-reading running average for the previous alpha reading. This eliminated the false launches without substantially increasing the true launch opportunity angle from the actual minimum angle, as shown i n Figure 4.11. A small increase i n angle occurs because the averaging introduces a slight delay in the measurements.  A similar running average o f 4 readings was implemented as well to clear up the  small amount o f noise i n the vertical accelerometer readings.  4.3.6  Basic L a b o r a t o r y Testing Tests were done in the lab to ensure the program operated properly for conditions when  only tilt or acceleration was present. Figure 4.12 shows an arrangement created in the laboratory where the device is subjected to a modest amount of vertical acceleration. A s can be seen, the program fires at every possible minimal alpha, except those few very close to 5 degrees, where filtering removes them as opportunities. The program no longer considers the buoy's position i n the wave trough because it is not a concern in very shallow waves.  94  6.0  I  1  r  1  r—v  Q uJ <d  -2 0' ' 0.  1  10.  1  20.  1  1  1  30.  40.  50.  60.  TIMEw F i g u r e 4.12: L a u n c h Opportunities N o V e r t i c a l M o t i o n Figure 4.13 shows the opposite case, where tilt is kept to a minimum, but vertical accelerations are present. The important feature to note is that the program does not only order launches at minimum alpha readings, but at any opportunity it sees where the buoy is in a wave trough. The program is not as consistent for this case, as it misses about every second wave trough due to the filtering. Decreasing the filtering causes more false launches, so it becomes a balance between sensitivity and optimizing the launch angle.  In addition, the accelerations  provided in the laboratory are very small and close to the cut off where the program takes vertical motion out o f consideration. It is not simple to duplicate the level o f motion caused by a buoy bobbing in 3-9 meter waves.  Given the number o f launch opportunities provided in a  minute by these borderline conditions, and the fact only one good launch is needed, it is not a problem that the program misses one in three opportunities.  95  TIME  (s)  Figure 4.13: Launch Opportunities Minimal Tilt In general, the testing showed that the algorithm appears to work correctly for general cases and the specific cases o f no vertical acceleration, or not tilt motion.  4.4  Experimental Validation Once the program was tested in the laboratory setting using an L E D to represent a rocket  launch, it was then implemented in a series o f outdoor experiments.  The purpose o f these  experiments was to verify the program could be used to fire an actual rocket. In turn, this would demonstrate that the launch control system is a feasible solution to the problem o f launching rockets vertically on the Rocketsonde Buoy System.  96  4.4.1  Experimental Set Up and Procedure The outdoor experimental apparatus, shown i n Figure 4.14, had three main components.  First was the computer required to implement the launch control algorithm. The computer used in the experiments was the same as that used for the indoor laboratory tests. This was done for convenience, but the program could easily be adapted to run off a more portable computer, such as a laptop, in the future.  Figure 4.14: Outdoor Experimental Apparatus The second component was the ignition system. The main ignition source was a 20 volt portable power supply, modified to allow direct connection to an ignition circuit without triggering the short-circuit safety. This supply is the same as that used to ignite the rockets at the Harrison Lake launch site.  The ignition system is triggered by a signal from the computer's  parallel port closing a relay.  97  The final component o f experimental apparatus was the launch platform, shown in detail in Figure 4.15. The platform was suspended from a steel structure by elastic cord, to allow it to make both rocking and bobbing motion, similar to the motion o f a buoy in the ocean.  The  platform had a rocking period ranging from 1-3 seconds, which is more severe than that required of the launch control algorithm.  It proved a challenge to design the platform so that it had a  minimum period o f 5 seconds, even after arms were added to the platform to suspend weights and slow down the period o f the rocking. However, it was determined that testing could be done using this platform because successful tests for 3 second wave periods would ensure the algorithm would work for 5+ second wave periods. If the tests were not successful, the platform would be again adapted to have a greater period before conclusions were made regarding the functionality o f the launch control system.  F i g u r e 4.15: Close up of L a u n c h P l a t f o r m For each experiment, a model rocket, 20cm in length, was loaded with an A-level engine and mounted on a launch rod at the centre of the bobbing platform. The ignition circuit would be assembled, and the computer algorithm reset. One person would initiate rocking and bobbing  98  motion i n the platform until warned that the program was to be activated. Ideally, the platform would be given a three second period and 10-15 second amplitude when the program was activated, although this was occasionally difficult to maintain. Once that person was clear, the program would be initialized, and the rocket would launch at the first opportunity. Two digital video cameras were set up perpendicular to both the launch platform, and to each other, to record the launch. These videos could then be examined to find the platform angle when the rocket was ignited. When two cameras were not available, rocking motion was only initiated in one plane only.  4.4.2  Results and Discussion The main purpose o f the experiments was to verify that the system launched rockets in a  near-vertical direction. Therefore, the important data collected was the actual platform angle relative to the earth when the program ignited the rocket.  This was found by advancing the  digital videos o f the launch, and capturing the image for each video when the rocket first ignited. This frame was usually easily detected by a small puff o f white smoke. A vertical reference for the earth frame is determined i n the captured picture by finding a vertical line on a building i n the background, such as a window frame or corner.  A second line is then extended from the  bottom surface o f the bobbing platform, as shown in Figure 4.16. The angle between these two lines, minus 90 degrees, is the actual launch angle in one dimension. If only one camera is used, motion was one initiated i n only one dimension, so this is the actual launch angle alpha. If two cameras are used, alpha is calculated from combining the platform actual launch angle in the two separate images.  99  F i g u r e 4.16: Determining A c t u a l L a u n c h A n g l e There are two major sources o f error in this method o f determining the actual angle o f launch.  First there is a +/- 1 degree error from the sensor, as determined in section 3.2.  In  addition, there is a reading error o f approximately 0.5 degree i n the process o f reading the launch angles from the digital camera images. A l s o , i f the camera was not at the same height as the platform, then parallax errors would alter the angle as recorded by the camera. Combined, the actual launch angles, as determined from the images, have an error o f +/-1.5 degrees.  It is important to note that the angle determined here is the angle o f ignition, not the angle of the rocket's flight when leaving the platform. A s discussed in section 4.3, it is safely assumed that for the final design, using K motor rockets, these angles w i l l be equivalent. However, this experiment uses 1/2A and A class motors, which take up to 0.5 seconds to clear the launch rod. During this time, the platform continues moving such that a rocket ignited at 4 degrees may fly  100  at a 7 degree angle after leaving the platform. For this reason, it is the platform angle at ignition that is found, not the angle o f the rocket flight.  Tests with No Angle Optimization Two sets o f experiments were performed.. First, seven launches were performed without using the optimization algorithm. The purpose o f these tests was to determine the functionality of the launch opportunity algorithm independent o f the effects o f the optimization subroutine. The maximum launch angle was set at 5 degrees for safety, given the confined space o f the launch site, instead o f being passed by the probability algorithm. The results for these tests are shown in Figure 4.17.  0 -I  1  ,  1  ,  2  ,  3  4  Launch + —  p r o g r a m  r.  i  a n g l e  5  6  7  Number — • — a c t u a l  a n g l e  Figure 4.17: Experimental Results No Optimization  101  A l l rocket launches occurred within the 5-degree limit determined by the program, indicating that the system effectively follows the safety limit.  More significantly, for all  experiments, the actual angle o f launch was less than, or within error tolerance, of the launch opportunity angle recorded by the program.  From these results, it appears that the basic algorithm works well. The rocket actually ignites at an angle which corresponds to the launch angle recorded by the computer.  This  demonstrates the functionality o f the 3 D M - G sensor in this application. In addition, these tests show that the algorithm succeeds in finding safe and optimal launch opportunities. The majority of the actual launch angles are 3 degrees or less, which is much better than the 5 degree limit set as a maximum angle limit.  Tests w i t h a A  n r  Optimization  second set o f experiments were performed using the complete launch control  algorithm, including the angle optimization component.  In these experiments, the optimization  algorithm was run for 1-3 minutes at the beginning o f the experiment. The platform was given continuous motion by a person pulling a series o f pulleys. The aim was to provide repeatable rocking motion throughout the experiment. The optimization algorithm recorded all the launch opportunities. For these tests, P, the launch probability, was set as a constant i n the algorithm, rather than x, the acceptable number o f launch opportunities i n the launch window. A n example set of data is shown i n Figure 4.18. A n ' x ' on the second graph marks the a program.  102  o p  calculated by the  •<o.  r/M£(s)  F i g u r e 4.18: L a u n c h Angle P r o b a b i l i t y G r a p h (P =0.6)  Five launches were performed using the optimization algorithm. The results are detailed in Figure 4.18.  In the case o f the first four launches, the actual launch angle is within error  tolerance o f the program launch angle.  This again supports the results o f the previous tests.  However, the last launch occurs at 10 degrees, which is much higher than the angle limit determined by the optimization program, and clearly not within error bounds o f the angle the program determined.  This result is perplexing until digital video is further examined. During  this rocket launch, the platform had a period o f approximately 1 second with a 17 degree amplitude.  Both these specifications are quite outside the limits for which the launch control  system is designed.  A l l other tests have a period i n the order o f 2-3 seconds, for which the  program has much less problems.  This test proves that the current system is therefore appropriate for rocking motion with a period o f three seconds o f more, but w i l l not work for periods o f 1 second or less. This is not an issue for the proposed Rocketsonde-Buoy System, as the N O M A D buoy w i l l never have this  103  small a period.  However, it might be a constraint i f the system was ever implemented on a  smaller buoy with different characteristics.  0  -I  1  1  1  1  2  3  1  4  5  Launch  —•— program angle —«— actual angle  Op Alpha  Figure 4.19: Experimental Launch Using Optimization  In Figure 4.19, the program launch angles are all below the optimized minimum launch angle. In addition, the a  o p  set by the history algorithm is also much less than 10 degrees in all  cases, demonstrating that this probability subroutine is beneficial to achieving better nearvertical launches than just using a set maximum angle.  Problems did occur when the launch history recorded very few launch opportunities during the period the launch optimization algorithm was run. U s i n g a constant probability, P, became a major problem, as the program would sometimes decide it was sufficient to have only one probable launch opportunity i n the launch window. If rocking conditions changed even a  104  small amount, that opportunity would disappear and the program would not fire at all. This occurred during 3 to 4 trials.  It was decided after these tests to implement the algorithm with the criteria of having a minimum number o f launch opportunities in the launch window, and to use this criteria to determine the launch probability P as part o f the algorithm. This is how the algorithm has been written in Section 4.3, and future tests w i l l use this criteria for the launch angle optimization subroutine.  4.4.3  E x p e r i m e n t a l Conclusions In general, the experimental results showed that the launch-control system is successful  in igniting a rocket at an optimal angle.  In all but a single test, the actual launch angle  corresponds to the launch angle generated by the 3 D M - G sensor and launch control software. In all cases the launch angle was below the 10 degree safe launch limit. optimization program correctly determined  appropriate  The launch angle  new launch control angles, which  improved the system's ability to launch close to vertical. The launch control system is, however, limited to rocking motion with a minimum 3-second period as higher frequency motion causes large errors in the launch angle.  4.5  Conclusions and Final Launch Control Design A major concern regarding whether the Rocketsonde B u o y System feasibility was how  the rockets could be safely launched vertically off a rocking buoy in the North Pacific. The main purpose o f the launch control system detailed in this chapter is to achieve near-vertical,  105  autonomous rocket launches for buoy motion conditions up to rocking with a 15° amplitude and 5 second period. This corresponds to the worst-case winter storm weather i n the North Pacific.  The system was implemented using a MicroStrain 3 D M - G 3-axis orientation sensor to provide the buoy motion relative to an inertial earth frame.  The sensor uses three orthogonal  accelerometers, 3 orthogonal magnetometers and 3 angular rate sensors to provide the rotation matrix M between the inertial earth frame and the rocking launch tube frame. The raw sensor outputs can also be accessed to get angular rate and linear acceleration data. Preliminary tests showed that the sensor is not subject to drift from being on an angle for an extended time, corresponding to a permanent' list in the buoy, and that the presence o f metals does not significantly affect the angular readings. Tests also showed that the sensor has an accuracy o f ±1°, which is slightly larger than the ±0.5° accuracy specified by the manufacturer, but still considered reasonably small for the purpose o f the launch control system. The MicroStrain was determined to be a good choice for the system.  For convenience, the launch control software was written i n F O R T R A N . A lower-level language would likely be used for practical implementations. The basic format o f the program was first to initiate a 2-20 minute launch angle optimization program, followed by the actual launch algorithm which fires a rocket at the first opportunity. The optimization program records all the launch opportunities for the given time period and creates a probability spectrum for the launch angle a.  It uses these data and a given desired probability o f launch to determine the  optimal launch angle,  a t, op  which becomes the maximum launch angle at which the rocket w i l l  launch. The purpose o f this program is to prevent the rocket from firing at a high angle when  106  current wave conditions suggest a much better launch opportunity w i l l be available within the launch window.  The actual launch algorithm then looks for launch opportunities where the launch tube angle is less than a  opt  , and this angle is not still decreasing, which would indicate a better  opportunity w i l l be available shortly.  In addition, it looks for these opportunities when  accelerometer readings indicate that the buoy is in a wave trough, where there is less wind to interfere with the rocket flight. This program also considers special cases where there is no tilt or no vertical motion o f the buoy, and responds appropriately.  Experimental tests using this system to ignite real model rockets demonstrate that the system is capable o f firing rockets within 5 degrees o f vertical off a rocking and bobbing platform with a period as small as 3 seconds and amplitude o f 10-20 degrees. This indicates that the system should work for the proposed N O M A D buoy, which has both a smaller amplitude and larger period. The system also respects the 10 degree safety limit by refusing to fire i f the launch tube is at a greater angle.  The angle optimization program is demonstrated to improve the  system's ability to launch near vertical and is a valuable component.  In general, the launch control system, as detailed above, fits all the system requirements and demonstrates that vertical launching o f rockets from the Rocketsonde Buoy system is feasible.  107  5.  Conclusions and Future Work The purpose o f this research is to evaluate the feasibility o f the Rocketsonde-Buoy  System. It focuses specifically on the major challenge o f designing an appropriate rocket and guaranteeing that it w i l l launch vertically even during the most severe winter storm conditions. Over the course o f the work, the feasibility o f this system has been examined through the discussion o f existing technology and its limitations, the simulation, design and testing o f various proof o f concepts and prototypes, and the analysis o f those test results.  Simulation and experimental results demonstrated that it is feasible to create a practical rocketsonde design for the Rocketsonde Buoy system.  Results indicated that the final design  should use a short burning, K class motor and a dual-diameter body tube design. This is a small enough design that it is practical to consider placing on a buoy.  Experimental launches also  show that tube-launching does not impair the rocket's flight or affect the motor. Tube launching provides a convenient way to protect the rocket from the elements, which makes it feasible to operate this system in storm conditions. A 6 fixed-fin design for the dual-diameter body tube w i l l allow the rocket to fit i n a 4 inch diameter tube, which is small enough to place several hundred on a buoy. In addition, since a fast burn motor w i l l be employed, the tubes only need to extend a metre longer than the rocket length. This is low enough not to alter the buoy's centre o f gravity significantly and cause instability.  Trade-offs, however, must also be made between the desired altitude and the number o f rockets on the buoy, rocket mass and rocket diameter.  108  Rocketsondes can be designed to reach a  6km altitude, but only 200 can be stored on a N O M A D buoy. The system in this case could still be serviced once a year, but it would only have daily launches half the year, during the F a l l Winter-Spring storm season.  If 400 rockets are desired, the rocketsondes w i l l only be able to  reach 3-4km. In addition, results show that the currently available short-burn motors cause the rocket to exceed the acceleration limit. However, a medium to fast burn motor, for example a K 8 0 0 or K900, should still have the desirable fast burn, while keeping the rocket acceleration lower. Future work is being planned with Cesaroni Technologies to develop a dual-pulse rocket motor. This motor would start with a fast burn which would allow the rocket to launch from a short tube and be unaffected by wind conditions, then to ignite a slow-burn component before the rocket reaches the 5 0 G instrumentation acceleration limit. Cesaroni is optimistic about this project and i f successful, it w i l l provide a solution to the problems with the acceleration limit. Also, the possibility o f strengthening the sonde to withstand 70-80 G forces has been discussed with Vaisala.  The research has also demonstrated that it is feasible to achieve near-vertical autonomous rocket launches for buoy motion corresponding to the worst-case winter storm weather in the North Pacific.  A launch control system implemented using a MicroStrain 3 D M - G 3-axis  orientation sensor was capable o f firing rockets within 5 degrees o f vertical off a rocking platform with a period as small as 3 seconds and amplitude o f 10-20 degrees. The system also respects the 10 degree safety limit by refusing to fire i f the launch tube is at a greater angle. A n additional optimization program records all the launch opportunities for the given time period and creates a probability spectrum for the launch angle a. Experimental results showed that this program improved the system's ability to launch near vertical.  109  Tests also showed that the  sensor is not subject to drift from being on an angle for an extended time, and the presence o f metals does not significantly affect the angular readings.  In brief, the launch-control system fits  all the system requirements and has demonstrated experimentally that near vertical launching o f rockets from the Rocketsonde Buoy system is feasible.  In conclusion, research has shown it is feasible to design a battery o f rocketsondes to fire autonomously from a buoy, and to guarantee that these rockets w i l l launch close to vertical. Additional work, however, needs to be done to either improve the acceleration limit o f the sonde electronics past 5 0 G , or to develop a custom rocket motor that w i l l keep the rocket acceleration under 50 G while still providing a high initial impulse. Additionally, it has been determined that it is not possible for the system to contain 400 rockets capable o f reaching 6 k m . The effect on forecasting o f having daily launches at 4 k m needs to be examined and compared to the effect o f having 6km launches only half o f the year, to determine which solution is preferred. Once these two issues have been resolved, there should be no major technological or theoretical barriers to adapting and implementing the Rocketsonde-Buoy system on a larger scale.  Possible future work on this project, now that the system has been determined to be feasible, would be the creation and testing o f a small prototype, capable o f autonomously launching mid-power rockets. This would ideally consist o f a real commercial buoy and provide a means to test the launch control system i n actual wave conditions, rather than motions created by a bobbing platform. A l s o , the launch control program could be expanded in this prototype to include additional inputs, such as meteorological data, transmitted signals, and interfaces with the rocketsonde receiver. In such a prototype would require the issue o f weather-proofing the  110  system would need to be addressed.  It would also likely reveal problems related to marine  implementation o f the system not previously considered. protected waters, then at a deep-ocean testing site.  111  The system could be tested first in  References Langland, R . H . et. al. "1999: The North Pacific Experiment ( N O R P E X - 9 8 ) : Targeted observations for improved North American weather forecasts". Bull. Amer. Meteor. Soc.,80: 1363-1384. 1  Stull and Spagnol. "Rocketsonde Buoy System Observing System Simulation Experiments". Submitted to Monthly Weather Review (February 2003)  2  Resolutions of the Union and of the Associations, Geophysics, Hamburg Germany, 1983. ,  3  Final Buoy Field Service Plan:2002-1, O D A S Buoy Service Trip, M a y 2002.  4  International U n i o n o f Geodesy and  Environment Canada Pacific Y u k o n Region Annual  McTaggart-Cowan, R. "Research Notes for Tethered Guided Balloon and Rocketsonde Launch System". Report. University o f British Columbia, 1998.  5  Shapiro, M . A . and Thorpe, A . J . "Report on the Current Status o f The Observing-system Research and predictability experiment (THORpex)". Report. 2002. 6  Cole, H . L . et. al. "Development o f an Advanced Balloon Platform and Dropsonde for use in The Hemispheric Observing System Research and Predictability Experiment". Report. National Centre for Atmospheric Research, 2001. 7  Holland, G . J . et. al.. "The Aerosonde robotic aircraft: A new paradigm for environmental observations". Submitted to Bull. Amer. Meteor. Soc,, 2000. 8  The Aerosonde System. Aerosonde Pty Ltd. Available at <www.aerosonde.com> See Appendix A .  9  North Nomad (C46184) February 2002. Marine Environmental Data Service. Department o f Fisheries and Oceans. Government o f Canada. <www.meds-sdmm.dfo-mpo.gc.ca> 1 0  Readyhough, C . " O D A S Buoy 2002 Yearly Servicing Trip Report" University o f British Columbia, A p r i l 2002. 11  1 2  6 Metre N O M A D Buoy Data sheet. A x y s Environmental Systems. Published 1997.  1 3  Clark, D . D . " A meterological rocket sonde". Sci. Instrum., 42 (1965): 733-36.  1 4  R K 9 1 L o w Altitude Rocketsonde Technical Data Sheet. Vaisala Inc. Published 2001.  112  Sea Launch: Cruising into Orbit, <www.sea-launch.com> See Appendix A . 1 5  Gilhousen, D . B . " A Field Evaluation of N D B C Moored B u o y Winds, Journal of Atmospheric and Oceanic Technology". 4 (1987): 94-104. 1 6  1 7  Makher, G . The Flight of Uncontrolled Rocket. N e w York: Pergamon Press, 1964.  Barrowman, J.S. "The Practical Calculation o f the Aerodynamic Characteristics o f Slender Finned Vehicles". Dissertation. School o f Engineering and Architecture o f the Catholic University o f America, 1967. 1 8  Design of Aerodynamically 14-5-133..  19  Stabilized Free Rockets, U S Department o f Defense, 1990, pp 5-  2 0  Stine, G . S . Handbook of Model Rocketry, Chicago:Follet Publishing Company, 1970.  2 1  Warren, F . A . Rocket Propellants.  N e w York: Reinhold Publishing, 1958.  Hypertek H y b r i d Rocket Technologies. Cesaroni Technologies Inc. <www.cesaroni.net/hypertek.html>. See Appendix A .  2 2  2 3  Readyhough, C . " P M E L Trip Report". University o f British Columbia, M a y 2001.  Model 904-T "Clinometer Pack" Data sheet. Applied Geomechanics. Published 2000. T r i m Cube Data Sheet. M D L Measurement Devices L t d . 2 4  Readyhough, C . " O D A S B u o y Depot Trip Report". University o f British Columbia, M a y 2001.  2 5  2 6  T R I A X Y S Wave Sensors Data sheet. A x y s Environmental Systems. Published 1997.  2 7  3 D M - G 3-axis Orientation Sensor Data sheet. MicroStrain Microminiature Sensors.  113  A P P E N D I X A : Reference M a t e r i a l from Webpages  114  The Aerosonde System The Aerosonde U A V is developed and operated globally by Aerosonde Pty L t d (AePL) and Aerosonde North America ( A e N A ) . It has been undertaking operations for about 7 years, was the first U A V to cross the Atlantic Ocean, has flown for over 32 hrs i n one stretch and has undertaken continuous operations with relay aircraft extending over several days. The great flexibility o f the Aerosonde, combined with a sophisticated, command and control system, enables deployment and command from virtually any location. The Aerosonde U A V The Aerosonde was designed to a specification that would allow long endurance and economical flights anywhere i n the World. The specifications are provided i n Table 1. Table 1. Specifications of Mark 3 Aerosonde UAV Specifications (Weight, wing span _ _ §27-30 lb, 10 ft _ Engine J24 cc, 1.2 kw, fuel injected using premium unleaded petrol [Navigation [GPS Operation [Staff for Launch and Recovery J3 people: Controller, Technician, Pilot/Maintenance [Staff for Flight Operations 51 Person for up to 3 aircraft [Ground Equipment [Proprietary Staging Box, personal computer (laptop), GPS antenna, ; [aviation and local communications radios Flight [Fully autonomous, under Base Command Launch and Recovery [Launch from car roof rack (catapult option), land on belly, : [Autonomous or with pilot [Ground & air communications [UHF or Satcoms (Iridium) to Aerosonde, VHF to field staff and [ -other aircraft, internet to command center and users. Performance 18 - >32 ms" , Climb >2.5 ms"' at sea level 1  [Speed, Climb  [Range, Endurance with no [additional payload Altitude Range [Payload Standard Instrumentation Temperature, Pressure, [Humidity, Wind  Up to 20,000 ft (medium weight) [Maximum 5 lb with full fuel load h. Vaisala RSS901 Sondes for temperature, pressure and humidity, :and a proprietary wind system.  Aerosonde Payloads In addition to the standard meteorological instruments i n Table 1, a range o f other payloads has been flown on the Aerosonde and further additions are anticipated. These are described in Table 2. Table 2. A e r o s o n d e P a y l o a d s  Status Operational  Payload Type Surface Temperature: KT11 IR Sensor  Mission  Still Camera: Olympus  Surveillance, environmental,  Meteorological and environmental  115  Operational  Video Camera: Various Electronic Warfare Comint: Various instruments Chemical Sensor: Sulfur and carbon compounds, NASA JPL Radio data relay Magnetometer: Purpose built for Aerosonde  biological. Surveillance, environmental, biological. Surveillance and deterrence Surveillance and deterrence Volcanic plume and atmospheric chemistry Support field operations Mineral survey  Operational in fixed mounted mode Under development, with initial test flights accomplished Under development Flight tests completed Under development Instrument nearing completion Suitable unit identified and funded for integration Unit beingflighttested  SAR  Surveillance  Cloud Physics: NCAR Heymsfield  Meteorological and environmental  Laser Altimetry  Surveillance, ice and topographical  Under consideration  Extra capability w i l l be required for the envisaged missions under the A M R F , especially in regard to surveillance and reconnaissance payloads. Whilst the details w i l l require feedback from potential users, a generic estimate is included under development requirements. The Aerosonde Deployment and Command System Aerosonde U A V s are fully autonomous and capable o f making sophisticated operational decisions. They are operated in full accordance with civil aviation regulations. Through the Aerosonde Global Reconnaissance Facility ( A G R F ) , Aerosondes are deployed from designated launch and recovery sites, operate from a specified command center (with communications through the Iridium satellite system) and provide data to users through the Aerosonde Virtual Field Environment ( A V F E ) . The Command Center is the focus o f the entire operation, providing for mission coordination, a monitoring, regulatory and safety watch, the tasking o f multiple mission requirements, and monitoring the flow o f data. A global command center has been operational in Melbourne, Australia, for several years and another command center is under consideration for Hawaii. For the U S operations, a permanent command center would be located on the A M R F Base and a mobile command center would be moved to support operational requests. For example, the mobile command center could be located at a field program headquarters, or with the staff in control o f an emergency situation. Figure 1. The command center is based on PC technology and easily located in an office environment.  116  Launch and Recovery Sites are established on an as needs basis. The Figure 2. A minimalist Launch and Recovery Site requirements are relatively simple, a large field or open area, a small shed or other protection from the elements and a vehicle from which the aircraft is launched. W h e n on mobile operations the entire operation can be undertaken out o f a vehicle or a tent, subject to weather conditions. The Aerosonde Virtual Field Environment is a web-based data  Satellite or Radio Co mm "^k*  L E O Satellite  Ground itato  Real-time  Sat'land II n* or ht«rn«t  Visualisation, mission planning, analysis capacity  Figure 3. A typical set up and display for both the command center and virtual field environment, showing an Aerosonde in operation off the US east coast during the NASA CAMEX, overlain on satellite imagery. display system that enables users to access data and monitor the progress o f missions in real time from their own personal computers or office workstations. Users can also develop modified missions and upload these to the command center for implementation. Aerosonde Field Services includes all training, maintenance and specialized field operations. We have participated actively in the development o f the Australian U A V regulatory environment (the first in the world) and all operators w i l l be certified under these new regulations. We are also actively involved i n the establishment of F A A guidelines for U A V operations.  117  a  13 In addition to heavy lift performance capability of 4,500 - 6,000 kg, S e a Launch offers superior value, operational and cost advantages. Our marine operations reduce launch infrastructure, minimizing operational cost. Our continued focus is on customer satisfaction, mission assurance and evolutionary growth with emphasis on high performance, streamlined integration and efficient operations.  • • •  Launch to all inclinations from a single launch pad Our equatorial launch site provides the most direct route to orbit, offering maximum lift capacity for increased payload mass or extended spacecraft life Independent launch range scheduling and excellent environmental conditions  Proven, reliable components from the world's premier companies have been combined to create a revolutionary satellite launch service that maximizes payload capability, extends spacecraft life and delivers outstanding injection accuracy.  •  Our marine-based operations and highly automated systems, coupled with a customized launch location, are designed for a performance capability of up to 6,000 kg.  •  From analytical integration to spacecraft encapsulation to vehicle integration to automated launch processing, the S e a Launch partnership provides a complete launch service package, backed by a half century of experience and best practices.  118  IT •  •  In addition to high performance extending spacecraft life, S e a Launch offers operational and cost advantages and West Coast satellite processing facilitates integration operations for U.S. manufacturers. Orbital placement accuracy results in a reduction of on-board fuel consumption for final on-orbit maneuvering. With the ability to use the extra fuel, satellites can expect extra years of life.  The Concept: • • • • •  Launch commercial satellites to orbit from a platform at sea. Modern, accessible, user-friendly payload processing. Automated launch operations. All-inclination launch capability. Affordable, reliable, new-generation launch vehicle, comprised of capable, flight-proven components. Facilities and amenities of a U.S. launch site.  •  The Partners: • • •  Boeing Commercial S p a c e Company, Seattle, Washington, U S A Kvaerner, Oslo, Norway RSC-Energia, Moscow, Russia  •  S D O Yuzhnoye/PO Yuzhmash, Dnepropetrovsk, Ukraine  Sea Launch Rocket: • • • • • • •  Stages 1 & 2: Zenit-3SL. Stage 3: Energia-produced Block DM-SL. Payload enclosure and interfaces: Boeing. Widest diameter 14 feet. Overall length 209 feet. All stages kerosene/liquid oxygen fueled. Capacity to geosynchronous transfer orbit: 6,000 kg.  Assembly & C o m m a n d Ship: • • • • • • • •  Kvaerner. Modified roll-on, roll-off cargo vessel design. Rocket vehicle assembly facilities below decks. Launch control facilities on upper decks. Customer and crew accommodations for 240 people. Approximate length 660 feet. Approximate width 106 feet. Approximate displacement of 34,000 tons.  Launch Platform: • • • • • • •  Kvaerner. Modified, self-propelled, ocean oil-drilling platform. Rocket hangar, transporter-erector-launcher system, fuel storage/supply system with threestage launch capacity. Accommodations for 68 crew and spacecraft personnel. Approximate length 436 feet. Approximate width 220 feet. Approximate displacement: Surfaced - 30,000 tons. Submerged - 50,600 tons. 120  a  0  j  s 4a  a  ^ H y p e r T E K - Hybrid R o c k e t Propulsion S y s t e m  [JCesaroni Technology Incorporated (CTI) is announcing that it is now nmanufacturing and distributing the HyperTEK Hybrid Propulsion -•System under license from Environmental Aeroscience Corporation =keAc). H y p e r T E K is a modular, hybrid propellant rocket motor system Jthat uses readily available nitrous oxide (N20) and thermoplastic fuel Ugrains. Designed primarily for use in launching small experimental ^payloads by universities, colleges, research institutes and sport rocketry enthusiasts, the system is also intended to be a technology ^demonstrator for CTI's and eAc's hybrid propulsion technology. Extensive research and development by CTI and e A c has yielded a high tech, low cost rocket motor alternative to the current solid propellant rocket motor. Hybrid motors have several safety advantages over solids. The fuel and oxidizer are isolated from each other until just before ignition, thus eliminating the requirement for a low explosive permit during transportation and storage of rocket motors. The HyperTEK hybrid propulsion system uses a remote fill/fire launch system as well as 100% pyrotechnic free ignition system. The H y p e r T E K flight motor consists of three major parts: the oxidizer tank, the injector bell and the fuel grain. The fuel grain is molded from thermoplastic and has an insert-molded silica/phenolic nozzle. This monolithic grain functions as both the fuel and the combustion chamber. Complicated assembly of a reloadable system has been eliminated by incorporating the fuel, nozzle and motor case into one injection molded component. The fuel grain simply screws onto the injector bell prior to flight and the spent fuel grain is unscrewed and discarded of after each flight. The injector bell has field interchangeable injector orifices for tailoring the time/thrust profile to the needs of the rocket. Potential applications include small sounding rockets, R A T O boosters and more.  121  Features and Benefits •  Completely non-pyrotechnic and inert  •  Fuel grain is non-explosive / non-hazardous  •  Extremely clean burning and stable  •  High mass fraction  •  Unrestricted storage and transport  •  Low cost per flight  •  Indefinite shelf life  •  Non-toxic  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items