@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Applied Science, Faculty of"@en, "Civil Engineering, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Miyasaki, Grant W."@en ; dcterms:issued "2010-07-21T18:09:07Z"@en, "1987"@en ; vivo:relatedDegree "Master of Applied Science - MASc"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """A large percentage of automobile accidents in city traffic occur at speeds below 15 mph. Unfortunately there is a scarcity of experimental crash data at these low speeds to help investigators to reconstruct accidents. Accident reconstruction experts have consequently attached a low level of confidence to speed predictions from vehicle crush at the low end of the speed spectrum. The need for more experimental crash data, especially in a low speed range, has repeatedly been mentioned by accident investigators. The University of British Columbia Accident Research Croup has constructed a crash test facility in conjunction with the Insurance Corporation of British Columbia to address this need. The lCBC-UBC barrier is a low speed crash test facility. A description of the ICBC-UBC crash barrier, its systems and crash testing techniques at the ICBC-UBC facility are presented in this thesis. Also multiple impacts on the same vehicle are investigated to see if this technique provided accumulated crush data that reproduced known high speed crashes. In addition, the preliminary findings are presented on the impact speed to initiate permanent crush and subsequent implications toward vehicle crush characteristics in a low speed range."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/26722?expand=metadata"@en ; skos:note "AUTOMOBILE CRASH TEST FACILITY AND PRELIMINARY ANALYSIS OF LOW SPEED CRUSH CHARACTERISTICS by GRANT W. MIYASAKl B.A.Sc, The University of Toronto A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DECREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF CIVIL ENGINEERING We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September, 1987 © Grant W. Miyasaki, 1987 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 of Cli/ll £t\\«JI*.€tfi\\4 The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date OcJoljlr' f /?^ 7 DE-6(3/81) ABSTRACT A large percentage of automobile accidents in city traffic occur at speeds below 15 mph. Unfortunately there is a scarcity of experimental crash data at these low speeds to help investigators to reconstruct accidents. Accident reconstruction experts have consequently attached a low level of confidence to speed predictions from vehicle crush at the low end of the speed spectrum. The need for more experimental crash data, especially in a low speed range, has repeatedly been mentioned by accident investigators. The University of British Columbia Accident Research Croup has constructed a crash test facility in conjunction with the Insurance Corporation of British Columbia to address this need. The lCBC-UBC barrier is a low speed crash test facility. A description of the ICBC-UBC crash barrier, its systems and crash testing techniques at the ICBC-UBC facility are presented in this thesis. Also multiple impacts on the same vehicle are investigated to see if this technique provided accumulated crush data that reproduced known high speed crashes. In addition, the preliminary findings are presented on the impact speed to initiate permanent crush and subsequent implications toward vehicle crush characteristics in a low speed range. ii TABLE OF CONTENTS ABSTRACT ii LIST OF TABLES v LIST OF FIGURES vi ACKNOWLEDGEMENT ix 1. INTRODUCTION 1 2. LITERATURE REVIEW 4 2.1. First Full Frontal Barrier Impact Test 4 2.2. UCLA-ITTE Crash Test Facility 5 2.3. Ford Motor Company Test Facility 6 2.4. Road Research Laboratory Test Facility 10 2.5. General Motors Corporation Automotive Crash Test Facility 11 2.6. Crash Test Techniques at Fiat 12 2.7. Transport Canada Motor Vehicle Test Centre 13 2.8. A Few Other Crash Test Techniques 14 2.9. Conclusion 14 3. ICBC-UBC CRASH TESTING FACILITY DESIGN AND CONSTRUCTION 15 3.1. Introduction 15 3.2. Overview of the Crash Testing Facility 16 3.3. Test Site 18 3.4. Crash Barrier 20 3.5. Propulsion System 24 3.6. Speed Measurement 26 3.7. Recording of the Impact Event 29 3.8. Release Mechanisms 29 3.8.1. Main release mechanism 30 3.8.2. Redundant release mechanism 34 3.9. Data Collection 39 3.10. Safety Considerations , 41 4. VALIDATION OF ICBC-UBC CRASH TESTING FACILITY 42 4.1. ICBC-UBC Crash Testing Facility and SAE Recommendations 42 4.2. Validation of ICBC-UBC Crash Test Facility Results 43 5. HIGH SPEED CRASH DATA FROM REPEATED IMPACTS 53 5.1. Introduction 53 5.2. Campbell's Derivation of Energy Absorbed in Residual Crush 54 5.3. Absorbed Energy in Crush from Multiple Impacts 60 5.3.1. Force-residual crush response of multiple impacts 62 5.3.2. Quantification of crush energy from multiple impacts 64 5.4. Repeated Crash Tests 68 iii 6. BUMPER PERFORMANCE LEVEL AND LOW SPEED ENERGY ABSORPTION CHARACTERISTICS 73 6.1. Low Speed Bumper Tests 74 6.2. Graphical Interpretation of Crush Energy in the Low Speed Range .... 75 6.3. Crush Energy of the Two Regime Model 83 6.4. Three Regime Model of Vehicle Crush Characteristics 84 7. APPLICATIONS TO THE CRASH3 PROGRAM 88 8. CONCLUSION 98 9. FURTHER AND FUTURE RESEARCH 101 9.1. Further Research 101 9.2. Future Research 103 BIBLIOGRAPHY 105 APPENDIX I 109 APPENDIX II 115 APPENDIX III 119 APPENDIX IV 123 APPENDIX V 126 iv LIST O F TABLES Table 4.1 : Errors of fit/prediction 48 Table 4.2 : Honda test data 52 Table 5.1 : Double impact results 68 Table 5.2 : Double impact equivalent speed 69 Table 6.1 : Bumper tests 74 Table 7.1 : Speed-residual crush coefficients 91 Table 7.2 : Force-residual crush coefficients 91 Table 7.3 : Energy of deformation across the residual crush profile 94 v LIST O F FIGURES Figure 2.1 : UCLA-ITTE test setup 7 Figure 2.2 : Ford Motor Company crash barrier construction 9 Figure 2.3 : Towing arrangement 9 Figure 2.4 : Fiat catapult 12 Figure 3.1 : Towing arrangement 17 Figure 3.2 : Site plan view - original and prepared site 19 Figure 3.3 : Acceleration-speed-distance relationship 21 Figure 3.4 : Crash barrier 23 Figure 3.5 : Speed trap-remote electronics unit & sensor 27 unit Figure 3.6 : Main release mechanism 31 Figure 3.7 : Main release dynamic load test results 33 Figure 3.8 : Main release static load test results 35 Figure 3.9 : Main release calculated static release force - 36 maximum spring travel/spring constant/ static release force Figure 3.10 : Main release calculated static release force - 37 incline angle/friction coefficient/ static release force Figure 3.11 : Redundant release mechanism 38 vi Figure 3.12 : Redundant release dynamic load test results 40 Figure 4.1 : Crash test results of 1975 to 1981 Honda Civics 45 Figure 4.2 : Residual plot of lCBC-UBC, Transport Canada, 46 and Strother et al data Figure 4.3 : Frequency distribution of errors of fit/pred. 47 (Honda Civics) Figure 4.4 : Crash test results of 1971-1972 full size C M . 49 cars Figure 4.5 : Residual plot of ICBC-UBC and Campbell data 50 Figure 4.6 : Frequency distribution of errors of fit/pred. 51 (Full size C M . cars) Figure 5.1 : Force-residual crush response 56 Figure 5.2 : Measured force-crush responses 58 Figure 5.3 : Impact speed-residual crush curve 61 1971-1972 C M . vehicles Figure 5.4 : Crush energy of a vehicle repeatedly impacted 63 Figure 5.5 : Possible deviation of multiple impact data from single impact data 67 Figure 5.6 -. Multiple impact test results-1977 Honda Civic 70 Figure 5.7 : Multiple impact test results-1971 C M . Cutlass 72 Figure 6.1 : Single regime force per unit width-residual 77 crush model Figure 6.2 : Two regime speed-residual crush model 78 vii Figure 6.3 : Two regime force-residual crush model 80 Figure 6.4 : Higher order force-residual crush function 81 Figure 6.5 : Two regime energy plot 82 Figure 6.6 : Impact speed vs residual crush for full frontal 85 barrier tests for C M . Citations 1979-1982 Figure 6.7 : Three regime speed-residual crush model 87 Figure 7.1 : Impact speed vs residual crush 90 1975-1981 Honda Civic Figure 7.2 : Force per unit width vs residual crush 92 1975-1981 Honda Civic Figure 7.3 : Damage pattern 93 Figure 1-1 : Triangular deceleration pulse 112 Figure I-2 : Effect of wall movement on earth pressure 112 113 Figure I-3 : Crash barrier free body diagram Figure IV-1 : Force response for elastic and plastic crushing 124 viii ACKNOWLEDGEMENT This thesis would not have been possible without the assistance of many people. 1 would like to thank Dr. Francis Navin for his advice and guidance in preparation of this thesis, and for the coordination of the crash barrier project. I thank Mike MacNabb for his contributions to the design and coordination of the facility. His assistance was responsible for much of the success of the test facility. Thanks are extended to the staff at the Insurance Corporation of British Columbia - Material Damage Center. The assistance of John Gane, Bob Wilson, Larry Kenmare, and lain Saville in the construction and preparation of the test facility and help in testing of vehicles is greatly appreciated. I would like to thank Rod Nussbaumer for designing the speed trap electronics and also John Nepomuceno for building the speed trap switch box. In addition, I would like to acknowledge the work and assistance of Richard Postgate and Max Nazar in building the release mechanisms. Financial assistance came from the Natural Science and Engineering Research Council. ix 1. INTRODUCTION At the scene of an automobile accident, one of the most important pieces of information to the accident reconstruction expert is the damage pattern on the vehicle since this provides a vital clue to the speed change during impact. The residual crush profile provides a measure of the energy dissipated during the collision. More precisely, it indicates the quantity of the vehicle's kinetic energy that is absorbed in residual crush. The energy absorbed in residual crush is necessary for predicting speed changes of the vehicles involved in the accident. Residual crush is the permanent crush on the vehicle or plastic deformation. In order to determine the energy absorbed in residual crush, the crush characteristics of the vehicle must be known. The vehicle's front structure behaves like a plastic spring during a collision. The pre-compression force (force to begin displacement) and the increase in force with displacement (spring constant) are the parameters that describe the characteristics of a spring. Similary, these two parameters define the crush characteristics of the vehicle. The force to initiate permanent crushing and the increase in force with further residual crushing define the crush characteristics of the vehicle's front structure. Different vehicles absorb different quantities of energy per unit of residual crush. A vehicle that has a stiff front structure absorbs a greater energy per unit of residual crush than a vehicle with a soft front structure. In summary, the crush characteristics of a vehicle enable the quantity of energy absorbed to be calculated from which the speed change of the vehicle is determined. 1 2 The crush characteristics of the vehicle's front structure are derived from full frontal barrier collisions. In a full frontal barrier impact, the energy absorbed in residual crush is easily determined since all of the vehicle's kinetic energy is consumed in deforming the structure. To determine the stiffness of a specific vehicle model, many full frontal barrier impacts are conducted at different speeds. From the impact speed and residual crush data of these tests, the vehicle crush characteristics are derived. The source of most crash data is the compliance tests conducted for Canadian and U.S. Federal Motor Vehicle Safety Standards. These tests are full frontal barrier collisions in a medium impact speed range of 48 to 56 km/hr. Since most crash data is in this narrow speed range, there is a scarcity of experimental data at the low and high end of the speed spectrum. Consequently, speed estimates from the residual crush in a .low speed range are being challenged because of this lack of data. In response to this need for crash data, a test facility was constructed by the University of British Columbia Accident Research Group in conjunction with the Insurance Corporation of British Columbia (ICBC). The research for this thesis consists of two major topics. The first topic is design and construction of a reliable, low cost, low speed crash test facility. The calibration of the barrier was done by collecting impact speed and residual crush for a few different vehicles and compare them to known results. The second part of the research involves the analysis of data produced at the facility. A program of crash tests were conducted to investigate the following 3 two topics: •Determination of frontal crush characteristics from repeated impacts on the same vehicle - To reduce the cost of data collection, one method suggested is to use the same cars in multiple impacts. This technique was investigated because of its potential to save research money. The theory behind repeated impacts will be presented and crash tests performed to evaluate this theory. •Examine the energy absorption capacity of bumper systems - Tests were conducted to determine the energy absorbing capacity of the Honda Civic bumper system and/or impact speed required to initiate permanent crush. Chapter 2 is a literature review of the systems and techniques used to conduct vehicle crash tests. In Chapter 3 is a full documentation of the design and construction of the ICBC-UBC test facility, and Chapter 4 has the validation data of the ICBC-UBC test facility. A preliminary analysis of the multiple impact technique is presented in Chapter 5. An investigation of the threshold speed and crush characteristics at the low end of the speed spectrum is presented in Chapter 6. An application of the proposed crush behaviour model of Chapter 6 is outlined in Chapter 7 where an equivalent barrier speed is predicted from a example accident case. Chapter 8 is the conclusion and Chapter 9 identifies areas of further research. 2. LITERATURE REVIEW Vehicle impact testing has been quite extensive and encompasses a wide variety of different crash configurations as well as many different objects being struck. In addition, impact testing includes individual components as well as full scale vehicles. This review is confined to full scale barrier impact testing. The evolution of barrier impact testing is sketched from the early primitive, yet effective techniques of the 1950's, to the present sophisticated procedures. Different crash test facilities in North America and Europe are described with a focus on the following topics: •Crash barrier design '•Propulsion of the test vehicle •Directional control of the test vehicle •Speed control of the test vehicle •Instrumentation and data acquisition In addition, the motivation behind undertaking the full scale barrier testing will be outlined. 2.1. FIRST FULL FRONTAL BARRIER IMPACT TEST Barrier impact testing began in 1934 when C M . crashed a vehicle into a retaining wall. In that test the vehicle was driven at a low speed to allow the 4 5 driver to jump out just prior to impact. This technique was very simple since the driver controlled the vehicle. The vehicle's own engine was used for propulsion and a driver steered the vehicle and controlled the speed. However, the technique was limited to very low speeds, and the safety of the driver remained a problem. 2.2. UCLA-ITTE CRASH TEST FACILITY An automobile crash test facility was constructed by the University of California in the 1950's. The purpose of the crash tests were to compare the restraining features of the chest-type, lap-type and shoulder-type safety belt. In addition, the collapse characteristics of automobile structures were investigated. More specificially, the deceleration pulse was examined and the causes for departure of this deceleration pulse from the ideal uniform deceleration pulse were investigated. This crash barrier consisted of large diameter utility poles sunk 2.4 m into the ground and supported by cross members and braces. The barrier was backed by earth fill to provide additional resistance. The impact surface was 2.4 m high and 4.3 m wide. The test vehicle was propelled toward the barrier by pushing it with another vehicle (control car). The control car would push the test vehicle until it was 30 m from the barrier then break away, decelerate and stop 12 m short of the barrier. 6 Directional control was maintained by remote control steering that utilizes a selsyn motor. A selsyn motor is held in contact with an auxiliary steering wheel in the control car. Movement of this steering wheel rotates the shaft of the selsyn motor which changes the output from the selsyn. This output is transmitted to an identical motor in the test car which is held in contact with its steering wheel. The output produces a similar movement of the selsyn shaft and steering wheel. A person in the control car steers the test vehicle with the auxiliary steering wheel. Trial runs were conducted so the desired impact speed could be achieved. Trial runs showed that by pushing the test vehicle to 48 km/hr at 30 m from the barrier it would decelerate, while coasting, to the desired impact speed of 40 km/hr. Figure 2.1 is a diagram of the test arrangement. The test vehicle was fully instrumented. An anthropometric dummy was placed in the driver seat and instrumented with accelerometers in the chest cavity and head. Strain guages were mounted to record the impact forces on the foot rest, seat and chest level safety belt. The signals from these detectors were sent to the recording equipment on the control vehicle by an 30 m cable. High speed cameras were used to record the impact event. 2.3. FORD MOTOR COMPANY TEST FACILITY A crash test facility of the Ford Motor Company has provided full scale collisions during the 1950's and 1960's. Tests were conducted to determine what happens to occupants during a collision. The dynamics of the dummies and injury 7 Source : Severy, D.M. and Mathewson, J.H. \"Automobile-Barrier Impacts\" Figure 2.1 UCLA-ITTE test setup 8 mechanisms were investigated by observing the damage sustained by the dummy from striking the interior of the passenger compartment. The crash barrier is constructed of 0.6 m diameter logs embedded vertically 1.8 m into a trench that is packed with a concrete fill. The barrier is backed by a sand pile and the impact surface is faced with oak planks. The barrier is 5.5 m wide, 1.8 m high and 4.5 m long as shown in Figure 2.2. The test vehicle is towed by a tow car with a cable toward the barrier. The test vehicle is displaced 3.3 m laterally and 10 m behind the tow car. This arrangement is shown in Figure 2.3. Trial runs are performed on an open area to determine the steering wheel position which provides accurate tracking and correct impact direction. A release mechanism which is actuated manually, releases the tow cable just prior to impact. The towing distances for various impact speeds are established when the tow car is operating at full throttle while towing a 1800 kg car. Setting these towing distances allows control of impact speed. The test vehicle is instrumented with accelerometers mounted on the frame and floor pan of the passenger compartment to measure decelerations during impact, and tensiometers are mounted on the seat belt to measure loads during impact. Dummies with accelerometers in the head and stomach cavity are placed in the front and rear seats. The signals from these transducers are sent through an electric cable to an instrumentation van which follows alongside the test vehicle. High speed cameras placed at the side of the barrier record the impact, and high speed cameras mounted on the the test vehicle record the movements of the PLAN LOGS-2 It. DIA. 12 ft. IONG OAK PIANKIHG DIRECTION OF lUPACTsOf) 6 h.HIGH ROAD SURFACE | . ;~ PACKED Fill CONCRETE \\\\WM' ELEVATION Source : Fredericks, A.L. \"Automobile Crash Research\" Figure 2.2 : Ford Motor Company crash barrier construction INSTRUMENTATION VAN CRASHING CAR TOW BIRD Source : Haynes, A.L. Fredericks, R.H. and Ruby, W.J, \"Automotive Collision Impact Phenomena\" Figure 2.3 Towing arrangement 10 dummies. 2.4. ROAD RESEARCH LABORATORY TEST FACILITY A crash test facility at the Road Research Laboratory track in Crowthorne, England built in the 1960's has provided crash information on European cars. The crash barrier is constructed of 56 concrete blocks that are held together with tie rods. The barrier weighs 91,000 kg and is 1.8 m high, 3.7 m wide and 6.4 m long. The crash car has its ignition switched on, and the clutch and appropriate gear engaged. A control car pushes the crash car until its engine starts, and thereafter the crash car accelerates under its own power. The control car follows 'a short distance behind since the vehicles are connected by the electrical cables of the instrumentation. Directional control is maintained by remote control steering with a selsyn motor. A person in the control car remotely steers the crash car. The desired speed at impact is achieved by setting the throttle opening, the distance of the crash car from the barrier and the speed to which the crash car is pushed. Trial runs are conducted to determine these settings for the desired impact speed. The crash car is instrumented with accelerometers, event markers, displacement meters and strain guages. The acclerometers are mounted on the floor to measure decelerations of the passenger compartment. The event markers are contact strips which are placed on the front bumper to record the instant of impact. They are also placed behind the engine block to record the time from first 11 impact and the passenger compartment moving forward to the engine block. Displacement meters are placed in the passenger compartment to measure movement of the steering wheel column. The strain guages are clamped to the seat belts to measure the loads during impact. Dummies with accelerometers in the chest cavity are placed in the front seats. High speed motion cameras are used to record the deformation of the car and movement of the dummies. 2.5. GENERAL M O T O R S C O R P O R A T I O N A U T O M O T I V E CRASH TEST FACILITY General Motors Corporation built a crash test facility in the 1%0's to provide data for development of crashworthy automobiles and testing for compliance to U.S. Federal Motor Vehicle Safety Standards. The barrier at this facility consists of a formation of concrete packed with sand with a subgrade foundation. The mass of the barrier is 64,000 kg and the impact face is 2.1 m high and 3 m wide. The tow system consists of an endless cable that forms a continuous loop. The cable passes through a friction drive system composed of pulleys, drive-drum and tensioners. The cable is wrapped around the drive-drum and maintained in position by guide pulleys. A vehicle mounted on ground stands powers the drive-drum with a shaft connecting the rear axle to the drive-drum. Directional control of the test vehicles are maintained with a dolly and guide rail. The test vehicle is chained to the dolly which slides along the flange of the guide rail. Speed control is acheived by manually balancing the power input to the system against the system speed. A tach generator connected to the drive-drum provides a measurement and display of speed for the person controlling the speed. 1.6. CRASH TEST TECHNIQUES AT FIAT 12 Fiat has tried a number of different and unique techniques for crashing cars. One of these techniques is a catapult which launches a vehicle with springs at speeds up to 40 km/hr. This test rig is shown in Figure 2.4. Barrier Release dampers t ravel2 .3m max Source : Franchini, E. \"Crash Testing Evolution at Fiat\" Figure 2.4 : Fiat catapult Another technique attempted was to push the test vehicle with another vehicle. The test vehicle was steered by a long steering column which extended through the rear window to a person in the pushing car. As the vehicle nears the barrier, the pushing car brakes and the steering column slides off so the test car travels freely to the barrier. A more elaborate method of conducting crash tests was to radio control 13 the whole vehicle operation. The test vehicles were equiped with pneumatic jacks to steer, shift gears and actuate the clutch, accelerator and brake. Radio signals sent to the test vehicle actuated the various actions of the vehicle. Fiat's new Safety Center built in the 1970's is a versatile crash test facility capable of staging many different collision configurations. It has a 2 million kg solid concrete block for testing trucks of up to 18,000 kg and a smaller 320,000 kg concrete block. Vehicles are propelled by a two ton trolley rolling on rails in a culvert below the surface. The trolley is driven by a closed loop cable powered by an electric motor. The motor speed is controlled by a central computer. 2.7. T R A N S P O R T C A N A D A M O T O R VEHICLE TEST CENTRE A crash test facility in Blainville, Quebec was built for compliance testing of vehicles to Canadian Federal Motor Vehicle Safety Standards. Vehicle testing began in 1979 and to date over one hundred vehicles have been tested. The crash barrier is a solid 180 metric tonne reinforced concrete block that is 4 m high, 5.8 m long and 3.4 m wide. It rests on a concrete pad that is supported by piles. The test vehicle is propelled by a cable that is driven by a winch powered by an electric motor. Directional control is maintained by a guide rail. The test vehicle is chained to a dolly which slides along the guide rail. 14 Test vehicles are fully instrumented with anthropometric dummies and accelerometers. High speed cameras are mounted overhead, below and at both sides of the vehicle. 2.8. A FEW OTHER CRASH TEST TECHNIQUES Daimler Benz has staged collisions by propelling the test vehicle with a steam rocket which pushes the vehicle. The present test rig that Daimler Benz uses for propulsion and guidance of test vehicles is a linear motor drive which accelerates a trolley along a channel. The Swedish State Power Board used gravity to propel the test vehicles. In this technique, vehicles were dropped from a crane. Volvo has also used gravity to propel vehicles wherein test vehicles descended down a sloping track. 2.9. CONCLUSION 1. There are as many different methods used to run the tests as facilities. 2. Most barriers are becoming progressively more costly to build and complex to operate. 3. Proprietory data limited the availability of the results. 3. ICBC-UBC CRASH TESTING FACILITY DESIGN AND CONSTRUCTION 3.1. INTRODUCTION A need for low speed automobile crash data had motivated the UBC-Accident Research Croup to design and construct a crash test facility. The whole facility is comprised of six different design components. The components are: Site Layout Barrier design Propulsion system Speed measurement High speed video camera Release mechanisms Data collection Safety considerations Many of the components had, a number of alternatives considered before arriving at the final design. In this section the design and construction of each of the facility components will be described. The low cost of the facility should be emphasized. Unlike other test facilities, the ICBC-UBC facility was not intended to have expensive instrumentation and control systems because impact speed and residual crush is the primary information sought. Simplicity was sought throughout the design and construction. The design and construction of the testing facility was a joint project between the University of British Columbia - Accident Research Croup (UBC-ARC) and the Insurance Corporation of British Columbia (ICBC). The role of the UBC-ARC was to design and construct the testing facility while ICBC would provide the site, maintain the facility, and supply vehicles. Personell from both the UBC-ARC and 15 16 ICBC conduct the crash tests. 3.2. OVERVIEW OF THE CRASH TESTING FACILITY The configuration of the crash testing facility consists of a tow vehicle positioned directly behind the barrier on a roadway as shown in Figure 3.1. The test vehicle is positioned on an approach way at an acceleration distance from the impact surface of the barrier. Connection of the tow vehicle to the test vehicle is with a wire rope that is attached to the tow-vehicle having a main release mechanism and attached to the test vehicle through a redundant release mechanism. The wire rope is threaded through a conduit that extends the length of the barrier. The test facility is configured such that the tow roadway, approach way, and barrier location are aligned along a straight path. This arrangement was found most suitable for the geometry of the test site as well as being the simplest towing arrangement since it obviates pulleys or snatch blocks as required by other towing arrangements. The actual crash testing procedure once the pre-test vehicle preparation and facility checks have been performed consists of the driver accelerating the tow vehicle from its position immediately behind the barrier and towing the test vehicle from its position on the approach way. Once the test speed is attained, the driver stops accelerating and maintains a constant speed until impact. At impact the main release mechanism detaches the tow cable from the tow vehicle. The driver of the tow vehicle then deccelerates to a stop. REDUNDANT RELEASE MECHANISM MAIN RELEASE MECHANISM CRASH BARRIER / TOW CABLE \\ ^ -fy-TOW-VEHICLE / ( SAND FILL CONCRETE BLOCKS TEST VEHICLE Figure 3.1 : Towing arrangement 18 3.3. TEST SITE A plan view of the ICBC Surrey test site is shown in Figure 3.2. It is 195 metres long and 30.5 metres wide and is bounded by King George Highway at the west end, a residential lot at the east end and both commercial and residential lots on the north and south sides. An existing building that is used for vehicle repairs occupies 93 m of the west portion of the lot. The section of the lot behind this building is the testing area which measures 102 m long and 30.5 m wide. Preparation of the site involved clearing and levelling, and spreading a gravel surface for the tow road. A 15 cm deep \"approach\" depression was excavated for the barrier placement, see Figure 3.2. The testing site arrangement has the approach way to the barrier starting at the back face of the building and ending at the barrier impact surface 27.5 m away. The barrier extends 9 m from the end of the approach way , and the tow road continues from the back of the barrier to the east property line which is a distance of 65.5 m. The approach way and barrier lie within an existing chain link fence which provides a secure area from unauthorized personell entering the testing area. At the outset it was decided that the tow system would be designed to accelerate the test vehicle at 0.10 C . This acceleration governed the length of approach way and tow road. At an acceleration of 0.10 C over a 27.5 m approach way, the test vehicle can attain a velocity of 26 km/h. This approach length is sufficient for our earlier tests which will run at less than 26 km/h. For later tests at 3 PJ rt 03 King George Hwy. TJ n> TJ 0) n rt) Cb 3 0 . 5 m O 0) in rr 01 rr l O C n> L O a c rt) D 3 rt) H »- 3 o id 3 o 0) If n x MM % H y, x ' TJ •-1 rt) TJ 0J rt) a cn It a R e s i d e n t i a l Lots 6L 20 speeds between 26 km/h and 40 km/h, the approach way can be extended by opening two garage doors on the east and west sides of the building and towing the test vehicle through the building. The 65.5 m tow road allows enough distance for the tow vehicle to accelerate to 40 km/hr then decelerate at 0.5 C to a stop with excess roadway as a margin of safety. The curves of Figure 3.3 shows the acceleration-speed-distance relationship used in determining the roadway lengths. 3.4. CRASH BARRIER The low speed crash barrier is a rigid non-moveable wall designed for a 40 km/h impact of a 1820 kg vehicle. It is constructed of concrete blocks and earth fill. The blocks are solid concrete which interlock and are typically used for retaining wall systems. Each block measures 0.75 m wide, 1.5 m long and 0.75 m high, and weighs 1995 kg. The blocks interlock with a set of keys on the top and matching slots on the underside. Interlocking concrete blocks were chosen instead of a poured concrete wall for two reasons: the material and construction cost of an interlocking block barrier is much less than one of poured concrete. Also the blocks may be easily moved if necessary. 3 A total of 30 blocks and 23 m of earth fill are used in the barrier pictured in Figure 3.4. Eighteen blocks are stacked to form the barrier wall that is Figure 3.3 : Acceleration-speed-distance relationship 22 3.0 m wide, 2.3 m high and 2.3 m thick. Behind the barrier wall are two side block walls that are 2.3 m high and 3.0 m long. The side walls contain the 23 3 m of earth fill as well as providing resistance to movement. The whole barrier is set in a 15 cm deep excavation for increased resistance to movement. The overall dimensions of the barrier are 9.1 m in length, 2.1 m high and 3.0 m wide. The impact surface is 3.0 m wide and 2.1 m high and faced with 3/4 inch plywood in accordance with SAE Recommended Practice J850a. Two blocks were specially cast with a hole for placement of the 8 cm pipe through which the tow cable passes. The pipe is flared at both ends to reduce abrasion of the tow cable. The total mass and configuration of the barrier was designed by consideration of the applied forces during impact and resistance forces. A triangular deceleration pulse was used to approximate the vehicle deceleration on impact. The applied force during impact was then calculated utilizing Newton's second law: Force = mass * acceleration The resistance forces come from friction between the concrete blocks and ground, and the passive pressure of the earth fill. A net force during the impact event imparts a velocity to the barrier and as the impact force subsides a net resistance force deccelerates the barrier. Appendix I shows the calculations of the barrier movement. According to these calculations, the barrier is sufficient to limit movement to less than 1% of the permanent crush of the vehicle as specified in SAE F i g u r e 3.4 : C r a s h b a r r i e r 24 Recommendations (J850a). The total mass of the barrier is 100,000 kgs which also meets the recommendations of approximately 98,000 kgs of compacted earth fill or equivalent which is stated in an earlier version of the SAE Recommendations (J850). Although the barrier meets SAE Recommendations the crush results obtained will still require a barrier validation to ensure compatibility of results to other barriers. This involves comparing ICBC-UBC crush results to those of other crash barriers under similar test conditions. 3.5. P R O P U L S I O N SYSTEM As previously mentioned, the test vehicle is pulled by a tow vehicle with a wire rope. The tow vehicle is a Chevrolet pickup truck capable of providing sufficient power to accelerate a test vehicle at 0.10 G. From initial tests, the gravel tow road surface was found to provide sufficient traction to accelerate at 0.10 G. The acceleration of the tow vehicle is actually governed by the withdrawal force of the main release mechanism. The tow cable is a 9.5mm (3/8 inch) diameter, 12 x 24 construction wire rope that is 45.7 m long. A 12 x 24 construction was chosen since strength is required rather than flexibilty. The 6.4 tonnes breaking strength is more than adequate to sustain the towing force. Initially guidance of the test vehicle was to be accomplished by two means. 25 1. The steering is locked with the wheels directed along the line of travel. 2. The tow cable pulls the test vehicle in the direction of travel. The tow cable pull is maintained along the approach way centerline by the guidance conduit in the barrier. However, a trial test indicated that these two guidance methods were inadequate. It is difficult to set the wheels precisely directed towards the barrier and at a low speed the cable pull does not keep the test vehicle aimed at the barrier. These problems resulted in the test vehicle wandering off the approach way at the start of its acceleration. The simple solution was to have a person running alongside the test vehicle with one hand on the steering wheel. When the test vehicle reaches a speed of approximately 20 km/hr the person guiding the vehicle lets the vehicle travel the remaining distance on its own accord. At speeds higher than approximately 20 km/hr and the shorter length of tow cable, the test vehicle maintains better directional stability than at lower speeds and thus the directional pull of the tow cable is sufficient to keep the vehicle on course for the remaining distance. This guidance method has produced mixed success because of the gravel approaches. A guide rail system will be installed to improve the directional accuracy and reliability when a concrete approach is built. In this system a steel U-channel beam will be laid along the length of the approach way and the test vehicle's wheels will be guided in the trough of the U-channel. This type of system has been used successfully by most facilities in North America. 26 3.6. SPEED MEASUREMENT Impact speed measurement was made with three different devices. The primary speed measurement is obtained from a speed trap. The speed trap has a sensor unit and a remote electronics unit. The sensor unit is housed in a box that is 1 metre in length and width and 10 cm deep with three switch levers on top spaced 30.58 cm apart. The unit is set flush to the approach way surface 3 metres back from the barrier face, and positioned to one side of centerline for the front tire of the test vehicle to pass over. A picture of the speed trap is shown in Figure 3.5 . The remote electronics unit contains the timer and microprocessor for interpreting the signals sent from the sensor unit via wires. It is equipped with a digital display giving an immediate reading of speed in miles per hour and is accurate to within one percent. The speed trap is powered by 12v DC current provided by a 12v car battery. When the test vehicle's front tire trips the first switch an electronic counter starts, tripping of the second switch records the counter reading, and tripping of the third switch stops the counter. Thus, two timer readings are obtained which provides two speed measurements. Comparison of the two speed measurements indicates whether the vehicle was accelerating or decelerating prior to impact. If the two measurements differ significantly the vehicle's speed is changing and therefore the actual speed at impact is somewhat different than indicated by the speed trap. It is important to have an accurate speed measurement since even F i g u r e 3.5 : Speed t r a p - r e m o t e e l e c t r o n i c s u n i t & s e n s o un 11 28 a small difference in speed results in a substantial difference in the vehicle's kinetic energy. Two redundant impact speed measurements are available from a video camera and from a fifth wheel attached to the tow vehicle . The video camera is located on one side of the crash site perpendicular to the vehicle path at impact. The camera has an internal timer to record the time to 1/100 of a second. Impact speed is determined from playback of the video showing the time and movement of the vehicle past markers on the approach way. The fifth wheel looks much like a bicycle wheel. It is attached to the tow vehicle by an arm extending from a set of wheel forks and is rolled alongside the tow vehicle. An optical sensor sends a pulse signal - which has a frequency proportional to the angular velocity of the wheel - to an electronics unit on board the tow vehicle. The signal is processed and the tow vehicle's speed is recorded on tape. The tow vehicle's speed should be very close to the velocity of the test vehicle. Since the fifth wheel measurement gives an indirect measurement of the test vehicle's speed, it is the least accurate, however the measurement is valuable if the other measurement methods fail. Both the speed trap and fifth wheel give an immediate velocity measurement while the speed measurement from the high speed camera is obtainable after playback of the video. 29 3.7. R E C O R D I N G O F THE IMPACT EVENT Crash testing facilities typically capture the vehicle impact with high speed cameras that have a resolution of 200 to 1000 frames per second as specified by SAE recommendation J850a. These cameras are very expensive and require a professional camera operator. These cameras cost approximately $12,000 and the rental rate is $300 per day plus approximately $300 per day for a camera operator. Such equipment was too costly for this project. Standard commercial video cameras are positioned on both sides of the site perpendicular to the vehicle path at the instant of barrier contact. An overhead video camera is mounted on a boom directly above the barrier contact point. The standard video cameras are used for visual recording of the impact only. The scanning rate of 30 frames per second is not sufficient for micromotion analysis of the impact. 3.8. RELEASE MECHANISMS Two release mechanisms were designed to separate the test vehicle from the towing vehicle. The main release mechanism on the tow vehicle will usually release with each test, however if it fails a redundant release mechanism will detach the tow cable from the test vehicle. 30 3.8.1. Main release mechanism The main release mechanism is bolted to the rear bumper of the tow vehicle and is designed to release the cable upon test vehicle impact. Having the cable release from the tow vehicle prevents a whipping cable and need to rethread it through the barrier conduit after each test. The main release is shown in Figure 3.6 . The main release has two rollers: one roller is fixed in position while the second roller is free to slide up and down in a slot. Both rollers are free to rotate. The roller that is free to slide up and down is spring loaded with two springs that push the rollers together. A wedge with notches on each side is pushed between the two rollers' forcing them apart. The rollers seat into the notches on each side of the wedge. The force of the springs on the rollers resists the outward movement of the rollers from the notches when a withdrawal force is applied to the wedge. The incline angle of the wedge is shallow allowing the wedge to slide into position with minimal force. The steep incline angle in the notches produces a high resistance force against outward movement of the wedge. The tow cable is connected to the wedge with cable clamps. At impact the tension generated in the tow cable pulls the wedge from the roller assembly. Two required characteristics of the release mechanism are to withstand a high static load before releasing and yet release with low energy expended. Retension at a high static load is required to prevent premature releasing of the cable while the test vehicle is accelerated. A low release energy is desirable so that F i g u r e 3.6 : M a i n r e l e a s e mechanism 32 a minimal amount of energy is added by the release mechanism to the test vehicle's kinetic energy at impact . These two properties of the mechanism can be controlled by the altering the pre-compression of the springs, the incline angle in the notch and the spring stiffness. An impact test apparatus was set up to determine the release energy. This involved mounting the mechanism on a test stand and applying an impact load. The impact loading system consisted of a round steel guide bar hanging from the wedge with a stopper plate at the bottom end of the bar. Different sizes of steel weights dropped at different heights along the guide bar produced different impact loads and energy applications. The potential energy in the drop height and weight is converted to kinetic energy which is applied to the release mechanism. The wedge has a incline angle of 40 degrees and the springs have a stiffness of 1750 N/cm (1000 lb/inch). When the maximum spring travel is set to 1.3 cm, the release energy is 90 N*m (66 ft-lbs). Compared to 4700 N*m (3500 ft-lb) kinetic energy of a 1400 kg car travelling at 25 km/hr, the release energy represents only an 1.9% increase. Results of the release energy tests is shown in Figure 3.7 . A 2000 N force is the calculated static tension in the cable to pull the test vehicle at 0.10 G, however , a substantially greater force is generated when slack in the cable is taken up. Towing tests were conducted and the release mechanism settings sufficient to prevent premature releasing were determined to be: •1750 N/cm (1000 lb/in) spring constant *1.3 cm (0.5 in) maximum spring travel Figure 3.7 : Main release dynamic load test results U l 34 *40 degree incline in the wedge notches Static load tests conducted with the above release settings gave a withdrawal force of 4000 N (9001b). Figure 3.8 displays the results of static load tests using a 530 N/cm spring. Figure 3.9 is a plot of the calculated relationship between maximum spring travel, spring constant and static release force. Figure 3.10 is a plot of the calculated relationship between incline angle, friction coefficient and static release force. These latter two plots were used as a guide for selecting the release mechanism settings during the design stage. 3.8.2. Redundant release mechanism The redundant release mechanism acts as a link between the trailing end of the tow cable and the test vehicle. Two chains connected to the redundant release mechanism are hooked to the underside of the test vehicle and the wire rope is engaged into the release mechanism. A picture of the release mechanism is shown is Figure 3.11 . The principle behind its operation is a brass pin holding the tow cable to the release mechanism which will shear on impact only if the main release fails. The release mechanism has two steel plates between which a closed swaged socket is positioned. The closed swaged socket is crimped to one end of the tow cable and held in place between the plates by a shear pin. The brass shear pin is slid through 9.5 mm (3/8 inch) holes in one plate, the swage and the other plate, thus securing the swage and tow cable in place. The tow cable is released by shearing of the brass pin at the two interfaces between the plates and swage. The brass Figure 3.8 : Main release static load test results on Figure 3.9 : Main release calculated static release force -maximum spring travel/spring constant/static release force Figure 3.10 : Main release calculated static release force -incline angle/friction coefficient/static release force F i g u r e 3.11 Redundant r e l e a s e mechan i sm 39 shear pin is 9.5 mm in diameter and at the shear locations the pin is notched to a smaller diameter. The pin diameter at the notches is machined to 6.75 mm and will not shear at impact unless the main release fails. The closed swage socket is 1 11/16 inch at its widest point which will allow it to pass through the 3 inch diameter conduit in the barrier. Both static and dynamic load tests were performed on the shear pins using the same testing apparatus as the main release. The shear pin can resist 14,000 N which is 10,000 N greater than the withdrawal force of the main release. The energy to shear the pin is 156 N*m which is 66 N*m greater than the energy required by the main release. The shear pin size is sufficient to retain the tow cable during towing and does not add a significant amount of energy to the test car when the main release mechanism fails to operate. Figure 3.12 shows the relationship between shear area and energy to shear the pin derived from the dynamic load tests. 3.9. DATA COLLECTION The raw data collected from the tests are impact speed and residual crush. The residual crush measurements are accomplished by taking measurements before and after the collision. Prior to testing, the vehicle's dimensions are measured. The dimensions recorded are front or rear overhang, wheelbase, front and rear track width, and front or rear vehicle width. After impact, the same dimensions are recorded from which the residual crush is determined. The data form in Appendix II shows the additional vehicle information recorded. 1 6 0 0 . 8 Shear p i n area Figure 3.12 : Redundant release dynamic load test results 41 Measurement of elastic rebound and rebound velocity will be made with the video cameras. During the impact, four images are recorded from which maximum deformation just prior to elastic rebound is determined. This gives the dynamic crush, and the difference between the dynamic and residual crush is the elastic rebound. 3.10. SAFETY CONSIDERATIONS The test facility was designed with a number of safety features such as a 2.5 m high chain link fence encompassing the barrier and approach way, and a redundant release mechanism. A recommendation of all safety procedures to be followed and safety equipment required are outlined in the manual of Appendix III. 4. VALIDATION OF ICBC-UBC CRASH TESTING FACILITY 4.1. ICBC-UBC CRASH TESTING FACILITY AND SAE RECOMMENDATIONS Systems and components of the crash test facility were designed in accordance to SAE Recommendation J850a. The purpose of the recommendation is to \"establish sufficient standardization of barrier collision methods that results of tests at different facilities may be compared\". The specification for the barrier states \"The barrier size and construction should be sufficient to limit barrier face motion to less than 1% of the permanent crush of the vehicle\". Theoretically, the barrier mass is sufficient to meet this specification, however trial tests were required to determine the actual movement. The following factors were identified as possible reasons for data that is not comparable. •Excessive movement of the barrier •Speed measurement not accurate •Accelerating test vehicle at impact •Vehicle does not strike the barrier at exactly 90 degrees The extent of these factors were investigated in trial tests. Trial tests have shown a slight movement between the concrete blocks of the barrier. The whole barrier is not sliding on its foundation and there is no 42 43 permanent shifting of the blocks. However, any movement indicates that some of the kinetic energy of the vehicle is expended in movement rather than dynamic crush of the vehicle. Speed measurement accuracy was checked by comparing the measurements from the speed trap and high speed video. These two measurements were within 0.5 mph of each other (note: Imperial units are used from here on since speed-crush data has traditionally been presented in these units). An accelerating vehicle at impact will have inertial forces from the wheels which may increase the amount of crush. The speed trap gives two speed measurements from which acceleration at impact can be determined. i A vehicle was directed into the barrier at a twenty degree angle to the barrier face. The average residual crush was much less than a collision at 90 degrees to the barrier surface. Thus, directional accuracy is an important factor in full frontal barrier crash testing. If collisions occur at less than a 90 degree angle, then the test falls into the category of an angled barrier impact. From past crash testing experience, I estimate that an impact angle between 90°±5° is equivalent to a full frontal impact. 4.2. VALIDATION OF ICBC-UBC CRASH TEST FACILITY RESULTS Validation of the results from the test facility involves comparison to the data of other test facilities. A 1975 and 1977 Honda Civic and a 1971 CM. 44 Cutlass station wagon were chosen for validation purposes since abundant test data is available on these vehicles over a wide range of impact speeds. Figure 4.1 is a plot of impact speed versus residual crush of full frontal barrier impacts involving 1975 to 1981 Honda Civics. The data is from three sources: Transport Canada, Strother et al, and the ICBC-UBC test facility. The ICBC-UBC data is not in range of the Transport Canada and Strother et al data therefore direct comparison cannot be performed. Instead the ICBC-UBC data data can be compared to the estimated or predicted speed at the same crush levels. The predicted speed being a least squares fit of a linear model to the Transport Canada and Strother data. This estimated relationship between impact speed and residual crush is: V = 4.4 + 1.5C where C is residual crush in inches and V is impact speed in miles per hour. The close proximity of the ICBC-UBC data points to this estimated relationship indicates that the results are acceptable. The errors of fit of the Strother et al and Transport Canada data, and the error of prediction of the ICBC-UBC data values are shown in Figure 4.2. The ICBC-UBC values have a noticeably smaller error. In Table 4.1 are the errors of fit of the Transport Canada and Strother data and the errors of prediction of the ICBC-UBC data. The quality of the ICBC-UBC results is portrayed with a frequency distribution, as shown in Figure 4.3. The frequency distribution of the errors of fit and predictions is plotted along with a Normal distribution with the same variance. The distribution appears to be normally distributed. The number of standard deviations of the ICBC-UBC results from the predicted impact speed are shown in Figure 4.1 : Crash test results of 1975 to 1981 Honda Civics a 6 c o TJ OJ u a \\ 4J J-l o u a A ICBC-UBC • Transport Canada O Strother et al V-Impact speed (mph) Figure 4.2 : Residual plot of ICBC-UBC, Transport Canada, and Strother et al data ^ 0.4 -n 0 .35 -0.3 H E r r o r o f f i t / p r e d i c t i o n (mph) 4.3 : F r e q u e n c y d i s t r i b u t i o n o f e r r o r s o f f i t / p r e d . (Honda C i v i c s ) 48 Table 4.1. Figure 4.4 is a plot of impact speed and residual crush of 1971-1972 full size CM. vehicles from Campbell. A regression line is drawn and the ICBC-UBC impact result of the CM. Cutlass is also plotted. The ICBC-UBC data point lies outside of the speed range of the data, however assuming the estimated relationship is valid to a crush level of four inches then the closeness of the data point to the line indicates that it is also valid. The error of fit of the Campbell data and the error of prediction of the ICBC-UBC value is plotted in Figure 4.5. At a crush level of 3.8 inches, the estimated relationship over-predicts speed by 0.8 mph. The distribution of the errors of fit and the Normal distribution are shown in Figure 4.6. The ICBC-UBC result is 0.4 standard deviations from the expected impact speed. Table 4.1 : Errors of fit/prediction DATA SOURCE Transp. Canada Transp. Canada Transp. Canada Transp. Canada Transp. Canada Strother Strother Strother Strother Strother ICBC-UBC ICBC-UBC ICBC-UBC ICBC-UBC IMPACT ERROR OF STANDARD SPEED FIT OR DEVIATION (mph) PREDICTION (mph) 28.7 1.5 29.7 -0.4 29.2 4.9 34.9 2.7 29.3 -2.4 8.9 -0.6 13.9 -0.8 27.1 -2.1 27.1 -3.7 40.0 0.8 19.7 0.3 0.1 19.2 0.6 0.2 18.3 0.7 0.3 17.3 1.2 0.5 C-Residual crush (inches) Figure 4.4 : Crash test results of 1971-1972 f u l l size G.M. cars • ICBC-UBC Q Campbell rP. -TJB-• • • I 1 1 I 2 0 4 0 6 0 Predicted speed (mph) Figure 4.5 : Residual plot of ICBC-UBC and Campbell data (1971-1972 f u l l size G.M. cars) o Error of fit/prediction (mph) Figure 4.6 : Frequency distribution of errors of fit/pred. (Full size G.M. cars) 52 Repeatability is the ability of the testing facility to reproduce the same damage for the same test variables. To check the repeatability of the test facility, the Honda Civics were tested at impact speeds within a narrow range of 17 to 20 mph, as shown in Figure 4.1. The slight differences in residual crush is consistent with the slight differences in impact speeds. These tests confirm the repeatabilty of the test facility. Table 4.2 contains the Honda test data. TABLE 4.2 : Honda test data Impact Speed Residual Crush (mph) (inches) 19.7 19.2 18.3 17.3 10.0 9.5 8.8 7.8 The speed and crush data from the ICBC-UBC test facility are consistent with predicted values of Transport Canada and Strother data. Thus, the ICBC-UBC test facility produces results that can be compared with other test facilities. 5. H IGH SPEED C R A S H DATA F R O M REPEATED IMPACTS 5.1. I N T R O D U C T I O N There is a scarcity of barrier crash data at either low and very high speeds. Multiple low speed impacts performed on the same vehicle may be equivalent to a single higher speed impact. If a equivalence exists then relatively inexpensive low speed crash tests can be performed to generate high speed crash data. This chapter will quantify the speed-residual crush relationship of repeated low speed impacts and attempt to verify the the relationship by conducting a series of repeated low speed crash tests. Conducting high speed testing in a 40 to 60 mph speed range is a more costly undertaking than low speed testing. First, a testing facility to accomodate high speed crashes requires substantially larger components and systems. The barrier must be considerably more massive, usually requiring a subgrade foundation and a solid reinforced concrete block. A more powerful and sophisticated propulsion system is required or alternatively a longer approach and guidance track is required. Producing the damage pattern of a high speed impact from repeated low speed impacts has the added benefit of providing crash data over a range of speeds from a single vehicle. After each low speed collision, the progressive crush and equivalent barrier speed gives a new data value further along on the speed-residual crush curve. Three or four speed-residual crush values obtainable from a single vehicle drastically reduces the cost of test vehicles. 53 54 A high speed crash derived from repeated low speed crashes may be applicable when damage pattern and impact speed is the only information of interest. Obviously safety information such as occupant dynamics and deceleration histories during a high speed collision cannot be derived from repeated impacts. 5 . 2 . CAMPBELL'S DERIVATION O F ENERGY A B S O R B E D IN RESIDUAL C R U S H The following list defines the variables used in this chapter. C = Plastic deformation (referred to as residual crush) a 0 = Force per unit width to initiate residual crushing a x = The increase in force per unit width with residual crush f = Force per unit width W 1 = Width of residual crush W 0 = Vehicle width b„ = Impact speed to initiate residual crushing bi = Slope of impact speed-residual crush line V = Impact speed E = Absorbed energy Equivalent speed: the vehicle speed which has a kinetic energy which equals the total kinetic energy of the multiple impacts. 55 Campbell's pioneering work in quantifying the energy absorbed in residual crush provides the foundation to derive an equivalence between multiple low speed impacts and a single high speed impact (Campbell,1974). Campbell's crush energy model is based on residual crush, or plastic crush, which is the crush remaining after the impact. In this model, force per unit width is assumed to be linearly related to residual crush as depicted in Figure 5.1. f • a, + a xC (1) The parameters a 0 and a x define the force-residual crush response where a 0 is the force at which residual crushing begins and a x is the increase in force with plastic deformation. The force response cannot be fully defined in terms of residual crush alone and as such it is not an entirely correct representation of the force response during plastic deformation. However, the force response of equation (1) provides an effective relationship of the energy absorbed in residual crush. Appendix IV describes a more accurate force response in terms of elastic and plastic crush which would more closely emulate the true force response. Integrating the force per unit width (equation (1) ) with respect to residual crush and width gives the energy absorbed in residual crush. E = J / (ao+axOdC dw + constant a 0 0 The constant represents the energy stored in the initial elastic range. This is the fundamental model for crush energy adopted by accident reconstructionists and has been incorporated into the CRASH3 program. The accuracy of this model has been tested extensively and has shown to yield acceptable results when correct stiffness i 1 r R e s i d u a l crush Figure 5.1 : Force-residual crush response 57 parameters are specified (Woolley et al,1986). This model represents the current 'state-of-the-art' in determining the absorbed energy from vehicle damage. ln Figure 5.2 is the measured force response in terms of elastic and plastic crush of Torinos (McHenry,1987). Force is shown to be approximated by a linear function in both elastic and plastic crushing. This indicates that a linear force response chosen by Campbell may be a valid assumption in his crush energy model. Campbell derived the a 0 and a x parameters as follows. A linear relationship was selected between impact speed (V) and residual crush (C). V = b 0 + b xC ( 2 ) A linear relationship is assumed for the force per unit width (f) of the front structure as a function of residual crush. f = a 0 + a xC (3 ) In a barrier impact all the vehicle's kinetic energy is absorbed in crush of the vehicle's front structure. The integral of the force per unit width with respect to crush and width is the energy absorbed in residual crush, and the kinetic energy at impact is fmV2. Equating the vehicle's kinetic energy at impact to absorbed energy in crush gives: W. C x £mV 2 = ; / f dC dw + constant 0 0 Substituting equations (2) and (3) gives: Crush(inches) Source : McHenry, R.R. McHenry, B.C. 'A Revised Damage Analysis Procedure for the CRASH3 Computer Program'. Figure 5.2 : Measured force-crush responses 59 W0 C, tmCbo+biC) J= / / (a 0+ a iC)dC dw + constant (4) 0 0 The constant term is required to account for some initial energy absorbed in elastic deformation with no residual crush. In equation (2) an impact speed of b 0 is required to initiate residual crushing of the vehicle. This impact speed is the kinetic energy absorbed in the initial elastic deformation. The elastic deformation is comprised of deflection in the bumper system and elastic bending of the front structure. When the input energy exceeds the energy absorbed by elastic deformation, the front structure undergoes plastic deformation. The parameters a 0 and a 1 can be related to the coefficients b 0 and b x by integrating equation (4) and solving by comparison of like terms. a = S ^ (5) mb? a, - ( 6 , constant=tmbo (7) Thus, the force per unit width can be expressed in terms of the coefficients b 0 and b L which are derived from full frontal barrier collisions. The b 0 and bx coefficients are unique to each vehicle model or category of vehicles. f = a„ + a i C = J (b0bx+ bjc) (8) \"0 The integral of equation (8) with respect to residual crush and width gives the 60 absorbed energy + mb?C W0 )dC dW + tmbo where Cj is the residual crush and Wx is the width of residual crush. This general equation allows the absorbed energy to be determined for different damage patterns. The major assumptions of this model are: force is linearly related to residual crush and the stiffness is constant across the front of the vehicle. Also the residual crush must be across the full height of the vehicle's front structure for the model to be applicable. 5.3. ABSORBED ENERGY IN CRUSH FROM MULTIPLE IMPACTS As shown previously the key parameters a 0 and at of the crush energy model are expressed in terms of the coefficients b„ and b x which are derived from full frontal barrier impacts. The b„ coefficient is the impact speed intercept and the b x coefficient is the slope of the impact impact speed-residual crush line. These coefficients are derived from a least squares fit of a linear model to impact speed-residual crush data. To derive the coefficients of a specific vehicle model or category of vehicles, many vehicles must be crash tested to establish the impact speed-residual crush function. In Figure 5.3 the impact speed-residual crush data of 1971-1972 G.M. vehicles are plotted which exemplifies a linear relationship between impact speed and residual crush (Campbell,1974). If this same function can be established by repeatedly impacting the same vehicle then the cost for test vehicles Source : \"Energy basis for collision severity\", Campbell,K.L, 1974. Figure 5.3 : Impact speed-residual crush curve 1971-1972 CM. vehicles 62 can be reduced. The impact speed-residual crush function of Figure 5.3 was established by crash testing many vehicles once only. It may be possible that one vehicle repeatedly impacted will produce impact speed-residual crush data that follows the function derived from single impacts on vehicles. The b 0 and b x coefficient could then be determined for a particular vehicle model by conducting multiple impacts on as few as two vehicles. A postulated force-residual crush response of a repeatedly impacted vehicle will be presented which will produce the same linear impact speed-residual crush function of singly impacted vehicles. The assumptions inherent in this model will help in identifying possible reasons for deviations from the impact speed-residual crush function of single impacts. 5.3.1. Force-residual crush response of multiple impacts Figure 5.4 depicts the postulated force-residual crush response of multiple impacts. The energy absorbed in the initial elastic crush plus area 1 represents the energy absorbed in the first impact. The energy absorbed in the second impact and third impact is represented by areas 2 and 3 respectively. The total energy absorbed in the three repeated impacts is the initial elastic crush of the first impact plus areas 1, 2 and 3. Thus for the second impact, the same force-residual crush response is assumed to continue beyond the residual crush, d , of the first impact. The force-residual crush response models the energy absorbed during plastic crushing. For a single impact, the energy absorbed in the elastic range is the Figure 5.4 : Crush energy of a vehicle repeatedly impacted OJ 64 energy required to initiate residual crushing or plastic deformation. This energy is accounted for by the constant term in equation (1). In the multiple impact model the same force-residual crush response is assumed, however for the second and subsequent impacts the energy absorbed in the initial elastic crush is neglected. The validity of this model and the associated assumptions is tested with experimental data. The next section will derive the impact speed-residual crush relationship that is generated from the force-residual crush response of Figure 5.4. 5.3.2. Quantification of crush energy from multiple impacts Using Campbell's force-crush model a relationship for crush energy will be derived in terms of the residual crush from the first impact, C x, and the residual crush from the second impact, C2-The energy absorbed in two full frontal impacts on the same vehicle is E - J / ( a 0+ a iC)dC dw + tmbo (9) 3 0 0 where Ci+Cj is the total residual crush from both impacts and ^ mbj is the energy absorbed in the elastic range of the first impact. Substituting equations (5) and (6) for a 0 and a x gives W0 Ci+Cj jab b jfoiQ , E a - / J ( - ^ + )dC dW + *mb0 0 0 \"0 \"0 Integrating the above equation. mb?(C,+C,) 2 2 a 2 65 It can be shown that the relationship between the equivalent speed (E ) and cumulative residual crush is the same as the impact speed-residual crush function of single impacts. Equating the total absorbed energy of both impacts to the equivalent kinetic energy of both impacts (£mV^) yields mb?(C,+C,)2 , tmv^ = ttb.MCx+0 + I V l2 l ) + tmb, After some rearranging, this reduces to V e q = b 0 + MCj+C,) ( 1 0 ) The above equation can be generalized as follows n V = b e + b x I C ( 1 1 ) e q i = l 1 where n is the number of impacts and is the residual crush of the ith impact. In this model, the total residual crush from multiple impacts on a vehicle has an equivalent speed equal to the impact speed of a single impact with the same residual crush. The cumulative residual crush and equivalent speed from repeated crashes should lie on the impact speed-residual crush line of full frontal barrier impacts. Multiple impacts on the vehicle will give a trace of the impact speed-residual crush function from which the crush characteristics or parameters b 0 and b x can be determined. The underlying assumption of the multiple impact theory is the total energy absorbed in residual crush is the same regardless of 66 whether that residual crush level is produced in one impact or many impacts. For this major assumption to be valid, a number of other assumptions were made relating to the force-residual crush behaviour of the vehicle. The above force-residual crush model and assumptions will help in identifying possible reasons for deviations of the multiple impact results from the speed-crush function. The assumptions made in the model are: 1. The energy absorbed in the elastic crush for the second and subsequent impacts is negligible 2. The force versus residual crush is a linear relationship. If the first assumption is not valid and significant energy is absorbed in the elastic crush of the second and subsequent impacts then the multiple impact data may show a significant deviation from the impacts speed-residual crush line of single impacts. This deviation is depicted in Figure 5.5 where in multiple impacts a higher equivalent speed is required to produce the same crush as a single impact. This deviation results from the greater quantity of elastic energy absorbed from repeated impacts. Experimental tests on vehicles were conducted to determine whether the equivalent barrier speed versus cumulative crush will trace the impact speed-residual crush line of single impacts and provide the same b„ and bx coefficients. The next section presents the result of the tests. m u l t i p l e impacts R e s i d u a l crush Figure 5.5 : Possible deviation of mu.tip.e impact data from singie impact data 68 5.4. R E P E A T E D C R A S H T E S T S To verify the relationship between the equivalent speed and cumulative crush from repeated impacts, repeated crash tests were conducted on a 1977 and 1974 Honda Civic, and a 1971 CM. Cutlass station wagon. All of the vehicles underwent two full frontal barrier impacts test; the results of these tests are given in Table 5.1. Table 5.1 : Double impact results MAKE/MODEL IMPACT No. ACTUAL RESIDUAL IMPACT CRUSH SPEED(mph) (in) 1977 Honda Civic 1 19.2 9.5 2 18.8 4.3 1974 Honda Civic 1 18.3 8.8 2 15.2 4.6 CM. Cutlass SA/V 1 9.4 3.8 2 14.5 10.0 The impact speed and residual crush for the first impact is as shown in the table. For two impacts, the equivalent speed is the speed such that the kinetic energy is equal to the total kinetic energy of both impacts. Thus, the equivalent speed of the two impacts is calculated as follows: eq where V x is the impact speed in the first impact and V 2 is the impact speed in the second test. This reduces to v „ = u o X Residual crush Figure 6.3 : Two regime force-crush model Figure 6.4 : Higher order force-residual crush function 00 O \\ 63 o CM U JC in D L i O a, / (a , ) a 0/(aj) R e s i d u a l c r u s h Source : \"Crush Energy i n A c c i d e n t R e c o n s t r u c t i o n \" . S t r o t h e r et a l (1986), SAE 860371 Fi g u r e 6.5 : Two regime energy p l o t 00 83 less than Cj. However, at a crush level of C j , the two models converge and the energy absorption characteristics are the same. The advantage of a two regime model is more accuracy at speeds just above the threshold speed. Since a high percentage of collisions occur at these very low speeds, there is a definite need for this improvement. Also a single regime model cannot effectively replicate the energy absorption characteristics over the whole speed spectrum. Adaptation of a single regime model to the low speed range by making large shifts in the b 0 and b x parameters can adversely affect crush energy characteristics at higher speeds. Quality of prediction is sacrificed at medium speeds for better prediction at lower speeds. On the other hand adding a third parameter to the model substantially increases complexity. 6.3. CRUSH ENERGY OF THE TWO REGIME MODEL The parameters a 2 and a 3 in the two regime model can be expressed in terms of the coefficient b 2 and b 3 . _ mbiba _ mxb Bb 3 a 2 = = W Q W Q mbj a ' = w0 where x=^*- . The solution for these two parameters are shown in Appendix V. b 0 The force per unit width in the low speed regime is: f _ nixboba + mb2C W#) WQ In frontal collisions, the energy absorbed in crush at crush levels less than 84 Cj is: W, C E = J / (a2 + a3C)dC dw + $m bf a 0 0 For crush levels exceeding Cj, the crush energy is: W° Cj E = J / (a2+a3C)dC dw + tmbf a o 0 W0 C + / / (ao+axOdC dw ° C I The b 2 coefficient or threshold speed is easily determined from bumper tests. With a threshold speed, the b 3 coefficient or slope can be established from one valid crash test at a speed just above the threshold speed. Alternatively a lower limit of the medium speed regime can be assumed and the intersection of the low speed regime with this lower limit will establish the b 3 coefficient. 6 . 4 . THREE REGIME MODEL OF VEHICLE CRUSH CHARACTERISTICS A paper by Strother et al(1986) on crush energy in accident reconstruction suggests a two regime model in which the force per unit width reaches a saturation level. Such a model has been identified for 1979 to 1982 G .M. Citations wherein the high speed crash data indicates a flattening of the impact speed-residual crush function at crush levels above 22.5 inches. Figure 6.6 shows the two regime speed-crush function for G .M. Citations. Combining the high speed regime suggested by Strother et al(1986) and 85 C-RESIDUAL CRUSH (in.) A Strother et ol.( 1986 ), Figure 5 • Hight «t o l . (1985) , Figure 4 * Transport Conodo data Source : \"CRASH3 and Canadian data\". Navin, F (1986) Figure 6.6 : Impact speed vs Residual crush for f u l l frontal barrier tests for G.M. Citations 1979-1982 86 the low speed regime proposed in this research produces a three regime model, ln this model the vehicle exhibits three different crush characteristics which are obeyed for a specific crush range. Figure 6.7 depicts the impact speed-residual crush function of the three regime model. Regime II reflects the crush characteristics of Campbell's original single regime model. A two regime force response is a proposed model of the energy absorption characteristics at low residual crush levels. Other factors may also significantly affect the amount of residual crush and energy absorbed. At low impact speeds, the energy level is relatively low and as a result the vehicle rotation or pitch during impact may represent a significant quantity of energy that is not absorbed in residual crushing. Residual crush Figure 6.7 : Three regime speed-residual crush model 03 VI 7. APPLICATIONS TO THE CRASH3 PROGRAM CRASH 3 is a widely used computer program for reconstruction of vehicle accidents. Contained within it are two methods for determination of accident speeds: reconstruction based on the damage and reconstruction based on the vehicle trajectories and conservation of linear momentum. Reconstruction on a damage basis has been criticized for its large errors in speed prediction due partly to the lack of vehicle specific crush characteristics. In addition, the predictive capability in a low and high speed range has been challenged due to the lack of crash data in these ranges. The CRASH3 program categorizes vehicles according to wheelbase because of a lack of vehicle specific crush characteristics. All vehicles within a category are assumed to have the crush characteristics of the overall group. However, large errors have been demonstrated and attributed to this categorization (Woolley et al,1986). Some vehicles exhibit markedly different crush characteristics than the characteristics assigned to its size category which results in large speed prediction errors. As a result it has been suggested that vehicles be further stratified by their relative stiffness (Navin,1986). Within each size category, vehicles would be assigned to stiff, normal and soft categories. One step further would be to determine the crush characteristics of each vehicle model, however the data requirements and cost may not be commensurate with the improved accuracy. A multiple impact technique, as presented in Chapter 5 shows much promise as a method of producing inexpensive crash data. This technique can be 88 89 of great benefit for producing the data requirements to further stratify vehicle crush characteristics according to relative stiffness. Presently, the CRASH3 program predicts accident speeds on a damage basis with a single regime force-residual crush model. Crash data from tests in a medium speed range are used to define the crush characteristics (a„ and ax parameters), however these characteristics may not be valid at low speeds and as a consequence the accuracy of the CRASH3 program can be challenged in the low speed range. Preliminary tests at the ICBC-UBC facility indicates that there is a definable low speed regime. The two regime model for the crush characteristics presented in Chapter 6 will be utilized to demonstrate its application to the CRASH3 program. Using the test data on the Honda Civics from Strother et al, Transport Canada and ICBC-UBC, the parameters of a two regime force-residual crush function will be derived, and an equivalent barrier speed will be predicted for a Honda Civic involved in an accident. This prediction will be compared to the equivalent barrier speed prediction from the CRASH3 program. The equivalent barrier speed is defined by Campbell as the \" vehicle velocity at which the kinetic energy of the vehicle would equal the energy which was absorbed in plastic deformation.\" Figure 7.1 shows the two regime speed-residual crush model for 1975-1981 Honda Civics. A least squares fit to the data determined regime II. The ICBC-UBC data suggests that a residual crush of 8 inches is the lower limit to regime II. The y-intercept or threshold speed of regime I was determined from crash tests. The y-intercept and slope of regime I and regime II is tabulated in Table 7.1. 91 Table 7.1 : Speed-residual crush coefficients Regime I Y-intercept (b 2) 10.8 mph Slope (bj) 0.7 mph/in Regime II Y-intercept (b 0) 4.4 mph Slope (b x) 1.5 mph/in The parameters for the two regime force-residual crush function were calculated using equations (1) to (4) in Appendix V which are tabulated in Table 7.2. The mass (m) is 5.7 lb*sec 2/in and the vehicle width is 58.5 inches. Figure 7.2 shows the force-residual crush function of 1975-1981 Honda Civics. Table 7.2 : Force-residual crush parameters Regime I Y-intercept (a 2) 227.8 lb/in Slope (a 3) 14.7 Ib/in*in Regime II Y-intercept (a„) 199.1 lb/in Slope ( a i) 67.9 lb/in*in In Figure 7.3 is the residual crush profile of a Honda Civic involved in a frontal collision with a rigid immovable object such as a bridge abutement. The six residual crush dimensions in inches are C^ O.O, C2=2.3, C3=6.2, C4=8.0, C5=9.3, C,=14.8. Using a trapezoidal approximation to the area within the profile, the residual crush for the first interval is (C1+c2)/2. Since the crush for this interval lies within the range of regime I, the force-crush parameters are a 2 and a 3. The energy of deformation is: Figure 7 . 3 : Damage pattern is U ) 94 E = / / (a3+a3C)dC dw + ^ o o 2a 3 r W / „ a3C2 _ aJW = / (a,C + ~Z— )dw +-*— o 2 2a, where W is the width of residual crush. Substituting the crush of interval 1 ((Cx+C,)/2 ) for C gives E = ;W l( a 2(C 1 +C 7,) + a ^ C ^ C , * Cf) +a|W1. c o 2 8 2a3 where W x is the width of residual crush interval 1. Integrating over the width of interval 1 gives: . ( a.CC.+C,)^ a3(C?+2C,C,+ Cj) +a]j^ c 2 2 1 2a3 E = 22081 lb*in c These calculations are repeated for intervals 2 and 3. For intervals 4 and 5 which has crush levels in regime II, the a 0 and a x parameters are utilized for calculating the energy of deformation. Table 7.3 summarizes the crush energy for each interval. Table 7.3 : Energy of deformation across the residual crush profile Interval Residual Width of Crush Crush (in) Crush (in) Energy (lb • in) 1 1.2 10 22081 2 4.3 10 29581 3 7.1 10 39291 4 8.7 10 45539 5 12.1 10 76209 Total - W= 50 EN=212700 95 The crush energy in Table 7.3 assumes the forces producing the deformation acts normal to the edge of the vehicle. If the principle direction of force is inclined to the edge of the vehicle the following correction factor must be applied to (Smith, R.A. and Noga, J.T., 1982). Correction factor = (1 + tan Ja) where a is indicated on Figure 7.3. Thus, the corrected energy of deformation is: E • (1 + tan*a)E N = (1 + tan*(15°))(212700) = 227972 lb-in The line of action of the force vector does not pass through the center of mass. The non-central impact factor (7) used is 0.97. This factor is determined by the relationship where K is the radius of gyration about a vertical axis through the center of mass and H is the moment arm of the line-of-action of the average force about a vertical axis through the center of mass (Smith, R.A. and Noga, J.T., 1982). The delta-V for a vehicle in a two-vehicle crash is computed in the CRASH3 algorithm by the relation: 7 = K * / ( K * + H 2 ) 96 Since the Honda Civic struck a rigid immovable object, the energy absorbed by the object (E 2) is zero and its mass (m 2) can be considered infinite. Therefore delta-V for the vehicle is: where E x is the energy absorbed by the vehicle in crush. Delta-V for the Honda AV X = 279 i n / s e c • 15.8 mph Using the CRASH3 computer program, the delta-V for the Honda Civic is: AV X = 16.1 mph The CRASH 3 computation of delta-V utilizes a single regime force-residual crush model with average stiffness parameters of vehicles in the \"micro\" category. There is a small difference in the delta-V prediction which is largely due to the difference in crush coefficients utilized by CRASH3 and those derived from Honda Civic crash tests. When the same calculations are repeated using a single regime model with the Honda Civic crush coefficients, the delta-V is 14.5 mph. This is a 1.3 mph difference between the two regime model and single regime model. The Honda Civic crush is a hypothetical accident case created to exemplify the application of the two regime model in computing delta-V. Since the true delta-V is not known, the accuracy of this model cannot be judged. Full scale staged collisions should be conducted to determine the accuracy of the speed Civic is: 97 predictions. 8. CONCLUSION A very inexpensive and reliable low speed crash test facility has been built which has proven successful for controlled crash tests. The test results have been verified indicating the facility does replicate results from other barriers. At present, the facility has the following capabilities: •The tow system can accelerate the test vehicle to 30 km/h. •the crash barrier has been designed for an impact of an 1800 kg car at 40 km/h. •The speed trap provides a speed measurement with an accuracy of ± 1 % . •Impact speed can be controlled to within ± 2 km/h. •Both full frontal and full rear impacts can be performed on vehicles. •Data derived from each test is impact speed and residual crush. At present the facility does not have a high speed camera for making dynamic crush measurements. •A high test rate of one crash test per hour can easily be achieved. There has not been any known experimental investigations of repeated 98 99 impact tests on the same vehicle. Repeated impact tests were conducted to verify this technique of generating impact speed-residual crush data. The findings from these test are: •A second low speed impact on the same vehicle produces speed-crush data (ie. crush characteristics) that is comparable to a single equivalent higher speed impact. •The standard deviation of the errors of prediction indicate that the double impact results are equivalent to a single higher speed impact. Low speed crash tests were conducted to determine the speed to initiate residual crush of vehicles with an energy absorbing bumper system. The following conclusions were drawn. •The threshold speed of Honda Civics with an energy absorbing bumper system is much higher than 8 km/h (5 mph). This is an indication that vehicles with an energy absorbing bumper system (ie. isolators) will exhibit threshold speeds above 8 km/h. •A least squares fit of a linear model to speed-crush data in a medium speed range (48-55 km/h) may not be valid at speeds below 20 to 25 km/h. •At the low end of the speed spectrum, a second speed-crush regime exists for specific vehicle models. The crush characteristics at low crush levels is different than at higher crush levels. •The accuracy of delta-V predictions in low speed collisions may be improved when a two regime model is utilized. 9. FURTHER AND FUTURE RESEARCH The crash testing facility opens up many areas for research. There is still further research to be pursued on the topics covered in this thesis, as well as other areas of future research on low speed collisions. The research interests of many different organizations can be pursued through low speed crash testing. The University community, insurance industry, accident investigators, and small automobile manufacturers can benefit from the test facility. In this chapter, further research on the topics covered in this thesis will be identified and the future research interests of ICBC and UBC Accident Research Group will be outlined. In addition, the research interests of private firms will be outlined. 9.1. FURTHER RESEARCH Further tests should be conducted into full frontal repeated impacts on the same vehicle. A comprehensive testing program should be established to complement the tests already completed. In particular repeated impact tests should be conducted on a vehicle of each size category as defined by the CRASH 3 program. The number of impacts and impact speed of each vehicle tested should provide equivalent speeds that cover the whole speed spectrum. The equivalent speed has been defined as n x V = ( Z V?) 2 where is the impact speed of the impact. For example, four impacts each at an impact speed of 32 km/h (20mph) is equivalent to four impacts at impact 101 102 speeds of 32 km/h, 45 km/h, 55 km/h and 64 km/h (20 mph, 28 mph ,34 mph and 40 mph). This program of tests is needed to further validate the multiple impact technique. More crash tests are needed to determine the crush characteristics in the low speed regime for different vehicle models. The threshold speed for vehicles of different mass and with different bumper systems should be established. The slope of the speed-crush function for different vehicle models is also needed, ln addition, the accuracy of prediction over the low speed regime should be evaluated. At low impact speeds, the kinetic energy and crush energy is relatively low and consequently some of the assumptions may produce significant errors in prediction which would otherwise be insignificant at high impact speeds. For example, it is assumed that the force-crush characteristics does not vary across the vehicle width. In an offset low speed impact with an immovable object, a slight variation in stiffness across the vehicle width may have a significant effect on speed prediction. By setting up the ICBC-UBC facility for pole impacts, the variation in stiffness across the front of the vehicle can be tested. At low speeds more consistent speed predictions may be derived by considering both static and dynamic crush rather than only static or residual crush. By determining the percentage of total crush that is dynamic crush - for different vehicle models and at different crush levels - the total crush can be determined from the level of static crush. Prediction of equivalent barrier speeds can then be based on both static and dynamic crush. Crash tests recorded with a high speed camera would provide dynamic crush data from which the dynamic crush percentage as a function of the static crush level can be determined. 103 9 . 2 . F U T U R E R E S E A R C H The future research interests of ICBC and UBC Accident Research Group are: • Repair Strengths - Damage to repaired vehicles in a controlled barrier impact can be investigated to assess how the repaired area alters the energy absorption characteristics of the vehicle during collision. This has a direct bearing on the risk of injury to the occupants. • Repairability and Damagability - The repair costs of different vehicle makes and models can be evaluated to assess premiums for the different vehicles. • Side Impacts - In a low speed side impact the passenger compartment is subjected to some intrusion. Conducting side impact tests at low speeds can provide information on injury risks to occupants for different vehicles. • Effect of Corrosion on Structural Integrity - Corrosion degradation in structural frame rails of older vehicles can be investigated to assess the effects on vehicle stiffness. • Seat belt tests - Once a suitable dummy is obtained, tests will be undertaken on four point seat belts to assess their effectiveness on injury reduction. Three point seat belts will also be tested. 104 Private firms have expressed an interest in conducting test for the following research: • Accident reconstruction experts have a significant percentage of cases involving collisions at less than 5 mph. Consequently, they are interested in the elastic deflection of the energy absorbing bumper system and its relation to impact speed. • An automobile manufacturer has expressed a need for conducting 30 mph crash test on prototype vehicles to gain prior knowledge on their ability to pass Canadian Motor Vehicle Safety Standards. B I B L I O G R A P H Y Appleby, M.R. and Morris, A.G.R. \"Automobile Damageability and Insurance Costs\", SAE paper 740305, 1974. Bartee, E.M. Engineering Experimental Design Fundamentals. Prentice-Hall lnc.,Englewood Cliffs, N.J., 1968. Bigg, G.W. \"Uses and Abuses of Speed Change Results from Crush Analysis\", Proc. of the Canadian Multidisciplinary Road Safety Conference IV, May 26-28, 1985. Bleckert, H. and Mack, G. \"Linear Drive for Crash Test Facility\", Siemens Review, No. 7, 1977. Campbell, K.L. \"Energy Basis for Collision Severity\", SAE paper 740565, 1974. Franchini, E. \"Crash Testing Evolution at Fiat\", SAE paper 660165, 1966. Fredericks, R.H. \"Automobile Crash Research\", SAE paper S363, 1962. Guenther, D.A., Goodwin, L.G. and Bookwalter, J.C. \"Barrier Equivalent Impacts Speeds: An Analytical Review With Some Experimental Validation\", Proc. of the Canadian Multi-disciplinary Road Safey Conference III, May 27-30, 1984. Haynes, A.L., Fredericks, R.H. and Ruby, W.J. \"Automotive Collision Impact Phenomena\", Highway Research Board Bulletin V. 142, 1956. Hight, P.V. and Lent-Koop, D.B. \"Correlation Between Delta V, Barrier Equivalent Velocity and Crush in Automobile Collisions\", Proc. of the Canadian Multidisciplinary Road Safety Conference III, May 27-30, 1984. Jones,!.S. and Jennings, P.W. \"Development and Evaluation of the CRASH2 Program for Use Under European Conditions\", 105 106 SAE paper 810473, 1981. Lundstrom, LC. \"The Safety Factor in Automotive Design\", SAE paper 660539, 1967. Macmillan, R.H. \"Vehicle Impact Testing\", SAE paper 700404, 1970. Macaulay, M.A. \"Vehicle Frontal Barrier Impacts\", TRRL Report LR 481, 1972. McHenry, R.R. \"A Comparison of Results Obtained With Different Analytical Techniques for Reconstruction of Highway Accidents\", SAE paper 740565, 1974. McHenry, R.R. and McHenry, B.C. \"A Revised Damage Analysis Procedure for the CRASH Computer Program\", SAE paper 861894, 1987. Meriam, J.L. Dynamics. 2d ed. John Wiley, 1975. Miller, P.M., Ryder M.O., and Shoemaker, N.E. \"Crash Energy Management in Subcompact Automobiles\", SAE paper 740572, 1974. Myers,R.H. Classical and Modern Regression With Applications. Boston Massachusetts: Duxbury, 1986. Navin, F., Navin N., and MacNabb M. \"CRASH 111 and Canadian Data\", SAE paper 870499, 1987. O'Neill, B. \"Bumper Perfomance Level and Insurance Loss Experience\", SAE paper 840224, 1985. Scott, D. \"Versatile Crash Test Facility has Cable Drive\", Automotive Engineering, V.85 No. 7, July 1977. Severy, D.M. and Mathewson, J.H. \"Automobile-Barrier Impacts\", Highway Research Board Bulletin 91, 1953. 107 Severy, D.M. \"Automobile Collisions on Purpose\", Human Factors V. 2 No. 4, Nov. 1960. Sinke, R.A. and Prevost, T.C. \"An Automotive Crash Test Facility\", SAE paper 700527, 1970. Smith, R.A. and Noga, J.T. \"Accuracy and Sensitivy of Crash\", National Highway Traffic Safety Adminstration Technical Report No. DOT-HS-806152, March 1982. Snider, H.P. \"Vehicle Instrumentation for Crash Testing\", IEEE Transactions on Industrial Electronics and Control Instrumentation, V. IECI-11, 1964. Stonex, K.A. \"Single-Car Accident Problem\", SAE paper 811 A, 1965. Strother, C.E., Woolley, R.L., James, M.B. and Warner, CY. \"Crush Energy in Accident Reconstruction\" SAE paper 860371, 1986. Wilson, R.A. \"A Review of Vehicle Impact Testing: How it Began and What is Being Done\", SAE paper 700403, 1970. Woolley R.L., Warner, CY. and Tagg, M.D. \"Inaccuracies in the CRASH3 Program\",SAE paper 850255, 1986. A Review of Canadian Bumper Standards, Insurance Corporation of British Columbia, 1986. Research Trends, Cornell Aeronautical Laboratory Inc., Summer-Autumn 1972. Society of Automotive Engineers Handbook, \"Barrier Collision Tests\", SAE J850a, 1986. Foundation Engineering. 1974. \"Crash Testing at Fiat\", Engineering, Aug 1966 \"Controlled Vehicle Impacts-Instrumentation Test Procedure\", RRL Report LR92, 1967. APPENDIX I Crash Barrier Calculations 109 Crash Barrier Calculations 110 The barrier movement was calculated by considering the impact force of the vehicle and the resistive forces from the passive earth pressures and friction between the concrete blocks and foundation. During impact, the deceleration of the vehicle is assumed to increase to a peak then decrease to zero. The deceleration of the vehicle multiplied by its mass is the force applied to the barrier according to Newton's second law. The deceleration history is approximated by a triangular pulse with a peak deceleration of 40g as shown in Figure 1-1. The passive earth pressure generated by the gravel fill behind the barrier wall increases as the movement of the wall increases. This is shown in Figure I-2 where the ratio of horizontal to vertical stress (K) increases with horizontal movement. The friction force between between the concrete blocks and foundation is assumed to be fully developed with the slightest movement. Since the impact force and passive resistance are a function of time and barrier displacement, the net force, barrier velocity, and acceleration were calculated at time steps of 0.01 seconds. A net force is calculated which accelerates the barrier. Ti:e velocity and displacement of the barrier at the end of the time step is then calculated from the acceleration which is assumed to be constant over the time step. The displacement allows calculation of a new net force and subsequently acceleration in the next time step. These calculations are repeated until the barrier velocity reaches zero. Table 1-1 show the calculations of the barrier movement. 1 1 1 Figure I-3 shows a diagram of the barrier and a free body diagram. The friction force is: F. = M(Mass of blocks)a = 0.5 * 30 blocks * 1995 kg/block * 9.81 m/seca = 294 KN where F f = friction force between barrier and foundation u = coefficient of friction a g = gravitational acceleration The passive earth force is: F p - (7HiW + i-7H2W}K(d) = (18(0.74)*(1.5) + i(18)(1.6)*(1.5)}K(d) = 48-K(d) where W = width of gravel fill (m) j = unit weight of gravel fill (KN/mJ) H W H , = height of fill (m) K(d) = ratio of horizontal to vertical stress d = displacement (m) * 10.0 » 8.0 4.0 3.0 < u - 2.0 ' ec > o 2 O N CM: o x 1.0 0.8 0.6 0.5 0.4 0.3 0.2 < 0.1 0£ I-.1 — — 1 i i i i — r ^ • ^ ^ D E N S E -1 1 I _ H / • ^ L O O S E / • — 1 • \"••*\". . . // K P A C T I V E S T A T E T P A S S I V E S T A T E -— : L O O S E — ^ / J — > » v C O M P A C T — I D E N S E > ^ , L__ ! 1 l | 112 WALL ROTATION , — . H Source : Foundation E n g i n e e r i n g ,1974. Figure 1-2 : E f f e c t of wall movement on e a r t h pressure 70 60 -4 so H 40 -4 30 -J 20 H 10 H 20 40 60 80 100 120 Time a f t e r c o l l i s i o n onset (msec) F i g u r e 1-1 : T r i a n g u l a r d e c e l e r a t i o n p u l s e Elevation view Figure 1-3 : Crash barrier free body diagram 114 Table I - l : B a r r i e r c a l c u l a t i o n s TlK Tlas lncrcaeot . DUUoct Velocity lipact force I FujlTt resist Frlclioa lei force Accelerilloo l«c) (see) l«/J) IB] force (III force (•) |U) 0.00 0.0 0.0 0.01 lit 1.0 -46 -130 0 0.0 0.01 0.0 0.0 0.01 0.0 0.0 0.01 249 1.0 -46 -201 0 0.0 0.02 0.0 0.0 0.02 ao 0.0 0.01 392 1.0 -48 -294 50 0.14 0.03 4.202-05 0.00M 0.03 • 4.202-05 O.0OH 0.01 552 1.0 -48 -294 210 3.51 0.04 3.022-01 0.043 0.04 3.022-04 0.043 0.01 113 1.0 -41 -294 311 6.20 0.05 1.042-03 0.105 0.05 1.042-03 0.105 0.01 534 1.1 -53 -294 161 3.12 0.06 I.21E-03 0.135 0.06 1.212-03 0.135 0.01 261 1.1 -53 -294 -60 -1.34 0.01 2.502-03 ft 123 0.01 2.508-03 0.123 0.01 0 (.2 -56 -294 -352 -5 .M 0.06 3.402-03 0.064 0.1)9 3.402-03 0.064 0.01 0 1.2 -56 -294 -352 -5.U 0.09 3.152-03 0.005 0.09 3.152-03 0.005 0.01 0 1.2 -56 -294 -352 -S.U 0.10 3.152-03 0.00 Movement o f b a r r i e r - 0.0 '038 m » 3 . 8 mm « 0 . 1 5 i n c h e s S t a t i c c r u s h o f a v e h i c l e a t 40 k m / h r » 508 mm • 20 i n c h e s Movement o f b a r r i e r a s a % o f c r u s h « ( 0 . 1 5 / 2 0 ) * 100 - 0 .75% APPENDIX II Data Collection Form i 115 116 I C B C - U B C T E S T F A C I L I T Y Date: Ambient Temp. : Times Impact Ho.: VEHICLE IXFORHATIOK VEHICLE DESCRIPTION Venule Hake: VIM : Date of Kanufac.: Vehicle Hodel: Klleaf e: CRASH TEST Location of Impact i Front Q Rear Q Front end Rear end Total tt of previous lmnaets * of previous Impacts i t ICBC teat f a c i l i t y VEHICLE DIHENSIOHS Le f t Side Before A f t e r Front Overhang _ _ _ _ _ _ _ _ _ _ _ _ Wheel Base Rear Overhang _ _ _ _ _ _ Front Rear width . Track Width _ _ _ _ _ _ _ Frame Type: unlbody ladder frame Enflne Displacement: Test Weight: Hanufac. Weight: Right Side Before A f t e r 117 SPEED HEASUREHENTS Speed T r a p V i d e o Replay ftverall SDeed : F i r s t M a r k e r Soaclna:: F i r s t Time I n t e r v a l : Second M a r k e r S o a c l n f : Second Time I n t e r v a l : F i r s t Time I n t e r v a l : Difference l n T i m e : second Time I n t e r v a l : Other Hethod S p e c i f y : Speed Measurement: COMMENTS 11 8 APPENDIX III Recommended Safety Procedures 119 120 RECOMMENDED SAFETY PROCEDURES FOR CRASH TESTING To ensure the safety of all on-site personnel during testing, the following safety procedures and equipment are recommended. Test Vehicle Preparation •Completely drain fuel tank, lines and carburator of gasoline •Remove battery •Put maximum allowable air pressure in test vehicle tires Site Preparation •Close and lock all gates and garage doors during testing to prevent people wandering onto the test area. •Line both sides of the approach road to the barrier with guard rails to prevent runaway test vehicles from leaving the site. •Erect appropriate fencing around the tow road to keep unauthorized people away from the test area. •A guide rail for directional control of the test vehicle is strongly recommended. Equipment Check •Check that the gap between the rollers of the main release is 121 not less than 1.1 cm (0.45 in). •Check that the rollers are clean of debris, free to rotate, and well greased. •Check that the redundant release has the correct shear pin (Only the designated brass shear pin is to be used) •Inspect the entire length of the tow cable prior to each test for frays, kinks and/or broken strands. •Inspect the wire rope clamps to ensure they are firmly in place. •Do not proceed with testing if there are deficiencies in any of the above. •The cab of the tow vehicle should be protected with a wire cage. Personnel Safety •During testing all personnel except the tow vehicle driver and test controller are to stay at least 15 m (50 feet) back from the tow lanes and barrier. •The driver of the tow vehicle shall wear an approved motorcycle helmet and a properly adjusted 3 point seat belt while operating the tow vehicle. •Heavy gloves on personnel handling test vehicles especially after 1 2 2 barrier impact. •No smoking around test vehicles. Test Procedure •During the testing only two personnel are to be in the test area: the tow vehicle driver and the test controller. All other personnel are to stay out of the designated test area. •Radio communication shall be set up between the driver of the tow vehicle and the test controller. •The test controller is responsible for ensuring that all personnel are in the safe location during testing. •The tow vehicle driver shall watch for unauthorized people entering the site behind the the barrier just prior to testing. •The test controller shall stand a safe distance to one side of the approach and tow roads. •The tow vehicle driver must wait for clearance from the test controller before commencing the test. •Once the test has commenced there is only limited abort capabilities. Should the test controller notice an unsafe situation he must immediately signal the tow vehicle driver to halt acceleration. APPENDIX IV Force response during elastic and plastic deformation 123 124 Force response during elastic and plastic deformation McHenry has suggested that a force response, in the initial elastic range and in plastic deformation, would be more accurately emulated with a zero force intercept as shown in Figure 1V-1. Elastic and plastic crush Figure IV-1 : Force response for elastic and plastic crushing In this force response, the applied force increases linearly from zero as the vehicle deforms elastically. The force continues to increase linearly as the vehicle deforms plastically. At maximum elastic and plastic crushing, the vehicle's velocity is zero for a barrier impact. This the approach period of the collision and the corresponding velocity change is called the approach velocity change. Once maximum crush is attained, the vehicle begins to rebound away from the barrier. 125 The applied force drops sharply as the elastic crush is restored. The restoration of elastic crush produces a rebound velocity change. The total change in velocity is the sum of the approach velocity change and rebound velocity change. During the approach period, the energy stored and absorbed in elastic and plastic crush is the area under the force response from zero to maximum elastic and plastic crush. This force response proposed by McHenry is more realistic than Campbell's force response based on residual crush, however it is less practical since the maximum elastic and plastic crush is not easily obtained. Following an automobile collision, residual crush is the only measurement available. A force response in terms of residual crush provides a practical and robust model of the energy absorbed per unit of residual crush. Wx Cx E = J J (a 0+axC)dC dw + constant a 0 0 McHenry pointed out that Campbell's crush energy model neglects the restored energy during the rebound period and only considers the energy absorbed and stored during the approach period. Consequently, the approach velocity change is determined in this model. A P P E N D I X V Solutions for Crush Parameters b Solutions for Crush Parameters a 126 127 Solution for the linear extension of the force-crush function The area under the linear extension of the force-crush function can be shown to represent the energy absorption capacity of the bumper system or the energy absorbed in the elastic range before permanent crushing occurs. First, the C dimension in the above figure will be established as follows: e C = ^ e a, The parameters a„ and a i are equal to: ^ _ mbub! ^ _ mbf W o Wp which can be substituted into the above expression. The area under the linear extension of the force-crush function E b - frCea0 Substituting for C g and a 0 gives: This is the energy absorbed per unit width in the elastic range. 129 Solution for crush parameters a c , a,, a 7 > a 3 Solution for Regime I parameters: W0 C t m V*K„= ' / (f)dC dw W„ C j-mCbj+bjC) = J J (aj, + a3C)dC dw + constant tmbl+mbjbjC + i-mbJC1 = a2W0C + JajC'W,, + constant Comparison of like terms gives: constant = fmb? mb2bi W. a, = (1) mbj, 130 (2) Solution for Regime II parameters: W„ C i-mv* = ; ; (f)dc dw 0 0 w0 c imCV.+biCC-C.))'- / / (f_+ axC)dC dw + E. 0 o i-mVj+mV.bxCC-C.)* imb?(C-C.)2 = f_W0(C-C_)+ta1W0 (C-C_)2 + E, Comparison of like terms gives: E.-i-mVj 1 ^ 0 mb? The intersection values f ^ and V_ are equal to: f_ = a, + axC_ V . = b 0 + bxCj which can be substituted into equation (3). Solving for a„ gives: a. = (5) "@en ; edm:hasType "Thesis/Dissertation"@en ; edm:isShownAt "10.14288/1.0062702"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Civil Engineering"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Automobile crash test facility and preliminary analysis of low speed crush characteristics"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/26722"@en .