UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Application of work sampling to short cycle assembly operations Villalobos Casillas, Humberto 1978-02-27

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
831-UBC_1978_A7 V54.pdf [ 7.42MB ]
Metadata
JSON: 831-1.0081017.json
JSON-LD: 831-1.0081017-ld.json
RDF/XML (Pretty): 831-1.0081017-rdf.xml
RDF/JSON: 831-1.0081017-rdf.json
Turtle: 831-1.0081017-turtle.txt
N-Triples: 831-1.0081017-rdf-ntriples.txt
Original Record: 831-1.0081017-source.json
Full Text
831-1.0081017-fulltext.txt
Citation
831-1.0081017.ris

Full Text

APPLICATION OF WORK SAMPLING TO SHORT CYCLE ASSEMBLY OPERATIONS by HUMBERTO ^VILLALOBOS CASILLAS B.Eng., Instituto Politecnico Nacional, Mexico D.F., 1973 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in the Faculty of Graduate Studies in the Department of Mechanical Engineering We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 1978 ^^Humberto Villalobos Casillas, 1978 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 representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Humberto VILLALOBOS Casillas. Department of Mechanical Engineering The University of British Columbia 2075 Wesbrook Place Vancouver, B.C. Canada V6T 1W5 June 26, 1978. ii ABSTRACT The objective of this work is to study the applicability of work sampling to a short cycle repetitive operation and to compare the economics of work sampling technique with a conventional time study. To make this comparison, work sampling studies were carried out concurrently with ongoing time studies done by personnel of the Industrial Engineering department in a local lock factory. The criteria for the study were : 1. - The base data would ensure that the predictions would be within + five percent of the true value with 95% certainty. These are standard confidence limits for most industrial applications. 2. - Preparation time for work sampling and time study were assumed to be equal. Two operations were studied, a line assembly operation and a bench assembly operation. The first study represents the work of a team of four people as a unit. The workers often changed their activities to achieve a better balance of the line, since the nature of the operation is such that it is practically impossible to have the workers at fixed positions performing the same activity all the time. iii The second activity represents numerous workers working independently at benches. The study proved that for assembly operations work sampling can be used to good effect and at viable cost when compared with standard stop watch study. The acceptability of the studies psychologically favours work sampling, as does the level of training required for the time study practitioner. It is generally accepted that a work sampling analyst could be trained in one-quarter of the time which would be required for time study. Approved by : Prof. N. Eley iv TABLE OF CONTENTS Abstract ................................... ....... ii Table of Contents iv Acknowledgements ..................................vi CHAPTER I INTRODUCTION TO WORK MEASUREMENT ............... 1 1.1 Productivity •••........... ............ ......• 1 1.2 Management ratios .......................... 5 1.3 Direct means of improving productivity ..... 4 CHAPTER II TECHNIQUES OF WORK HEASUREHENT ................. 7 2.1 Terms used in work measurement ............. 7 2.2 Stop-watch time study 10 2.3 Representative samples of times 12 2.4 Rating 14 2.5 Determining the number of readings 15 2.6 Determining production standards ...........17 CHAPTER III WORK SAMPLING ..................... ...... ....... 18 3.1 Introduction ............................... 18 3.2 An example of sampling ..................... 19 3.3 Confidence level ...........................20 3.4 Accuracy of work sampling ..................22 3.5 Example of accuracy . .... ............. ...... 23 3.6 Accuracy for a given number of observations 24 3.7 Determining production standards ...........25 V CHAPTER IV APPLICATION OF COMPUTERS TO WORK SAMPLING ......28 4.1 Introduction ...... •. . • •. •.•• •.• • •....• ••...28 4.2 Description of the line .29 4.3 The study .... ................... ........... 35 4.4 Bating in work sampling ....................40 4.5 Schmid's computer programme ................42 4.6 Random times for observations ..............46 4.7 A practical application of Schmid's programme ......................... 48 4.8 Results from stop-w,atch studies ............54 4.9 Time required to set the standard ..........55 4.10 Summary of this chapter ................... 56 CHAPTER V WORK SAMPLING IN A SHORT CYCLE OPERATION 57 5.1 Work sampling in department 41 Latch Assembly ...............57 5.2 Time reguired to set the standard time .....66 5.3 Summary of this chapter .................... 67 CHAPTER VI CONCLO SIONS .............•..«.... .......•....... 6 8 Bibliography ................ ••••»«••»•«•• ••• •'»»«»«» . 70 Appendix A ........................................ 71 Appendix B ...•.•..»«•••«•••••»»......»..•...• 75 Appendix C 80 Appendix D ........................................ 83 Appendix E 92 vi ACKNOWLEDGEMENTS The author wishes to express his profound gratitude to Professor Norman Eley for his guidance and encouragement towards the completion of this work. Mr. Eley has not only been an enthusiastic advisor, but also a good friend who has helped me throughout my graduate studies. Deepest appreciation is extended to the personnel of the Industrial Engineering department at Heiser Lock Co. of Canada for their cooperation during the collection of data for this thesis and also for giving access to their own records. The author is further indebted to Miss Katherine Kassay for her patience and effort in the preparation of this thesis. 1 CHAPTER ,1 INTRODUCTION TO WORK MEASUREMENT The measurement of human work is an integral part of productivity. 1.1 PRODUCTIVITY Productivity has been generally defined as: PRODUCTIVITY = (OUTPUT) / (INPUT) In principle, this productivity ratio is a simple and straightforward one, but if an attempt is made to apply such a measurement over too large an area then considerable complexity may result. Fairly simple examples may be the number of words typed per minute by a secretary or the number of cubic meters dug by a labourer per day. On the other hand, when productivity is applied to a factory, the input and output would have to be expressed in a common guantification factor. Problems arise in defining this factor. Not only is there difficulty with the evaluation of the input (manpower, materials and capital equipment) in homogeneous units, but also with the evaluation of the output. 2 There have been several attempts to evaluate the productivity ratio. Smith and Beeching 1 suggested a factor called 'man-year equivalents*. These were obtained by dividing the actual sum of money spent on the resources by the average industrial income at that time. Capital investment would, of course, require a further division by the years of amortization. By adding together eguivaletst man-years of manpower, materials and capital eguipment, an expression of input is derived which takes inflation into account. This sum was called the 'resources-man-years*. William T. Stewart in his article on productivity2 describes and gives a practical application of a different 'yardstick* for measuring the elusive productivity concept. Mr. Stewart uses the Keeney model of multiplicate multiattribute utility function, in which a representative management group establishes utility values for various departments to help determine productivity goals. * Dennis A. Hhitmore, WORK MEASUREMENT, (Great Britain, Butler S Tanner Ltd, 197 5) p. 2. 2 William T. Stewart: IE PRODUCTIVITY SERIES, Vol. 10 No. 2 of The Journal of Industrial Engineering (February 1978), p. 58. 3 This group selects the relevant ratios that affect productivity such as inventory turnover (cost of goods sold/average inventory), value added per direct labour hour, key machine efficiency (a percent that reflects the actual utilization of high investment machines), material identification and location accuracy, total operating quality cost per net sales dollar, and so on. These factors would depend entirely upon the industry under study. The representative management group would select and decide the number of ratios to take into account. The ratios are multiplied by the scaling factors (rank) and added together giving what Stewart calls the •composite productivity utility measure'. The total productivity is a broad long-term guide to the senior management describing progress of the company. It does not point out the specific areas that may need corrective action. This leads us to the next section in which an attempt is made to subdivide productivity. 1.2 MANAGEMENT BATIOS This is another way of measuring productivity. They are financial ratios rather than measurements related to a defined yardstick such as the •man-year*. 4 Dun & Bradstreet publishes the 'Key Business Ratios in Canada*, a yearly compilation of over 150 lines of retailing, wholesaling, manufacturing and construction businesses. Examples of these ratios might be: PROFIT SALES „ SALES -TANGIBLE NET WORTH SALES INVENTORY It has been claimed that management ratios derived in this way can be used to compare the *state of health* of different companies in the same line of business. From the above discussion, it becomes clear that in most manufacturing companies, the concept cf productivity measurement is a complex one. 1.3 DIRECT BEANS OF IMPROVING PRODOCTIVITY. Against this general background, it is now possible to define work measurement and understand its importance. Generally speaking, productivity can almost always be increased by heavy investment of money in new and improved plants or eguipment. This rise in productivity usually takes time in planning, searching, implementing and testing. 5 On the other hand, an increase in productivity can also be obtained by reducing wasted human effort and time in operating the manufacturing process or eliminating unnecessary movements. The systematic approach of reducing the work content of the manufacturing process is called method study, and the technique of locating ineffective time and setting standards of performance is called work measurement. The formal definitions are: 1 •Method study is the systematic recording and critical examination of existing and proposed methods of doing work, as a means of developing and 'applying easier and more effective methods and reducing cost12 •work measurement is the application of techniques designed to establish the time for a qualified worker to carry out a specified job at a defined level of performance * Method study and work measurement are closely linked. Method study simplifies the job and develops more economical ways of doing it, while work 1 These definitions were adopted from: International Labor Office: INTRODUCTION TO HORK STUDY REVISED EDITION {Switzerland, Impression Coleurs Heber, 1974), Appendix 5, p. 413 - 424. 6 measurement is concerned with the determination of how long it should take to do the job. Work measurement provides the management with a means of measuring the time taken to perform the operations in such a way that the ineffective time will show up and can then be separated from the effective time. Work study acts like a surgeon's knife, laying bare the activities of a company and their functioning, good or bad, for all to see. There is nothing like it for "showing up" people, and for this reason it oust be handled, like the surgeon's knife, with skill and care. Nobody likes being shown up, and unless the work specialist displays great tact in his handling of people he may arouse the animosity of management and workers alike, which will make it impossible for him to do his job properly. 7 CHAPTER II TECHNIQUES OF,WORK MEASUREMENT The existing techniques of work measurement are briefly described in this chapter. Before outlining the techniques of work measurement, the following terms should be defined : * 2.1 TERMS USED IN WORK MEASUREMENT Observed Time The time taken to perform an element or combination of elements obtained by means of direct measurement. Element .- A distinct part of a specified job selected for convenience of observation, measurement and analysis. Standard Performance or Standard Pace . ? The rate of output which qualified workers will naturally achieve without over-exertion as an average over the working day or shift provided they know and adhere to the specified method and provided they are motivated to apply themselves to their work. This performance is denoted as 100 on the standard rating. 1 These definitions were adopted from: International Labor Office: INTRODUCTION TO HORK STUDY REVISED EDITION (Switzerland, impression Coleurs Weber, 1974), Appendix 5, p. 413 - 424. 8 Bating . - The assessment of the worker's rate of working relative to the observer's concept of the rate corresponding to standard pace. , Basic Time or Normal Time .- The time for carrying out an element of work at standard rating, ±,.e;r Observed Time x Observed Bating Standard Bating Standard,Time . - The total time in which a job should be completed at standard performance, i.e. work content, relaxation allowances, unoccupied time and interference allowances, where applicable. Relaxation Allowances .- An addition to the basic time intended to provide the worker with the opportunity to recover from the physiological and psychological effects of carrying out specified work under specified conditions and to allow attention to personal needs. The amount of allowance will depend on the nature of the job. The available technigues may be divided into three major groups : stop-watch time study, statistical standards {work sampling) and predetermined motion time system. Stop-watch time study is a technique for recording the times and rates for working for the elements of a specified job carried out under specified conditions, and for analysing the data to obtain the time necessary 9 for carrying out the job at a defined level of performance. Statistical standards or work sampling, is a technique in which a large number of instantaneous observations are made over a period of tise of a group of machines, manufacturing processes or workers. Each observation records what is happening at that instant and the percentage of observations recorded for a particular activity or delay is a measure of the percentage of time during which that activity or delay occurs. Stop-watch and work sampling involve the measurement of actual observed time and its adjustment to obtain normal time by means of performance rating. Predetermined Motion Time System is a technique whereby times established for basic human motions (classified according to the nature of the motion and the conditions under which it is made) are used to build up the time for a job at a defined level of performance. 10 2.2 STOJ-IATCH_TIMJ_STUDY The 'Survey of Work Measurement and Wage Incentives'1 carried out by Robert S. Sice in collaboration with Patton Consultants and A.I.I.E. showed that the most prevalent approach to work measurement currently used by American and Canadian industries involves a stop watch time study and simultaneous performance rating of the operation to determine normal time. The procedure involves the following steps. 1.- Recording the standard method and identifying the unit of work. The following items are generally included : a) The department in which the job is performed. b) Job number. c) Product, material specifications, and identification as related to the operation. d) Work place layout and dimensions. e) Tool descriptions. .£) Feeds and speeds of machines. , * Rice R. : SURVEY OF WORK MEASUREMENT ANB HA€E INCENTIVES, Vol. No. 7 of The Journal of Industrial Engineering, (July 1977), p. 18 11 2. - Selecting the operator for the study. He has to be experienced and trained in the standard method. 3. - Determining the elemental structure of the operation for timing purposes. The criterion to be followed is that end points have to be easily detected and defined* 4. - Observing and recording the time required for each element. During the observations, the worker is being rated. 5. - Determining the number of cycles to study. This number depends on the duration of the basic time of the cycle, and on the dispersion of the basic times obtained during the study. A common practice is to take ten readings initially of basic time cycles, and from these readings calculate the number of cycles to study. (This topic will be covered in the next section). ., 6. - Computing normal time = average observed time x average rating factor/100 7. - Determining standard time = normal time( 1 • allowances in percent) V 12 2.3 REPRESENTATIVE SAMPLES OF TIMES Seme variation in the cycle time almost always occurs, even if the worker under observation is not attempting to vary his pace. This could be due to the random variation in the operator's pace, position of the parts worked with, position of the tools used, or simply variation caused by slight errors by the time study person in timing the worker. A reasonable limit in the number of readings is to take enough to make the chances 95 out of 100 that the observed average time will be * 5 per cent of the true average This is called in Statistics 95 per cent confidence with + 5 per cent of the standard error of the mean. This criterion has proved to be accurate enough if time studies are going to be used for establishing incentive wages, since • 5 per cent of the standard usually approximates a bargainable increment in wages.2 The sampling theory involves basically two formulae, Egs. 2.1 and 2.2. Both are based on the 1 True average is defined as the average value that a time study person would find if he were to spend his life timing the worker. 2 Hundel M. : HANDBOOK OF INDUSTRIAL ENGINEERING AND MANAGEMENT, SECOND EDITION, Prentice-Hall, 1971, p. 267. 13 assumption that random causes are controlling the length of time of the cycle. In most cases, this is a reasonable assumption *. S = [ /NII2 - (IX)*'.J /N (2.1) 8HEBE : S = Standard deviation. X = Bated individual readings. £ = Sum of homogeneous elements. N = Number of readings of an element. The standard error of the mean Sx indicates the probable variability of the averages of groups of N values of X about the obtained mean or average of all readings of an element. It can be computed by Eg. 2.2. Sx = syT (2.2) The property of this last measure is such that 95 per cent2 of the probable values of the average for the element will lie within • 2Sx of the true average. 1 These equations are derived and explained in Appendix A. 2 Actually 95.4 per cent, but it is usually rounded off to 95. 14 Therefore, if 2Sx is less than or equal to 5 per cent of the mean value, it may be said that the chances are at least 95 out of 100 that the average for the element is within + 5 per cent of the true average. Eqs. 2.1 and 2.2 are combined, and • 2Sx is set equal to 5 per cent of the mean value. The final equation 2.3 is solved for the number of readings N* required to ensure 95 per cent certainty that the obtained average reading will be within • 5 per cent of the true average cycle. »' = C 40 / N ZIX2 - { rxi'iz (2.3) where : N* = Number of readings needed computed from N. N = Number of cycles or readings taken at the time of computing N *. This eguation is easily translated into instructions for a portable programmable calculator.1 2.4 EATING In recent studies the rating factor has proved to play an important role in obtaining the standard time. *A programme for a Hewlett Packard Calculator model 25 is found in Appendix B. 15 Bating is that process during which the time-study analyst compares the performance {speed or tempo) of the operator under observation with the observer*s own concept of normal performance. Bating is a matter of judgment on the part of the time-study analyst, and unfortunately judgement is an indispensable factor in setting a time standard for an operation. Since rating is a subjective judgment entering the computation of the time standard, the immediate question that comes to mind is : "How accurately can experienced people rate?". Controlled studies, in which films from different industries showed workers performing at different rates, proved that experienced analysts could come to an agreement in rating within ••• • 3 per cent 1. 2.5 DETERMINING THE NUMBER OF BEAPINGS A PRACTICAL APPLICATION Consider the time study carried out in one of the local manufacturing firms. In this case, the cycle time was taken as one single element, due to the short elapsed time of the » Buffa E. : MODERN PRODUCTION MANAGEMENT, (New York, John Wiley 6 sons, 1973) p. 423. 16 constituent elements. The rating was posted as 100, since according to the time study man the worker under observation was performing at normal pace. The first ten readings in minutes were: .28 .35 .24 .31 .25 .25 .28 ,26 .25 ,28 Applying the Eg. 2.3 or making use of the programme mentioned.above, the first estimate of the necessary number of readings is obtained: N • =22 Readings, and. Average reading = .283 Minutes. More data were obtained. After each reading, the number of reguired cycles was found using a portable programmable calculator. *** In this case, 20 readings were enough to be included within the 95 % confidence limits within ± 5 % of the true mean. The average reading was .271 min. This average reading is called 'observed time*. The normal time is found by applying the following equation : NORMAL TIME = { OBSERVED TIME } X { RATING } (2.4) In this case, the analyst rated the operator at *** See Appendix C 17 100 %, so the normal time and the observed time were equal. 2.6 DETERMINING PRODUCTION STANDARDS The standard time is found by applying the following equation which includes the personal and fatigue allowances., STANDARD TIME = NORMAL TIME (1+R EL AX ATION ALLOWANCES) (2.5) In this case : Standard Time = .271 x 1.15 = .312 min. or STANDARD OUTPUT •— 192 Parts/hour. 18 CHftPTES III WORK SA BPIIBG 3 • 1 INTBODOCTION Work sampling is a method commonly used to estimate the portion of time that an activity occurs, without necessitating continuous observation. It is a fact finding method based on the laws of probability, especially the binomial distribution. The first article describing work sampling, known at that time as the "Snap Reading Method*, was published by L.H.C. Tippet in 1935. Work sampling is also called activity sampling, ratio delay method, and random observation method. Work sampling has the following uses : 1»~ Ratio,delay .- To evaluate the activities and delays of man or machine. 2. - Performance sampling Same as ratio delay including a rating factor during the productive portion. 3. - Establishment of standard time .- In some studies where the output and the elapsed time of the study are recorded, it is possible to estimate the standard time. 19 In this paper work sampling is used to establish standard times since both the output and the time at which the sampling is done are kept for further computations. work sampling is based upon the laws of probability. A sample randomly taken from a population tends to follow the same pattern of distribution as the population. If the sample is large enough, the characteristics of the sample will differ little from the characteristics of the group. Moreover, it is possible tc predict the accuracy of the difference with a degree of certainty . 3. 2 AH, EXAMPLE OF,,,SAMPLING. Suppose we want to estimate the proportion of time that a worker, or group of workers, spend working and not working. This could be accomplished by long term stop-watch studies in which either the productive or non-productive times are recorded. This is a lengthy and consequently costly approach. Instead, suppose that a large number of observations are taken at random times. The outcome of an observation would be : whether the operator is working or idle and then the observer would tally the results. This tally is shown in the following figure. , 20 TALLY I 1 NUMBER J//// //// //// //// ////I I//// //// //// //// ////I WORKING I//// //// //// //// ////| I//// //// //// //// ////I 98 I IDLE |//// /// TOTAL | 1 1 I I I 105 .i PEE CENT 93.33 6.67 100.00 Figure 3.1 Work Sampling Tally of Working and Idle Time. The percentages of the tallies under working and idle classifications are estimates of the actual percentage of time that the worker was working and idle. Herein lies the fundamental principle behind work sampling : "The number of observations is proportional to the amount of time spent in the working or idle state."1 3.3 CONFIDENCE LEVEL. The statistical principles of work sampling depend upon the binomial distribution as the outcome of an observation is either working or idle; success cr failure . i Buffa E. : MODERN PRODUCTION MANAGEMENT, New York, John Wiley £ sons, 1973 p. 431. 21 Such distribution is identified by: p = X = -NUMBER OBSERVED IN CLASSIFICATION (3.1) N TOTAL NUMBER OF OBSERVATIONS And p = Portion of time that the worker was found working. Sigma = Standard Deviation It is necessary to determine what level of confidence is desired in the final results. The most common confidence interval is 95.45$, which is usually rounded off to 95%, This concept can be visualized as the area under a normal distribution within the limits + 2 Sigma . (3. 2) Where N = Total Number of Observations -2Sigma •2Sigma Figure 3.2 22 3.4 flCCDBACY Of 8QBK SAMPLING The accuracy achieved in work sampling is affected by the number of observations. The purpose of the study will determine the number of observations reguired. The analyst can always estimate the number of observations from previous experiences or from a pilot study. One of the things the analyst must always keep in mind is the inherent variability of the people, machines or processes being measured. In some cases the output of a department may present an ideal situation for work sampling with raw material of reasonably uniform guality, low labor turnover, and good supervision. St other times the situation may not be as favorable, and in such cases the experience of the analyst plays an important role in the study. For many kinds of measurements an accuracy of ±5% is considered satisfactory. Assuming that the binomial distribution is used as a basis for determining the error, the number of observations reguired for 955S of confidence can be found taking ± 2 sigma : 23 Sp = 2 I yp(1-p)/S ] 43.3) Where : S = relative accuracy (• 5%) p = percentage occurrence of an activity N = total number of random observations The relative accuracy is a term commonly used in industry, and is defined as the ratio of standard deviation to the number of readings. From this expression, it is clear that even if the desired accuracy is known, there are two variables: p, the percentage occurrence, and N, the total number of random observations. In order to find N, p is generally assumed or estimated from a pilot study. 3.5 EXAMPLE OF ACCURACY Suppose se want to determine the percentage of productive time of a worker in the department of latch assembly by work sampling. Suppose that 105 observations were taken at random times, and that 7 times the operator was either idle cr out of the working area. Then the percentage of the productive time would be 93.33% (98 / 105 x 100 = 93.33$) And now N can be calculated, knowing that p = .933 and S = • .05. 24 N .05p = 2 [ N .002 5p2 p <3.4) In this case N = 1600 (1-.933) = 115 Observations .933 3.6 ACCURACY.fOB_A GIVEN NUHBEB OF 0 BS EB V A TIP N 5. In practice, the work sampling man covers several stations during his study, and most likely the percentages will not be the same for the different stations. The previous equation can be solved for the relative accuracy S, given the percentage p and the total number of observations N. Continuing with the above example where p = 93.3% and N = 105. S= • .052 In this case , it could be said that we are 35% confident that the operator under observation in the latch-assembly area was working 93.3% of the time within an accuracy of our estimate of • 5.2% of 93.3% V N 25 { 5.2% x 93.3% - 4.87% ), or the true value was between 98.17% and 88.4311, having as a most probable value 93.3%. 3.7 DETERMINISG PRODUCTION STANDARDS. Having found an estimate for the proportion of time that the worker spent doing productive activity, the following step is to compute the normal time. In doing so, the total time of the study in minutes is reguired. This time excludes scheduled delays such as coffee breaks, lunch time and clean-up time. Let us say that the 105 observations were made during a day . There is a 15 minute coffee-break in the morning and a 12 minute coffee-break in the afternoon. Lunch takes 30 min and clean-up time is 5 min. , Adding up all these scheduled delays, it comes cut to: First Coffee-Break 15 min. Lunch Time 30 min Second Coffee-Break 12 min Clean-up Time • 5 min. Total Scheduled Delays 62 min. Working hours are from 7:30 to 16:00 . That means workers spend 8.50 hours or 510 minutes in the factory. 26 Subtracting the scheduled delays (62 min.) from the time the worker spends in the factory (510 min.) gives us the total available time for production which is 510 - 62 = 448 rain. Another factor reguired in our computation is the average rating . In this case, the average performance rating was 10255, that means that for the analyst, the worker under observation was performing slightly above normal performance. Finally the output from this bench was 855 latches for the day. With this information, the normal time is computed from the following expression: (3.5) TOTAL TIME PRODUCTIVE AVERAGE OF STUDY X PERCENTAGE X PERFORMANCE NORMAL TIME = IN MINUTES FROM W.S. RATING TOTAL NUMBER OF PIECES PRODUCED Applying this eguation to the data mentioned above: (Time) (Percentage) (Rating) Normal time = 448 x .933 x 1.02 855 (Output) Normal time = .50 Min 27 The standard time is calculated from the nerval time and the allowances for delays, fatigue, and personal time. In this department such allowances are 15%. Placing these numbers in the following expression; STANDARD TIME = (NORMAL TIME)X{1 + ALLOWANCES) (3.6) It results: Standard time = |.50)x{1+ .15) Standard time = 0.57 min/latch Some manufacturing industries express their standards in terms of pieces/hour for accounting purposes. In this case the output rate is: (60min/hour)* (0.57min/latch) = 104 latches/hour. 28 C H API JB_IV APPLICATIONS OF COMPPTEBS TO-WORK SAoPLING 4.1 INTRODUCTION The work sampling studies which are the subject of this thesis were carried out in a local manufacturing plant, whose main products are domestic and security locks. Some of the constituent parts of the lock are made in the lock company, other parts are supplied by local firms and the rest come from branch plants, which are spread all over the world. The work sampling studies were done with the total cooperation of the personnel of the company, from the General Manager, the Industrial Engineering department, to the workers, who voluntarily agreed to take an active part in this study. The study presented in this chapter was done in the packing department. In this department the parts are supplied by indirect labour to the packing line which is attended by four operators. 29 4. 2 DESCRIPTION OF TBE..LINE A diagram of the line is found in Fig. 4.1. In the first station (Fig. 4.2), the operator opens the box in batches of ten or fifteen and places them on a moving belt. The same operator (Fig. 4.3) is in charge of inserting a small piece (strike) before the boxes arrive at the next station. The next operator gets the parts from an overhead conveyor in sets of two wooden boxes. She takes the outer trim from one of the wooden boxes, the inner trim from the other (Fig. ,4.4A), puts them together (Fig. 4.*}B) and places them in the box (Fig. 4.4C) . The third operator (Fig. 4.5) inserts the latch and screw package and closes the box. The last operator (Figs. 4.6 and 4.7) is in charge of opening and filling the shipping cases. These four operators do not remain doing the same activity all the time, but they change their places in order to achieve a better balance. 30 Operator 2 Assemble trims and place them in box. Operator 1 Open box insert strike 3' 9.5' Operator 3 Insert latch fe 9' Operator 4 Open and fill shipping cases fe 0 50' The numbers denote the operators DIAGRAM OF THE Figure PACKING LINE 31 The activity codes will be described in the following sections. They are presented here for convenience. These activities are illustrated in the following set of pictures. » . - Open or make box. activity Code 11 Figure 4.2 Place parts in box. Activity Code 12 Figure 4.3 32 33 • . - Insert latch. s .- Close the box. Activity Code 12 Activity Code 14 Figure 4.5 Figure 4.6 34 Figure 4.7 5 open or fill shipping cases, place then cn belt. Activity Code 15. * Beaove empty boxes, 7 Walking from stations prepare material. Activity Code 22 Activity Code 21. Figure 4.9 Figure 4.8 35 4.3 THE STUDY As a first attempt, a standard work sampling form depicted in Fig. 3.1 was used. During the study, it was felt that a refinement in the form of presentation of the information was highly desirable, since it would not only show whether the operator is working or idle, but it would also classify the working or idle activity.* This was particularly applicable to the packing line, since there was a lot of walking involved between stations., In order to classify all activities that any of the four workers could be found performing, the line was observed during the packing of a lock called A-151. The activities were divided as follows. 1 This form is found in Appendix B. 36 *. - Open box. 2. - Insert strike. 3. - Assemble outer and inner trim, W place them in box. 0 Insert latch R 5.- Fill shipping case. K close box. I Prepare material, N remove empty boxes. G 7.- Walk from stations. a.- Wait for parts. I Pick up parts from floor. ,. D Out of working area. L 1 *. - Converse. E Table 4.1 Classification of activities . In this case, there are 11 different activities which can be classified as productive or non productive. Activities 1 to 7 represent the working activity, whereas activities 8 tc 11 are non-working. It must be noted that •walking* was considered to be productive since it was a necessary activity in the line. It is customary in this kind of study, to use a table of random numbers cr random times. In this preliminary study, a programme was written for a 37 Hewlett Packard Calculator that generates random times between working hours excluding scheduled delays (coffee-breaks and lunch time)1. A pilot work sampling study was carried out using the improved recording forms and the random times generated by a programmable calculator. The operators were observed seguentially. The operator in Fig. 4.2 was first, the operator in Fig. 4.3 was second, the operator in Fig. 4.4 was third and finally the operator in Fig. 4.5 was fourth. It must be noted that the sequence refers to the operators, not to the stations, since as it was stated before, the operators did not remain at a fixed position. The results from this preliminary study are as follows. 1 The programme and a sample sheet for recording the random times for observations are found in Appendix E 38 ACTIVITY NO. OF PERCENTAGE ACCUMULATIVE OBSERVATIONS PERCENTAG1 Open box. 9 8.1 8. 1 2 Insert strike. 6 5.4 13.5 3 Assemble trims 8 place in box. 36 32.4 45.9 * Insert latch, screw packs. 7 6.3 52.3 s Fill shipping case, close box. 21 18.9 71.2 6 Prepare material, remove empty boxes. 2 1.8 73.0 7 Walk. 13 11.7 84.7 HOBKING TOTAL. 94 84.7 » Wait for parts. 0 0.0 0.0 9 Pick up parts from the floor. 2 1.8 1.8 »o out of area. 13 11.7 13.5 ii Converse. 2 1.8 15.3 IDLE TOTAL. 17 15.Table 4.2 Preliminary study of the packing line. .,-The total number of observations was 111., The output during the study was 82 shipping cases. Each shipping case contains 24 locks, so the output was 1968 locks. 39 The total length of the study was 2 hr. 31 min. or 151 min. Note the high percentage of total observations that represents activity 7, *Walking*. The study could be further divided into 'walking* and 'not-walking*. This new classification allows us to apply the principle of the binomial distribution (see Chapter 3) and to set probability limits in our findings. The same reasoning could be applied to each activity. The operators were not rated during the study. They were considered to be working at a normal pace. The number of readings required for + 5% is computed from Eq. 3.6 N« = 1600 x (1 - . 647)/.847 N* = 2 89 Observations. Hore observations were taken to satisfy the above conditicn, obtaininq the 'following values : Percentage = .870 Output = 223 Shipping cases = 5352 locks. Time of study = 400 min. By applying Eg. 3.7 and taking rating equals to 100%, we obtain : Normal time = 400 min. x .870 / 5352 locks Normal time = .06 min./lock or 923 locks/hr. The personal and fatigue allowances for this department are 15 %. By applying Eg. 3.8, the standard time is : 40 Standard time = (.06) x (1 + .15} = .07 min. / lock, or Standard output — 802 locks/hr. This standard is for the packing cf the lock A-151 when the line is attended by four workers. The relative accuracy is found using Eg. 3.5 S = { 2 / .870 } x { y .870 x (1 - .847) /295 } S =± 4.5 for 95% certainty. During the study, the analyst observed the different paces of the workers, and found that the improved form for recording the information was inappropriate for keeping track of these variations1. Moreover, it was noted that the operators* paces were not consistent. 4.4 BATING IS BOBK SAMPLING. A different procedure for recording the observations had to be devised in order to include the rating factor., Two records were required for each activity classified as working (see Table 4.1). One for the number of observations and another for the rating. For activities classified as 'Idle*, only a record for the number of observations was needed. 1 See form in Appendix B 11 To illustrate this, suppose that the analyst recorded the following ratings while observing a worker performing the activity number S, FILLING SHIPPING CASES AND CLOSING BOXES (see Table 4.1). 90 90 100 90 80 100 90 80 100 90 110 110 90 90 110 100 110 90 100 110 90 The number of observations was 21, which represents 18.9 % of the total observations. The average rating would be the sum of the ratings divided by the number of observations in this category. AVERAGE RATING = I!RATINGS/NO. RATED OBSERVATIONS (4.1) In this case AVERAGE RATING = 2021 / 21 = 96.2 The rated percentage could be easily evaluated from the following expression : RATED PERCENTAGE = AVERAGE RATING X PERCENTAGE (4.2) In the case of this preliminary study for activity 5, we have : Rated percentage = 96. 19/100 x 18.9 = 16.18 % At this point in time, it was becoming slightly difficult and tedious to keep separate records of the activities and the ratings. Moreover, since work sampling was intended to be used in different departments, new records for the job number and department number were reguired. 42 A computer programme was thought to be a suitable solution for the manipulation and processing of the data, since basically what we have here is numerous records that can easily be translated into arrays. The computations themselves are fairly straightforward, 4.5 SCHMID'S COMPOTES PBOGEABflE. , Before launching into the development of this programme, a review of the publications in this field was carried out. It was found that Schmid*s computer programme appeared to be similar and compatible with the ideas presented in this thesis*. Schmid's computer programme is one of the latest publications of a high caliber in work sampling. Thus, to avoid unnecessary repetition, the use of this programme seemed more logical than the development of a new one. Except for the slight modifications and necessary adaptations to the existing computing facilities at U.B.C. {because the programme was originally written for a Model 1130 Computer), this programme remains as designed by Dr. Schmid. * Schmid fl. : WORK MEaSUREMENT SAMPLING, University of Dayton, 1970. 43 For identification purposes. Dr. Schmid makes a further subdivision of 'working*, that is production and production-support. Idle is called non-prod active. Productive activities are given codes starting with 11. Production-support activities begin with 21, and non-productive activities start with 31. During the computations, the computer distinguishes only two major categories. Activities whose numbers are less than 30 are * working* and activities whose numbers are greater than 30 are 'idle* (see Table 4.3). The computer programme holds up to 1000 observations in a single run . It has several arrays to keep track of the number of observations and the rating for each activity. It applies Eg. 4. 1 to find the average rating for each activity, then uses Eg. 4.2 to find the rated percentage. During the calculations, the programme checks for non-productive activities (whose codes are greater than 30). If the search is successful, that activity is rated zero. The total rated activity is the sum of all rated percentages. TOTAL RATED ACTIVITY = ZRATED PERCENTAGE 44.3) where : RATED PERCENTAGE = 0, for activity whose code is greater than 30. This new classification is shown on the following page. 44 DEPARTMENT 49 - PACKING ACTIVITY CODES Activity Significance Code No. Production 11 Open or make box, put on labels. 12 Place parts in box. 13 Insert inner and outer trim and place them in box. 14 Close box or put on lid, apply transparent tape. 15 Open or fill shipping case, place it on belt. Production support 21 Remove empty boxes, prepare material. 22 walk. Non-productive 31 Wait for parts or release stuck wooden boxes. . 32 Pick up parts from floor. 33 Absent, personal time. 34 Talk. Table 4.3 45 From the above table, it can be seen that the activity codes describe all possible activities that the worker can be found performing. This breaking down of the activities into three major groups; (production, production support and non productive) is done by the analyst to achieve a better understanding of the balance of the assembly line (in this particular case packing locks), since for calculation purposes, the computer treats all these 11 activities in two major groups. One comprises all productive and production support activities, and the other includes non-productive activities., This classification into two groups leads us once more to having two conditions as in the case of the previous chapter (working or idle, success or failure). That is why, although there are 11 activities, the binomial distribution is still applicable. This computer programme has provisions for keeping track of the times at which observations are made. It also takes as input parameters the starting and finishing of the shift times, and the starting and length of the scheduled delays (coffee-breaks and lunch time). aith this information, the effective time of the study is calculated according to the following expression : TOTAL TIME STUDY = LENGTH OF THE STUDY - SCHEDULED DELAYS (4.4) 46 Finally, knowing the number of pieces produced, the normal time is calculated by applying Eg. 3.7. It must be noted that the product of «PRODUCTIVE PERCENTAGE FROM H.S. X AVERAGE PERFORMANCE RATING' has been called TOTAL RATED ACTIVITY (see Eg. 4.3), so this eguation becomes TOTAL TIME TOTAL OF STUDY X RATED NORMAL TIME = IN MINUTES ACTIVITY (4. 5) TOTAL NUMBER OF PIECES PRODUCED The standard time is found by applying Eg. 3.8. The computer programme has eight more records or arrays. They are used for the department, operator, job, special, interruption, day, shift numbers, and the size of the lot. 4.6 RANDOM TIMES FOB OBSERVATIONS. Work sampling requires that each individual moment have an equal likelihood of beinq chosen. In order to be statistically acceptable, the observations must be random, unbiased, and independent. A table of random numbers is commonly used in this type cf study to ensure the randomness of the times at which the observations are made. 47 In this paper, a computer programme was developed to generate random times during working hours excluding scheduled delays. This programme uses a function called RAND for generating random numbers. RAND is a function sufcprogramme written in /360 assembly language and callable from FORTRAN. This subroutine generates random numbers X, such as 0<X<1 under uniform distribution. It has a period of 230, and has proved to meet the stochastic properties.++• The random times are printed out in a format suitable for recording the observations. This programme and a run for two positions or stations are depicted in Appendix D. From the observation sheet (see Appendix D), the eight arrays that were mentioned in the previous section are as follow: PEP Nun stands for department number, CPR NOB, stands for operator or worker number. JOB MOMBER aeans the piece number or the assembly number. The SPECial and INTESruption codes are used when the operator stops or starts a particular job. They are also used when the operator shifts his position. These numbers are seldom used. ++• See "UBC RANDOM" Hrite-Op, April 1975 48 For the worker, Activity Code and Production RaTe were described in previous sections of this chapter, DAY-NOMber shall be used to keep track cf the length of the study. The SHIFT number was 1 for all cases in this paper, and the LOT-SIZe has to be posted only once. 4.7 A PRACTICAL APPLICATION OF SCBflID«S PROGRAMME Using observation sheets like the ones mentioned above, a pilot work sampling study was carried out on the belt. In this case, the observations were taken when the operators were packing locks type A-151. The observations are shown in the following three pages. 49 J IDENTIFICATION I | |DEP|OPH| JOB HE|MN|SC|NOMJNUMj NUMBER I I I i 1 | L | TIME I I I I N | I E | ISlIt WORKER t |S| | |P|N| | |DAY|Hj LOT| |E|T|A CJP RTJMUHIIJSIZJ PAGE NO.1 I C | B | I I 1 13 42 30 49 44 8 1 2 13 47 50 49 44 151 21 1 3 13 50 0 49 44 151 21 1 4 13 59 30 49 44 151 15 110 1 5 14 2 40 49 44 151 34 1 6 14 3 40 4S 44 151 14 1 7 14 4 50 49 44 151 14 070 1 8 14 5 20 49 44 151 33 1 9 14 6 10 49 44 151 14 085 1 10 14 6 50 49 44 151 33 1 11 14 7 0 49 44 151 14 085 1 12 14 8 30 49 44 151 34 1 13 14 9 0 49 44 151 13 090 1 14 14 10 40 49 44 151 14 1 15 14 11 50 49 44 151 14 075 1 16 14 12 0 49 44 151 14 110 1 17 14 12 10 49 44 151 13 095 1 18 14 12 20 49 44 151 33 1 19 14 12 40 49 44 151 14 080 1 20 14 15 10 49 44 151 15 110 1 21 14 15 30 49 44 151 11 1 22 14 16 10 49 44 151 14 070 1 23 14 20 20 49 44 151 14 070 1 24 14 21 0 49 44 151 14 110 1 25 14 22 50 49 44 151 13 095 1 26 14 24 0 49 44 151 14 095 1 27 14 29 10 49 44 151 14 075 1 28 14 29 20 49 44 151 33 1 29 14 50 20 49 44 151 14 095 1 30 14 52 50 49 44 151 14 080 1 31 14 55 0 49 44 151 33 1 32 14 56 30 49 44 151 15 105 1 33 14 58 20 49 44 151 31 1 34 15 2 40 49 44 151 31 1 35 15 5 20 49 44 151 33 1 36 15 6 20 49 44 151 13 095 1 37 15 8 10 49 44 151 14 095 1 38 15 8 20 49 44 151 14 090 1 39 15 9 30 49 44 151 33 1 40 15 12 10 49 44 151 13 095 1 41 15 12 50 49 44 151 14 090 1 42 15 13 20 49 44 151 21 1 43 15 14 30 49 44 151 33 1 44 15 15 0 49 44 151 9 11 110 1 |F| .121. 40 Table 4.4 - A 50 I I | TIME £ IDENTIFICATION |S| I| WORKER j | Sj | 1 I I 1 I IDEPJOPRJ JOB ipim i |DAY|H|LOT, 1 N |HR|MN|SCJNUMJNUM| NUMBER IE|T|A C|P RTJ NUM111SIZ\ 1 E I 1 I I I 1 1C | E1 | f 1 — .1 J L_ I JL _1_ J ISJ | ....... j z m i 45 7 51 30 49 44 8 3 1 46 7 53 30 49 44 151 14 080 3 1 47 7 54 30 49 44 151 14 3 1 48 7 54 50 49 44 151 15 3 1 49 7 55 20 49 44 151 12 3 1 50 7 57 10 49 44 151 14 090 3 1 51 7 57 40 49 44 151 13 3 1 52 7 58 10 49 44 151 33 3 1 53 7 59 0 49 44 151 14 080 3 1 54 7 59 20 49 44 151 14 075 3 1 55 8 3 30 49 44 151 14 090 3 1 56 8 4 0 40 44 151 14 3 1 57 8 4 30 49 44 151 14 095 3 1 58 8 7 10 49 44 151 14 3 1 59 8 8 20 49 44 151 34 3 1 60 8 8 30 49 44 151 32 3 1 61 8 8 40 49 44 151 14 090 3 1 62 8 11 30 49 44 151 13 3 1 63 8 11 50 49 44 151 14 095 3 1 64 8 13 10 49 44 151 11 095 3 1 65 8 13 30 49 44 151 14 090 3 1 66 8 15 30 49 44 151 13 3 1 67 8 17 30 49 44 151 14 095 3 1 68 8 18 20 49 44 151 11 095 3 1 69 8 19 20 49 44 151 14 090 3 1 70 8 20 20 49 44 151 14 3 1 71 8 20 40 49 44 151 33 3 1 72 8 21 0 49 44 151 12 090 3 1 73 8 21 10 49 44 151 14 095 3 1 74 8 21 30 49 44 151 13 080 3 1 75 8 21 40 49 44 151 33 3 1 77 8 22 30 49 44 151 14 090 3 1 78 8 24 30 49 44 151 13 090 3 1 79 8 31 10 49 44 151 14 105 3 1 80 8 33 0 49 44 151 13 090 3 1 81 8 34 0 49 44 151 34 3 1 82 8 35 10 49 44 151 13 3 1 83 8 38 0 49 44 151 11 095 3 1 84 8 38 10 49 44 151 13 095 3 1 85 8 38 50 49 44 151 33 3 1 86 8 39 50 49 44 151 15 3 1 87 8 41 50 49 44 151 14 095 3 1 88 8 47 30 49 44 151 14 095 3 1 Table 4.4 - B 51 I I | TIKE J IDENTIIICiTION_ | I | T I |DEP|QPBJ JOB | N |HR|I9N|SC| NUM| N0H| NUMBER I £ I I I I I I l___JL_l__J__J 1_„1_„_ mil WORKER I | SI | |P|M| | |DAY|H|LOTJ |E|T|A C|P BT|NOM|I|SIZ| IC i B| J | |E| | I >B> L J ill \ 89 8 48 0 49 44 151 14 095 3 90 8 50 30 49 44 151 34 3 91 8 55 30 49 44 151 14 095 3 92 8 56 0 49 44 151 14 095 3 93 8 56 3 0 49 44 151 14 090 3 94 8 57 30 49 44 151 14 3 95 8 58 40 49 44 151 34 3 96 9 0 40 49 44 151 31 3 97 9 2 29 49 44 151 31 3 98 9 2 50 49 44 151 31 3 99 9 5 10 49 44 151 34 3 100 9 5 20 49 44 151 31 3 101 9 9 0 49 44 9 3 42 PAGE NO. 3 »»>»»> END CF FILE Table 4.4 - C 52 Each observation represents a computing card. This data was submitted to an IBM 360 computer, giving the following results : JOB N P S STD. TIME STD. OUTPUT NUMBER {MIN.) (PABTS/HOUB) 151 42 80. 6% ±15% .075 803 151 55 80.6% ±13$ .068 877 TOTAL 97 80.6% ±10% .071 839 THEBE HEBE FOUR OPERATORS ATTENDING THE LINE. Table 4.5 Where : N - Number of readings. P = Percentage of time the worker spent working. S = Relative accuracy. The pilot study proved to be satisfactory as far as the methodology was concerned., It took the analyst a short time to get used to the forms and also to relate activities to codes. More work sampling studies were done in the packing line using the same activity codes already described. 53 The results are summarized in the following table: JOB NUMBER 01301020 11017002 11030100 1 103 3102 N 244 168 55 47 9 3.2% 92. 4% 87.3% 94.8% ±3.5% ±4.4% ±10.3% ±6.8% STD, OUTPUT (PARTS/HOUR) 247 179 ** 488 644 THE LINE WAS ATTENDED BY THREE WORKERS. Table 4.6 At this point in time, it became important to find out the effect of rating in the evaluation of the standard. Table 4.7 shows a comparison between the standard calculated including and ignoring the rating factor. JOB NUMBER 01301020 11017002 11030100 11033102 STD. CALCULATED WITHOUT HATING (LOCKS/HOUR) 237 172 478 652 STD. CALCULATED RATIO WITH BATING (LCCKS/HCUR) 247 .960 179 .961 488 .980 644 1.012 Table 4.7 ** An extra worker joined the crew half way through the study. 54 According to the results presented in Table 4.7, we can conclude that the rating factor is important in the setting of standards by work sampling. All these work sampling studies were concurrently carried out with an ongoing stop-watch study programme. 4.8 BESPITS FBOM STOP-WATCH STUDIES Making use of the data collected by the personnel of the Industrial Engineering department of the local lock factory, the following results were obtained. JOB STANDARD SET BY ACTUAL STANDARD SET EY NUMBER *W0BK SAMPLING *0UTP0T* *TIME STUDY 11017002 51 54 75 11030100 163 143 182 Table 4.8 The above figure shows that the standards set by means of work sampling were closer to the actual output reported by the foreman of the packing line,2 * This standard is measured in Locks/Hour/iorker for accounting purposes. 1 Eased on the average over a period of a month. 2 This point will be discussed in the conclusions. 55 4.9 TIME R EOHIBED TO SET TEE STANDARD TIME Table 4.9 shows the break down of an estimate of the time needed for an analyst to set standards following the two methods mentioned in this paper. TIME STUDY 1 HOURS Break down of the elements into components .25 Taking the readings 2.50 Carrying out computations 1.00 TOTAL 3.75 WORK SAMPLING2 HOURS Establishing activity codes .25 Taking the readings 5.00 Carrying out computations3 .25 TOTAL 5.50 Table 4.9 1 This estimate is an appraisal from the personnel of the Industrial Engineering based on their time studies. 2 Based on records kept during the preparation of this paper. 3 Actually, required time to prepare and punch the data, since the computations themselves are carried out by an IBM computer. 56 4.10 SUMMARY OF THIS CHAPTER From the studies presented in this chapter, the following conclusions were drawn : 1. - Work sampling is applicable to an assembly line operation. 2. - The rating factor seemed to be significant in fixing the standard time by means of work sampling. 3. - The time required to set standards was higher using work sampling than using time study. 57 CHAPTER V WORK SAMPLING APPLIED TO A SHORT CYCLE OPERATION In this chapter, a study carried cut in the Latch Assembly Department is described. A brief explanation of how the readings are recorded during a work sampling study is also given. A latch is an important constituent of a lock. In this factory, there are basically two types of latches. One of them is used in bedroom doors. The other one is found in house entrances or in some other places where more security is needed (see Appendix E). 5.1 W ORf_SAMPLING IN DEPARTMENT 41 LATCH-ASSEMBLY First of all, the analyst must understand the mechanics of the assembly, its difficulties and peculiarities. At this stage the analyst would have to decide the pace or tempo that constitutes a normal performance. Equally important is the establishment of activity codes. Figs. 5.1 to 5.3 illustrate some of the wcrker*s activities and the working place. v In this case, the parts for the assembly are supplied by indirect labour. It must be noted that the worker has all the pieces at hand and they are well organized; therefore a high percentage of production • production support is expected. 58 11.- Normal Production 21.- Testing latches Figure 5.1 Figure 5.2 This is the working place, a self staking bench Figure 5.3 Table 5.1 shows the activity codes for this study. 59 DEPABTMENT 41 - LATCH-ASSEBBLY ACTIVITY CODES Activity Significance Code No. Production 11 Normal production. Assembly of parts. Production Suppprt 21 Testing latches, 22 Changing boxes. 23 Repairing assembly or defective parts. 24 Unavoidable Delays: material is supplied to the bench, receiving instructions, recording production. Non-productive 31 Avoidable Delays: idle, talking 32 Absent. 33 Clean-up Table 5.1 60 In this case, the department number is 41, the operator number is the same as the bench number, and the job number is the assembly number at the time of observation. Since the stations are fairly close to each other, the analyst was able to observe eight benches in the same day*s study, taking 105 readings from each bench. The following picture shows the general layout of this department. Department 41 Latch Assembly Figure 5.4 The analyst gathered his data usinq observation sheets described in the previous chapter (page 46), and shewn in Appendix D . The followinq page is part of this study. 61 L 1 TIME I IDENTIFICATION >|S|I|. WORKERI 1S 1 I| | | |DEP|OPR| JOB N |HR:MN:S|NUM|NUM| NUMBER E I | | | | | _ JL 1_ 1 J_ 1. 1 |P|N| |E|T|i |C|E| _1_JBL 1 \ C|P 1 I 1 RT | 1 1 DAY|H|LOT NU13|I | SIZ |F| ITI •< 1 11 7:38:41 41| 111 4 1 I \ 1 1 1 311 1 5 111 1 -_2j_7i_Ui2l_ __1. 1 1 I 1 1 1 I 111 I 111 1 Jl_7:44i5l I 1 • • • • 1 \ 1 1 i 11| 90 | 11 1 1 _1_7_ 49^.51 I - -1 _ _ 1 1 1 1 t 1 11| 1 „;',-1.11 .'-,.,• • I SI 7:55:1J[ I I 1 I 1 1 II 111 I _11JL 1 ___l_8i_3i41 I - 1 -.- . , 1 1 1 I 1 1 11| I 111 1 71 8; 4;2| I . . 1 - . - ... 1 1 1 i II 11| I HI 1 _ _JL__i _.0JL_ i _1 ' _ 1 I 1 1 I j 1 11 " - 1 111 1 Jl_8:12;0|. I 1 •. 1 1 1 1 II 11| 105 | I _1_i__i1«:0| I ... 1 1 1 1 1 1 1 Hi 1 _11J^ I 111 8__2:21 I .-. 1 1 1 1 1 II 111 105| • 111-1 -I_l__i23i0l_ - 1 • . I- - -1 1 1 1 1 1 111 •: I •'i'-- HI 1 131 8:23:4| 1 „ r 1 • 1 1 1 1 1 1 111 I . .. n-i.:.,.,.. I 14j 8;25;5| 1 ..-.,J .. 1 I 1 I 1 1 321 1 . j 11 • • I _1_l_8l_1i31_ 1 1_i_8iJ2:3l_ __1. 1 _1__ . I 1 1 1 1 J_JL 1 1 i I 21| 1 in • — t.H,,;;':,,„;, I 171 8;33:0| 1 _ 1 1 1 1 1 1 1 211 1 11J -1 _18l_8i38:3j 1. _1_ 1 1 1 1 1 1 211 I 111 1 19J 8:40:21 1 ••7,1 ,• ;. ; -1 1 1 1 1 1 111 I 111_ 1 _201 8i4l£41 1 _JL _ _ • . 1 1 1 1 1 1 11| 1 : r .111 -„;., 1 2Jl_8i42i21_ 1 ; 1 I 1 1 1 111_ STATION NO. 1 PAGE NO.1 AUG,25,»77 COfiHEMTS Tafcle 5.2 62 Each line from the observation sheets is punched into an IBB computing card. The vertical line means that the information is to be repeated from the previous card. The listing of the first ten cards of the previous page is given in Table 5.3. c j TIME | IDENTIFICATIC N 1 s I| WOBKEJ | f S I A ! 1 I ! DEP|OPB| JOB P N ACTJ |DAY|H(LOT s HRJMNJS N0M|NUM| NUMBEB |E T COD|BATE|H0H|I|SIZ D t 1 I I C E 1 1 IEI —_ _l_ l_l I __!_. ' — R .___ 1 ____ 1. 121 ___ 1 7i38i4| _ iii 1 __L11 _. 4 _ _ _3JI_. l_5_|1|__ __2 7:41:2 -11I___1J_ ______ — _ii| _!___ili 3 _______ _i» 11 11| 4 _ _111 90 |_5_ |_1 |_— __4 _li___.5 i_j._i_iii_ 4 i _ _ _JLJ. 1 _____ 1 _5_ 1 JL! __ ___ 7:55:1 41J 11| n 1 _ _ 111 1 5 j1| __6 8: 3:4 i-ii I_J_I_. 4 1 _ _' 1JL1 1 _____ 1 Jl 1 __2 8: 4;2 -Jli_lli_ 4 ,- - -- _ '_" ___! 1 5 |1| __8 _______ l__Jll___l_ . 4_ . 1 _ _JII i_5_m 9 8:12:0 411 111 . n 111 1051 5 Hi __10 __i___0 1 411 111 4-- -' •• __UJ l_5_IJI Table 5.3 Note that whenever the analyst rates the operator at normal performance {100%), the rating is left blank. 63 After the computing cards are checked and any mistakes by the analyst or the key-punch operator are corrected, the data is submitted to the computer. The computer reads this data in the following manner: c ! _TIME | IDENTIFICATION - S| I| |S| A | | |DP|OPR| JOE P N I ACT| DAY I Hi LOT R HR|MN|SC|NM|NUM| NUMBER E T COD RATE MUM IUSIZ B 1 I 1 i 1 Cj E |F'| — 1—_l—.1 — 1 — — 1 — - _ _ — 1 ___ __—_ IU_ 1 7;38;401411 111 4 — — _311 — _5_ Hi 2 7i41:20j41| 11| • 4 - - — _UI 100 I 5-; ll 1 3 7-445.501411 111 1 — — _UI 90 I 5 in 4 7:49:50|41| 11| n t — —' 111 100 |_5_ 111— 5 7;55:101 411 111 . 4 .. . 111 109 | 5 111 __6 _8i_3 Jj40 J 4i|_ri I„ 4 - - — _JL1I _100 I 5 111 7 -Si-HlZQ. \ HI1 -111 „ __J1_ i ' — — -Ml -J0J3 _J2_ 111 8 8_;_9: 0|41| 11| 4 — 111 100 Hi 9 8;12: Oi 411 11J 4 111 105 5 ll 1 _1Q 8;14: Oj 411 | 4 111 100 | 5 in Table 5.4 Note that in the case of non-productive activity (31), the computer assigns zero to the rating factor. From this data the following results are obtained: HOBK MEASUREMENT SAMPLING SUMMARY •••*•«• COST JOB START PROD + PROD TOTAL AVG MAN CENTER POS NUM DAY SUPPORT OBS OBS HRS/PIECE 41 1 2 5 93 102 7.277 41 2 2 5 98 105 8. 1S8 41 6 2 5 104 105 8.513 41 8 2 5 83 105 6.692 41 10 1 5 94 105 7.907 41 11 4 5 101 105 8.468 41 12 1 5 91 105 7.6C9 41 13 3 5 103 104 8.265 *** WMS SUMMARY COMPLETE • STANDARD TIME FROM .'. HOBK MEASUREMENT SAMPLING JOB PERCENT RELATIVE AVG MAN TOTAL STD STD NUM P ACCURACY HRS. OUTPUT TIME OUTPUT (PARTS) (MIN.) PARTS/HR 2 90.35 ± 3.2 % 30.680 3335 .552 109 1 93.10 ± 3 1.7 ft 22.698 4550 .299 200 3 99.10 ± 2 !.0 % 8.265 1500 .331 181 4 96.19 ± 3 1.8 % 8.468 875 .643 93 END OF THE JOB EXECUTION TERMINATED Table 5.5 These are the summarized results. For complete printouts see Appendix E. 65 Table 5.6 shows a comparison between the standard calculated without and with the rating factor. JOB STD. CALCULATED STD. CALCULATED RATIO NUMBER WITHOUT BATING WITH BATING (LATCHES/HOUR) (LATCHES/HO UB) 0 2 0 109 109 1.00 0 1 0 205 200 1.03 0 3 0 181 181 1.00 0 4 0 98 93 1.06 Table 5.6 Table 5.7 compares the standard set by work sampling with the actual output and the standards set by time study. JOB STANDARD SET BY ACTUAL STANDARD SET EY NUMBER *WORK SAMPLING *0UTPUT* *TIME STUDY 0 2 0 109 105 119 0 1 0 200 197 190 0 3 0 181 186 191 0 4 0 93 103 116 Table 5.7 * This standard is measured in Latches/Hour. 1 Based on the report of January and February 1978. 66 In Table 5.7, it can be seen that the standards set by means of work sampling were closer to the actual output reported by the foreman of the latch assembly department, 5.2 TIME REQUIRED TO SET TEE STANDARD TIME The time needed by an analyst to set standards (computed in this chapter ) by (i) work sampling and (ii) time study can be broken down as seen in the following table. TIME STODY HOURS Breaking down of the elements into components .25 Taking the readings 4.50 Carrying out computations 2.50 TOTAL 7.25 WORK SAMPLING HOUBS Establishing activity codes .25 Taking the readings 7.50 Carrying out computations .50 TOTAL 8.25 Table 5.8 67 5-3 SUMMARY OF THIS CHAPTER From the study presented in this chapter, it can be concluded that: 1. r Work sampling is applicable to a short cycle repetitive operation such as the assembly of a latch, where the standard time varies from 0.3 to 0.6 minutes, depending on the type of latch (see Table 5.5). 2. - The rating factor seemed to have a small effect on the standard time obtained (see Table 5.6). 3. - The time reguired to set standards was 12% longer in the case of work sampling (see Table 5.8). 68 C HA PTE'S VI CONCIOSIOHS 1. - Instead of standard stop watch, work sampling can be applied to short cycle assembly operations without suffering any loss of accuracy. 2. - Work sampling appeared to be more expensive than time study in terms of actual man-hours reguired on tested applications studied. However, the accuracy achieved by means of work sampling was slightly higher than the accuracy obtained by time study since extra observations were taken. A feed back method such as a simple programme for a portable programmable calculator could be used to compute the number of necessary readings for a given confidence level after a given number of readings. If this is done, it is likely that the economics in terms of man-hours would favour work sampling because the study could be terminated when the desired accuracy level was achieved. 69 3. - Work sampling must be considered to be a more acceptable method psychologically and could be performed by a less skilled, less expensive technician . 4. - The question of the importance of rating is controversial. It seems to be less important in work sampling and this may be related only to the grouping of elements necessitated in work sampling technigues. 5. - Durlnq the work sampling study, the analyst has more time to observe the operation for methods improvement. 70 BIBLIOGRAPHY Whitmore D.A., WORK MEASUREMENT, The Institute of Practitioners in Work Study, Organisation and Methods, 1975, International Labor Office, INTRODUCTION TO WORK STUIY, four impression 1974, Geneva. Dun and Bradstreet, KEY BUSINESS RATIOS IN CASTA LA, 1S73, Merril Lynch Pierce Fenner S Smith Inc., HOW TO R f AD FINANCIAL REPORT, 1973. Journals of Industrial Engineering. A.I.I.E. Barnes R., MOTION AND TIME STUDY, Sixth Edition, John Wiley & Inc. 1968. Mundel M., HANDBOOK OF INDUSTRIAL ENGINEERING AND MANAGEMENT, Second Edition, Frentice-hall, 1971. Buff a £., MODERN PRODUCTION MANAGEMENT, Fourth Edition, John Wiley S Sons, 1973. Bertrand L. Hansen, WORK SAMPLING FOR MODERN MANAGEMENT, Prentice-Hall Inc, 1960. Schmid H., WORK MEASUREMENT SAMPLING, University of Dayton 1970. Schmid M., APPLICATIONS MANUAL - WORK MEASUREMENT SAMPLING, University of Dayton, 1970. Caruth D.L. GUIDELINES FOR ORGANIZING A WORK MEASUREMENT PROGRAM, 1971. APPENDIX A DETERMINING THE NDHBEB OF CYCLES TO STUDY 72 TIME STUDY The number of cycles which must be timed in order to attain a desired level of accuracy depends on the duration of the basic time of the cycle, and on the dispersion of the basic times obtained during the study. Eg. 1 gives a measure of the variability of data about its average. The variability is represented by S, the standard deviation, which is expressed as follows: d = X-X computed from each reading of the element separetely before sguaring and then summing. X = Individual readings of an element. X = Mean or average of all readings of an element. Z = Sum of like items. 8 •= Number of readings of an element. This eguation may be expressed as follows for machine computation (Friden, Monroe, Marchant, etc.): assuming that this represents the varability of a huge group of similar readings or the parent population (a commonly tenable assumption), another measure, Sx, the standard error of the mean (or average) , may be computed from Eg. 2, which indicates the probable variability of the averages of groups of N values of X (D 73 about the mean value. sx = s//T (2) The property of this last measure is such that 95 per cent of the probable values of X (average of the element) will lie within ±2Sx of the true average. Hence, if 2Sx is egual or less than 5 per cent of X, we may say that the chances are at least 95 out of 100 that our average for the element to which the rating will be applied is whithin +5 per cent of the true average representing the performance we observed. If the selected limiting condition is not met, we may work Eg. 2 backward, using the S we first obtained, setting 2Sx egual to 5 per cent of X, and solving for N*, which will indicate the number of the readings that will probabily be needed. Indeed, it is this last property that makes this test feasible, easy and convenient, and economical to use, after certain mathematical manipulations of the formulas have been made.; Combining Eg. 1 and 2, we may state: Sx = 1/ Setting 5 per cent of X egual to 2Sx, we get: 0.05X Where H* is the reguired number of readings. in NUMBER CF OBSERVATIONS IN WORK SAMPLING Work sampling or activity sampling is based on the theories of sampling and probability. The pilot study gives an estimate of the proportion (p) of time spent on a particular activity being studied. The limits of error are set two standard deviations (or, more correctly, two standard errors) from this estimate in order to ensure that the observer may be 95 per cent confident that the estimated error is correct. The standard-error formula for a binomial distribution is: Where N is the number of observations which must be made to ensure a certain reguired accuracy. The limits are set at two errors, therefore: Limits of error (L) = 2y p(1-p)/N Rearranging this formula produces one for estimating the number of observations reguired to attain a reguired error: N = 4p(1-p)/L* setting L = 0.05p N = 1600{1-p)/p APPENDIX B 1. - Improved observation sheet for Work Sampling study. 2. - Programme for a Hewlett Padkard calculator to generate random times during the available time for production. 3. - Form for recording the random times. WORK SAMPLING STUDY Sheet No. Date DEPARTMENT 49 - PACKING LINE ACTIVITY TOTAL PERCENTAGE 1. Open box 2. Insert strike 3. Assemble trims and place in box -4. Insert latch, screw packs 5. Fill shipping case, close box 6. Prepare material, remove empty boxes 7. Walking 8. Waiting for parts 9. Pick up parts from the floor 10. Out of area 11. Conversing TOTAL WORKING % OUTPUT =' TOTAL IDLE % P.F.A. = Title Fy\NDQ//\ TWAE. FOR A WORK 5MAPUN.C Page_J of__i Programmer U. yiLLt.1 np>n«; STEP INSTRUCTIONS INPUT DATA/UNITS KEYS OUTPUT DATA/UNITS I KEY IN PROCRA/A, II 1 a 3 STORE. SEED U_ STO II O I F I PGR/A STORE THE TI/AE FOR: II 1 IP II 3.1 START FIRST COFFEE BREAK START 1 STO II 1 1 1 1 3.3 END FIRSTCOFFEE BREAK END \ STO II Z 1 START LUNCH START Z STO ! 3 1 i 1 3.T END LUNCH END Z 1 STO In-3.1 START SECOND COFFEE BREAK END SECOND COFFEE BREAK START 3 STO 5 II END 3 LsiD__ fe ii CLEAN-UP START 4- LJHDL XIL. i GENERATE RANDO/A R/S 1 i 1 t TWAES I I 1 I i r II ir THE CALCULATOR WILL G.ENER CTE 1 IL _IL — AN INFINITE NUMBER OF II II RANDOM TI/AES. "II 1 __JL_ THE. ANALYST WILL QENERME AS ll 1 II 1 /AANY AS HE WILL NEED. 1 ii ll L AT THE. END OF THE RUN, 1 II II II U; IS RECORDED AND USED 1 F _l FIX, ll <\. IL __. FOR THE NEXT RUN . i RCL II 0 II II 1 u; 1 II II 1 . ... ' . _ j i i " i n i il ll l —-— ! 11 1 i ii •cr. T~] i 1 HP-25 Proqram Form 73 TitieM^mjjyAEi>^BJBSLR.VAI10NS DUilllCJl/ALAVAliABLLJQR page. Switch to PRGM mode, press [7J [ PRGM [ , then key in the program. PRODUCTION _L._of_L_. DISPLAY KEY ENTRY X Y Z T COMMENTS LINE CODE OO \\-.\V' IS73 " • • * • V-01 02 Tr Z.HOO RCL 0 Ui- 1 Tr 03 51 +• Ui-l-nr 04' 05T 5 5 05 1403 (U;-| + TTV 06 1501 q FRAC Ui 07 2300 5T0 0 Ui OS 09 0& 73 8 & 8. Mi UL __ -— 10 ' 05 5 8.5 Ui 11 12 <o\__ 8.5 U; 01 ~ 1 7 8.5U-, 13 73 7. 8.5Ui 14 05 5 7.5 8.5Ui 15 16 51 + 7.5+8.5U] a t A RANDO/A NU/ABE.R B ETWEfcN 7.5 4* Ho.O 2401 PXI_ V START 1 t > 17 i\ ~ X^y t START 1 18 19" 1344-F*<y QTO 44 t START 1 START I "> TEST F OR THE FU «T DELAY. 20 2402 RCL 2 END |i t 21 yfsr~ F*Vy END I t 22 1301 21 ~ 2403 ~ CTO 01 "* *< y END l~1 t END I j •- - -• 23 24 25 26 t "RCL 3 START 2 t 21 " **y ~ t START 2 V; 1441 F\<.y_ START 2 7 27 28 29 30 1344 C.TO 44 t START 2 VTEST F )R THE SE« rOND DELAY. 2.404" "RCL 4 END 2 t -- — -IHSI" Fx*y " ENDZ t 1301 CTO 01 • END 2 t L J -31 32 "33~ 34 2T~* 2405" 2.1 t END I RCL 5 START 3 X^Y _f_>i.y_ t START 3 mi _ _._t START 3 1 35 1344 _ _C«T0 44 . t START 3 -> TEST F RD DELAY. 36 240b 1W RCL (o ENO 3 t 37 ~Fx*y END 3 t 38 39 -----1301 CTO 01 END 3 t t ) 11 2407-*\y END 3 START 4-41 Tior~ Fx<.y START 4-' "t START 4-• TEST F( )R THE FO! flRTH DELAY. 42 43 "GTO" or STARTS' X*< y t 44 45 "1400 FH./AS t -s~ —me IF t AYED FOUR TI/AES. 46 47 _____ ^ _J^78: L.F»Vs: _i36"l—rGT0 61 t __ __ /"THE Tl// VE JS DlSPl 1 I ,;t) i ! REGISTERS 5TABXI. „t_iD__. R5  STML3... END 3 R 7  START H-H E WL ETT [hfjj I>AC K A R D U; - FRACTIONAL PART OF [CTT+U'rO5] \\mEKE O^Uil f,= 7.5 + 8.51V, TORK SAMPLING STUDY RANDOM TIMES FOR OBSERVATIONS 8 9 10 11 12 •n 14 15 : 10 1 0 1 0 1 0 Lunch 1 0 1 0 1 0 20 2 0 20 2 0 Time 2 0 2 0 20 3 0 3 0 • 3 0 3 0 30 3 0 30 7 30 40 1+0 40 40 it 0 4 0 Coffee 40 50 50 Coffee Break 5 0 . 50 5 0 5 0 50 5 0 6 0 6 0 6 0 Lunch Time 6 0 6 0 6 0 6 0 6 0 Clean-up PSEUDO RANDOM TIMES ' UNIFORMLY DISTRIBUTED o< U. < , l 1 U± = Fractional part of [ O + U. _ -) 5 ] °i-l U = o Date =. 80 APPENDIX G Programme for a Hewlett Packard calculator for finding the number of cycles reguired in a Time Study to achieve 95 % confidence within ± 5 % of error. HP-25 Program Form 81 Title^EQUlfitl-L^Ejy-dNCS FOR ^CONFIDENCE. WITHINtS^ERRflRp^ ) _of .1 Programmer _^_\aLXM,aBO^„X ^MBCM^.,„JS2£. STEP INSTRUCTIONS INPUT DATA/UNITS KEYS OUTPUT DATA/UNITS 1 KEY IM PROCRA/A 1 II i ! JNITJALIZE 1 F HPMM 3 ENTER CYCLE Tl/AE LUO=L&E_YS_ -. - .Mr ONLY IF THE 1 II II OPERATOR ISTO BE r ~IIII II .RATED, FOLLOW THE JL JL. JL NEXT Z STEPS; _HLZ.1I 11 -OTHERWISE SKIP THE/A. 1 I1 ! i| 5 PRESS ENTER - rrt" ir II II h ENTER THE RATING R I NO -II KEYS II 11 7 RUN FOR FIRST CYCLE :.R/S. II ll II 1 N 8 REPEAT STEP5 r i ir ~r 3 To 1 , 10 TI/AES r -"i 11 i .."S. FIRST EST1/AATE OF i II ir ~i ! N'.N N* FLASHES TWICE irnr r "> I 10 KEY FEEDING. DATA [ 1! .-II 11' "> 1 ll II II 1 N', N UNTIL N APPROACHES rJL Lzjm N' . • - JL JL .. I II THE PROCBAW . _•• -•_ HL_~]_"7J - Y>_ //VEAN : .STOPS WHEN I - ir nr~ 11 \ N>N* .'"" :\-XJ_LT~J. J _I.Z_ _FOR - A NEW CYCLE . . i JFT. JL 1 PRESS R/S BEFORE •r _L _J i ..ENTERING THE L JL _IL _JL 1 _„ FIRST READING ' ir r~ ir r - -NOTE TO THE USER: i ir ~i~ " ri L 1 IF" RATING IS NOT SPECI :IED r ~~I  i rII— • • ii THE PROCRA/A ASSU/AES 100 ~r irn b) THE PROGRA/A TAKES A MY rr jrnrjL j ..KINDS OF UNITS OFH/AE. [ .zirrx ZJLT J i II rr II i i i 11 1! l-l P-25 Program For 82 TitleJM-QL1A_EP READING, FOR S^/c, COMF\D_NCE-_\yjJJi_5l!/a ERRQ& Page_J of__L__ Switch to PRGM mode, press |T| [ PRGM | , then key in the program: /AARCH 2., \^12> DISPLAY KEY ENTRY X Y Z T COMMENTS LINE CODE OO WWWWv _N^N_*_ Pv X 01 02 x 5: y X R .1571. X 03_ ' I3QS croos LOOP FOR. THE 04" ~ M X. Rx 05"~ ~ 33 EX* l _ Rx "FIRST TEN CYCLES. 06l_ _T.OZ. 2" : IOO Rx 07 71 Rx 08 _l3lO CTO 10 Rx— 09"" 10~ — 22.— Rx _+ N 11 " —E-f— N 12— - - o» J 1 N . 13" ~~0O - 07 \o N 1'4_ 15 _ri2i xi y N lO FK^V N lO 16 1347 C.TO 47 N 10 17 _ _2HfO-T' N IO. N' IS COMPUTED 18 ~ - <bl X Nfcx1 to 19 " "2407 ; RCL 7 -,X ~ Nix1 : 20 ~ ~vsoi T£xY"H N£.xz IO 21 " ^r 10 . J_._.NA.«L_C&xi-22— —14-02. /X lO A »o -— .-- -23_ 04-" - 4. 4-24~ 61 X 4A IO 25 " —fcl 40A 26 "2407 RCL7 -X 40/\ . 27~" —-r ' r 4-0 A Mx 28 "~I502"~" gxV 29 " -J4"IT00~ RXO Nl 30— "1*14 F R/S N' 31 " 1474 FR/S N' 32~ "WT02 FIX 2 N' 33 - ""2.403 RCL 3 N N' 34 ' "~35-"""T4-4-1 ~1347 PK<. y N N' ! CTO 47 N N' . 36 ri421 FX X OMAN • r /AEANIREADJNC 37 30 " 74 1433 FREC, X READ X _j 39 1348 QTO 48 J__LT_.A JL_\yj^CLE._.__ 40 41 42 43 44 45 46 47 48~ 49 ~ -14-7I H434""' FR/S F STK. QTO 00 CLEAR THE. . . _ STACKS. REGISTERS R-7- £Rx. 40yN£,x1-(_x,)l\2 APPENDIX D IBS programme for generating random times under uniform distribution during time available for production for Work Sampling study. An example covering two positions or stations with forty two observations each. $SIGNCN WMSA F0RM=8X11 PAGES=50 PRIO=LOS *U B C UBC» 84 THIS IS NOT A PASSWORD I!!!! $RUN *FTN SCARDS=*SOURCE* PAH=NOSOURCE C ****************4*^ C C THIS PROGRAM FINDS RANDOM TIMES UNDER UNIFORM C DISTRIBUTION FOR A WORK SAMPLING STUDY. C C IT IS ARBITRARILY SET FOR WEISEE, WHEN THE C SHIFT IS CONSIDERED NUMBER ONE, STARTING AT C 7:30, AND FINISHING AT 16:00. C THERE ARE FOUR SCHEDULED DELAYS; AT 9:40, C 11:50, 14:30 AND 15:55. THE LENGTH OF C TBE DELAYS IS: 0:15, 0:30, 0:12 AND 0:05. C C »»> NOTE: C 9:20= 9.333 9.35= 9.583 C 9:40=9.666 9:55=9.916 C 11:45=11.750 12:15=12.250 C 11:50=11.833 12:20=12.333 C 12:30=12.500 13:00=13.000 C 14:30=14.500 14:42=14.700 C »»>15:55=15.916 16:00=16.000 C C THIS PROGRAM USES TWO SUBROUTINES FROM THE C LIBRARY, ONE IS CALLED »RAND», WHICH GENERATES C RANDOM NUMBERS, AND THE OTHER ONE IS CALLED C »SSORT», WHICH SORTS THE TIMES IN ASCENDING ORDER. C c >»»»»»>»»»»»»>»»»»»»>»>»»»» c C >»»»>NOTE TO THE USER OF THIS PROGRAM : C C THE NUflBER 3 IS ATTACHED TO A FILE. IT HAS TO BE C DECLARED IN THE RUNNING COMMAND C TBE ONLY CARD THIS PROGRAM REQUIRES IS, USING C COLUMNS 1 AND 4 THE NUMBER OF PAGES AND THE NUMBER C OF JOBS TO BE STUDIED. C >>>>>»»»»»»»»»»»»»»»»»»»»»»» c DIMENSION TIME (2187) ,IHR (2187) , MIN (2187) ,ISEC(2187) C C * READ THE SEED FROM FILE *S1ED» AND SAVE IT. C READ (3,1) SEED 1 FORMAT(F8.7) SAVSED=SEED C C * INITIALIZE COUNTER C J=0 c C * LOOKS FOR A VALUE FOR NUMBER OF C * OBSERVATIONS REQUIRED. IF IT DOESN'T C * FIND IT IN CARDS, IT ASSIGNS 1 TO C * NUMBER OF JOBS AND 1 TO THE NUMBER C * OF PAGES (27 OBSERVATIONS). C READ (5,2,END=3,ERR=3) NP,NJOB FORMAT (12, 1X,12) 85 GO TO 4 WRITE (6,25) FORMAT(1X,20X,19HERROR IN CATA****** ) NP=1 NJOB=1 THE TOTAL NUMBER OF OBSERVATIONS IS FOUND IF (NJOB.LE. 1) NJOB=1 IF(NP.LE.I) NP=1 N=NP*21*NJ0B TBI ELEMENTS OF THE ARRAY TIME ARE GENERATED CONTINUE SEED=RAWD(SEED) T=7.5+8.5*SEED 4****^ CHECK IF TIME IS HBITHIN PRODUCTIVE PERIOD >»»»»> FIRST COFFEE BREAK »>»»» 15 MIN=.25 HR. IF ( (T.GE. 9. 3333) .AND. (T. LE. 9. 5833) ) GO TO 5 »»»»» LUNCH TIME...... ..... 30 MIN=.50 HR. IF((T.GE. 11. 833).AND. (T.LE. 12.333)) GO TO 5 »»»»» SECOND COFFEE BREAK »>»»» 12 MIN=.20 HR. IF ( (T.GE. 14. 500) .AND. (T.LE. 14.700) ) GO TO 5 »»»»»> CLEAN UP TIME »»»» 5 MIN=.08 HR. IF (T.GT.15.916) GO TO 5 OBSERVATION 8ITHIN PRODUCTIVE TIME J=J*1 TIME(J)=T IF(J.LT.N) GO TO 5 THE ELEMENTS ARE SORTED USING A SUBROUTINE FROM THE LI CALL SSORT (TIME,N, 3) TIME IS CONVERTED TO •HR,HIN#SEC» DO 8 1=1,J IHR(I)=IFIX(TIME(I) ) XMIN=60.0* (TIME (I) -FLOAT (IHR (I) ) ) MIN (I) =IFIX (XMIN) ISEC(I)=IFIX(6.0*(XMIN-FIOAT (MIN (I) )) +0i 005) CONTINUE C 86 C * NOTE: ISEC(I) HEP RESENTS 10 SECONDS C c «««<««««««<<«<«<<«««<«««<<««<«< C THIS PART OF THE PROGRAM PRINTS THE FORMS FOR C WORK SAMPLING STUDY, C C IT ALSO PRINTS THE RANDOM TIMES FOR THE OBSERVATIONS. C C NOTE : THIS PARTICULAR FORMAT HAS BEEN DEVELOPED C FOB WEISER LOCK CO. IN CANADA C <<<««««««<<«««<««««««««<<««<«« c c * c 10 FORMAT<»1«, -55H - - • , -• • -.- - » -1X,7HSTATI0N ) 20 FORMAT(1X, -55H| L | TIME | IDENTIFICATION |S|I| WORKER | |S| | - 2X,3HNO. ,12) 30 FORMAT{••», -55H _; .__ „ , ) - 2X,3HNO.,12) 40 FORMAT{1X, -55fl| I | | | JDEPJOPRI JOB |P|N| | |DAY i H|LOT|, - 1X,8HPAGE NO.,11) 50 FORMAT(1X, -55H| N |Hfi:MN:S|SUM|NUM| NUMBER |E|T|A C|P RTJ NUM11|SIZ|) 60 FORMAT(1X, -55HI E | I | | II I CIE| - | | IF| |, - 1X,9HCATE__ ) 65 FOSMAT(1X, -55H| I I I I I I I IBJ \ I |T| I, - 1X,8HCCMMENTS ) 70 FORMAT{* + », ~55H -. • • - • - : •••• ) 100 FORMAT(1X,1H|,13,1H|,I2,1H:,I2,1H:,I1, 43H| | | I II I I 111 I) 110 FORMAT(1X, -55H| | IIIC C * INITIALIZE COUNTERS FOR THE LOOP C NFIBST=0 LAST=N-NJOB C C » PRINT THE FIRST HEADING OR RETURN AFTER PRINTING 27 LINES C DO 250 I=1,NJOB NSTOP=0 LINE=0 NPAGE=1 NFIRST=NFIRST*1 LAST=LAST*1 C C >»»PRINT THE HEADING C WHITE(6,10) WRITE (6,20) I WRITE(6,30) 87 WHITE(6,UO) NPAGE WRITE(6,50) WRITE(6,60) WRITE(6,65) WRITE(6,70) C C * »THE OBSERVATION SHEET IS PRINTED C DO 200 K=N FIRST,LAST,NJOB LINE=LINE+1 NSTOP=NSTOP+1 WRITE (6,110) WRITE(6, 100) NSTOP,IHR(K) ,HIN(K) ,ISEC(K) WRITE(6,70) C C * TEST FOR NEW PAGE C IF ( (LINE. GE. 21) . AND . (K. LT.LAST) ) GOTO 170 C GO TO 200 170 LINE=NPAGE=NPAGE*1 WRITE (6, 10) WRITE (6,20) I WRITE(6,30) WRITE (6,4 0) NPAGE WRITE(6,50) WRITE (6,60) WRITE(6,65) WRITE (6,70) 200 CON UNO E C C * c c 250 CONTINUE C 300 WRITE(6,350) SAVSED ,NJOB,NP,SEED 350 FOR0AT(,O',/ 1 20X,35HTHE ORIGINA1 SEED FOR THIS RUN IS :,1X,F8.7,/ 2 21X,7HF0R : ,12,14H STATIONS WITH,1X,I2,12H PAGES EACH.,/// 3 29X, 26HTBI SEED FOR NEXT RON IS :,1X,F8.7,/// 4 58X,14HEND OF THE RUN ////) C C * THE * SEED* IS SAVED FOR THE NEXT RUN C BACKSPACE 3 WRITE(3,1) SEED STOP END $Run -load 3=seed 7--temp T= 3 2/02 SSignoff 88 1 I-JUJL-1 I SEN TI f IC AT 10 N |S|I| I-| | T |DEP|OPR| JOB~ |P|N| N |HR:MN:S|NUH|NUM| NOHBER |'E|T| E 1 I I I | | 1C | E | J J 1_J_ _1 J ±_±RL _MORKER_| 1S J I |DAYjH|10T A C|P BT|HUH|I|SIZ 1 1 |F| 1 1 ITI 1 _ Jl_2ii2i0l_ 1 1 1 J. _ J_l_i. . -1 . • 1 111 1 2J 7:35:4! - 1_ I i 1 1 1 1 1 1 Ml 1 __3l 7:46i2l __1 I 1 1 1 _ 1 1 1 _1_ 1 1 4| 7:55:11 III 1 III 1 1 i 1J_ 1 _ 5i_2i59;1| , ; 1 1 1 1 1 III - -1 1 111 1 __6i_8iJ7i5l_ 1 __7i_8i48i3l_ 1 1 1 1 1 | JL _ JiH -J.1J _ 1 _1 1 111 1 81 8iJ3i2l_ _ JL_ 1 II 1 III . J . . I _111_ 1 -_9l_8i58i11_ _. 1 I 1 1 1-1 1 .  .:: 1 1 111 1 101 9: 4:p_l_ 1 III 1 1 1 1 :. 1 j 111 1 111 9i58:2l _ 1_ 1 i 1 JL _ _J._J._1. _1 j 1 12110: 6:41 1 1 1 1 1 _ 1 1 1 •1 - • 1 HI I _13l10j.14l3l„ 1 14,| 10:32:31 III 1 i 1 JL ill -111— __JL___ _ 1 •-- HI 1 15l10i_44;4| 1_ II 1 .1 J_J_± 1 1 111 I 16J10:46:11 __i 1 I 1 1 1 1 1 ..- ... |. . 1 • HI I _17l!Qi47i41_ I __i 1 1 1 JL- . Ill 1 . „ 1 1 I JL__. _1J111i_6i21_ __1 . _ JL _ 1 111 1 19111:22:11, —1 . I I 1 J. ___JL1 i. _ 1 1 .111-1 20111:29:31 1 1 1 1 1 1 1 1 .111 1 -UlllkHlllL- _ 1 1 I 1 J_ __ 1-1-1. _1_ 1 -111— STATION NO. 1 PAGE N0.1 DATE COBBENTS 89 L |_TIME _J IDENTIFICATION I| | J |DEP|OPR| JOB N |HB:MN:SjNOM|NOM| NUMBER E | | | | J | i_„L_-i-± J_L ._ I 22j11:45:4| I I S|I| WORKER | |S| P|N| | |DAY|H|LOT £|T|A C|P RT|NUM|I|SIZ C|E| | | IF| JL-JJU J I 12J I I I .nniimn I. i _J4iJ2i31i2J^ JL I 1 JL_1_JL I ^5U2i35i31_. I 26112:38:51 I I I 28112:59:51 1 l I 1 1 1 I HI 1 1 I_29jj3i_3i4j _l _i 1 1 1 1 •JL.-. I 111 1 1 1 30U3i20i5i_ i .„•{„•;; ;•„•. 1 1 1 1 1 • i_ •-• I- -111 1 1 1 31|13;36:3} 1 -1 _J 1 I 1 1 1 : _ 11J I I I 32113:44:4 1 -1 _ _i 1 1 1 j 1 111 1 1 .33113;56:2| I -J._ _. _J 1 1 1 1 1 111 1 1 1 34|14: 5:01 1 1 1 I 1 1 1 .••_JL__- 1 HI -1 1 1 35|14;14;4| 1 1 1 1 1 I . 11.J . I I l_J6U4i19i3j 1 1 1 1 . i j 1 1 l_J2lli»iMliil_„ II l_38l!4:51:5JL_ I J 1 1 I 1 J JL : I -1UL 111_ ; I I I^39l15i_6i0l . i 1 ' 1 1 1 1 . _JL . 1 .111 II I 40115:18:01 _i 1 1 1 1 1 1 1 . I __111_ 1 1 1 41|15:28:1| 1 _1 1 1 1 1 1 • I 11JL. 1 1 1 42115:32:31 1 _J 1 1 1 1 1 __1U I I JLJL J.UL .11JL .1___11±_. .1 1JJL_. ____1 111 STATION NG. , 1 PAGE NO. 2 DATE. .__ CGBMENTS 90 L I .TIME 1__IEJNTIFICJ1I0N | S 11 I J I | JDEPJOPBT JOB |P|U N |HR:MN:S|NUM|NUH| NUMBER |I|T E I I I I I I IC | E .—l—l—l-l—l—l^——.—^l-lM I .Jl_2:_5:il J L. I i i j._±. i I .-Hl-llMlU—l I I I .2l_8i_0_0i 1. I Jl_8: 55j_5l [ I I .-1-1. 1 I -_JL_1. I _JL _1. JL._. I 10( 9:13;1|, I I I I I _J3l10:2ii:2l_ L. JL i i JLrJL I J-U_______± 1. I J^llQiHllll—XL. I I I __JL I I .-1.1. I i 17|10:51:11 I I ', ,. I .JL_JL_ 1 1 111 , 1 1 18111:15:31 I __1 -I I II I 1 1 HI 1 i 19111:28:51 I _ j_ I I 1 1 1 1 Hi 1 1 | 20i11:31:_i I I I \ 1 1 1 I I t 2 1l11:i*3i2i I 1 1 1 \ 1 1 Ml WORKER | j S| | |DAY|H |I»GT A C|P RT|NUM|I|SIZ i i i Pi J J LU .11L J 11JL. J .111-. 111. 111. All STATION NO, 2 PAGE NO. 1 DATE COHHENTS 91 1 l_TIJJ__J IDENTIFICATION I I | i |D£P|OPfi| JOB N |HR:MN:S|NUM|NUM| NOEBER E I I I I i I .__1__1__1_1___1 L I 1,22111:47;t}| „ _ i •J .11 . 1 1 Hi ., •• 1 1 | 23i-12:29:3| I _1 i i j_l_ . .1 . 1 HI 1 1 I 21112x3 2iU I ._! _ j I i l I JL 1 11J_ I i I 25112:36:4 1 I ._! ' • • • ' i i .JL.1 JL_ . 111 1 1 I 26jJ2i46i11 1 i ._1 1 ! 1 II JL 1 __111_ 1 27U2i59-1| I _1 _ 1 1 1:1: - -1 | 1 I t I l_28iJJi_0i3l_. 1 I 1 1 .1_1__. .1 _ • 1 __111_ ! 1 1 29113: 3;5| i I j 1 1 11 :i:-„-:::. .. j ! I l_JfiJLUi29i11_. 1 1 I 31113:39:41 l_ „ L 1 1 "l 1 11 I ... |... I I J._ _^JLLJb__ t 1 l_32113:48231 I . *!_ _ : — -1. 1 1 .1-1 J. 1 lll_ i i 1 33113:57:01 I 1 .. : 1 1 1 1 1 1 1 111 1 1 1 341141 7x51 I _L 1 1 ll JL 1 Hi 1 1 L.3!l14iJ6: 51_. I _1 1 II I 1 • :ih ; r:Ml,-" 1 1 U36U4i20i3l_ __1__ . L 1 1 i i -J___ 1 | 1 37114:50:31 I • 1- , • i I I 1 JL-- — 1 • 111_ 1 1 l_38114x57141 I ._1 | II 1 I JL. . 1 .... ill..,.. 1 1 1 39115:17:31 I J 1 1 1 1 - 1 111 1 1 1 40115:23x21 I ._1 • • ] 1 1 .1_1__. J.--- I • 111. i i I 41l15x31:5JL_ I | 1 1 .1_J 1 1 __111_ I I I_ii2lJ5xl2xil_. 1. 1 1 J JL_ „111___ S|I P|N E|T C|E I JIORKEJLI 1 S | 1 )DA? |H |LOT 5 C|P RT | NOM |I | SIZ I I IF I 1 L___1T1___ STATION NO. 2 PAGE NO. 2 DAT E__ COMMENTS TBE ORIGINAL SEED FOR THIS RON IS : .576090 FOR : 2 STATIONS WITH 2 PAGES EACH. THE SEED FOR NEXT RUN IS : .011186 APPENDIX E Explosive view of the latches described in Chapter four. An example of the printouts obtained fron tbe computer. uu uu uu uu uu uu uu uu uu uu uu uu uu uu uu uu uu uu uu uu uuuuuuuuuuuu uuuuuuuuuu BBBBBBBBBBB BBBBBBBBBBBB BB BB BB BB BB BB BBBBBBBBBB BBBBBBBBBB BB BB BB BB BB BB BBBBBBBBBBBB BBBBBBBBBBB CCCCCCCCCC CCCCCCCCCCCC CC CC CC CC CC CC cc cc cc cc CCCCCCCCCCCC CCCCCCCCCC WW WW MM MM SSSSSSSSSS WW WW MMM MMM SSSSSSSSSSSS WW WW MMMM MMMM SS SS WW WW MM MM MM MM SS WW WW MM MMMM MM SSS WW WW MM MM MM SSSSSSSSS WW WW WW MM MM SSSSSSSSS WW WWWW WW MM MM SSS WW WW WW WW MM MM ss WWWW WWWW MM MM ss ss WWW www MM MM SSSSSSSSSSSS WW WW MM MM SSSSSSSSSS AAAAAAAAAA AAAAAAAAAAAA AA AA AA AA AA AA AAAAAAAAAAAA AAAAAAAAAAAA AA AA AA AA AA AA AA AA AA AA $RUN WMS.OB 5=PARAM(1,6)+DATA-41+PARAM(7) 6=*SINK* 7=*DUMMY* 8=-FILE8 9=-FILE9 10=-FILE10 ll=-FILEll EXECUTION BEGINS WMS- EDIT INPUT PARAMETERS ACTIVITY CODES 11 21 22 23 24 31 32 33 SHIFT NUMBERS 1 2 3 4 SHIFT START TIMES 750 9999 9999 9999 SHIFT END TIMES - 1600 9999 9999 9999 DELAY 1 DELAY 2 DELAY 3 DELAY 4 DELAY 5 START 1 933 1175 1450 1591 9999 0 9999 9999 9999 9999 9999 0 9999 9999 9999 9999 9999 0 9999 9999 9999 9999 9999 LENGTH 1 25 50 20 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ACCUM. DELAYS 1 25 75 95 103 103 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 GAMA = 0.80 READ DATA CARDS COMPLETE TOTAL DATA CARDS - 840 SORT FOR PART 1 COMPLETE *** WMS PART 1 COMPLETE SORT FOR PART 2 COMPLETE WMS PART 2 - END OF JOB vo WORK SAMPLING SUMMARY IDENTIFICATION CC POS 11 21 NO. OBSERVATIONS 41 1 82 11 ACT. (NOT RATED) 41 1 78.1 10 .5 ACT. (RATED) 41 1 75.7 10 .5 NO. OBSERVATIONS 41 2 80 13 ACT. (NOT RATED) 41 2 76.2 12 .4 ACT. (RATED) 41 2 78.3 12 .4 NO. OBSERVATIONS 41 6 90 12 ACT. (NOT RATED) 41 6 85.7 11 .4 ACT. (RATED) 41 6 85.8 11 .4 NO. OBSERVATIONS 41 8 64 17 ACT. (NOT RATED) 41 8 61.0 16 .2 ACT. (RATED) 41 8 60.9 15 .1 NO. OBSERVATIONS 41 10 89 0 ACT. (NOT RATED) 41 10 84.8 0 .0 ACT. (RATED) 41 10 87.3 0 .0 NO. OBSERVATIONS 41 11 97 2 ACT. (NOT RATED) 41 11 92.4 1 .9 ACT. (RATED) 41 11 94.9 1 .9 NO. OBSERVATIONS 41 12 89 0 ACT. (NOT RATED) 41 12 84.8 0 .0 ACT. (RATED) 41 12 86.7 0 .0 NO. OBSERVATIONS 41 13 101 0 ACT. (NOT RATED) 41 13 97.1 0 .0 ACT. (RATED) 41 13 97.2 0 .0 NO. OBSERVATIONS TOTAL 692 55 ACT. (NOT RATED) TOTAL 82.5 6 .6 ACT. (RATED) TOTAL 83.3 6 .4 WS OUTPUT FILE COMPLETE END OF JOB RUN PAGE 1 ACTIVITY CODES 23 24 31 22 1 1.0 1.0 0 0.0 0.0 0 0.0 0.0 0 0.0 0.0 2 1.9 1.9 1 1.0 1.0 1 1.0 1.0 0 0.0 0.0 5 0.6 0.6 0 0.0 0.0 1 1.0 1.0 1 1.0 1.0 1 1.0 1.0 3 2.9 2.9 0 0.0 0.0 1 1.0 1.0 1 1.0 1.0 8 1.0 1.0 2 1.9 1.9 4 3.8 3.8 1 1.0 1.0 1 1.0 1.0 0 0.0 0.0 1 1.0 1.0 0 0.0 0.0 1 1.0 1.0 10 1.2 1.2 1 1.0 0.0 1 1.0 0.0 0 0.0 0.0 8 7.6 0.0 5 4.8 0.0 1 1.0 0.0 5 4.8 0.0 1 1.0 0.0 22 2.6 0.0 32 8 7.6 0.0 6 5.7 0.0 1 1.0 0.0 14 13.3 0.0 5 4.8 0.0 3 2.9 0.0 8 7.6 0.0 0 0.0 0.0 45 5.4 0.0 33 0 0.0 0.0 0 0.0 0.0 0 0.0 0.0 0 0.0 0.0 1 1.0 0.0 0 0.0 0.0 1 1.0 0.0 0 0.0 0.0 2 0.2 0.0 TOT 105 89.0 105 95.4 105 99.1 105 77.9 105 92.0 105 98.7 105 88.6 104 99.1 839 92.5 WORK MEASUREMENT SAMPLING SUMMARY PAGE NO 1 COST START PROD + PROD TOTAL AVG MAN CENTER POS JOB NUMBER DAY SUPPORT OBS OBS HRS / PIECE 41 1 0 2 0 5 93 102 7.277 41 2 0 2 0 5 98 105 8.198 41 6 0 2 0 5 104 105 8.513 41 8 0 2 0 5 83 105 6.692 41 10 0 1 0 5 94 105 7.907 41 11 0 1 0 5 87 90 7.182 41 11 0 4 0 5 14 15 1.286 41 12 0 1 0 5 91 105 7.609 41 13 0 3 0 5 103 104 8.265 *** WMS SUMMARY COMPLETE STANDARD TIME FROM WORK MEASUREMENT JOB PROD + PROD TOTAL PERCENTAGE RELATIVE AVG MAN TOTAL OUTPUT STD. TIME STD. OUTPUT NUMBER SUPPORT OBS. OBS. P ACCURACY HRS. (PARTS) (MIN.) (PARTS/HOUR) 0 2 0 378 417 .91 + 3.1 % 30.680 3335 .552 109 0 10 272 300 .91 + 3.7 % 22.698 4550 .299 200 0 3 0 103 104 .90 + 1.9 % 8.265 1500 .331 181 vo oo 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0081017/manifest

Comment

Related Items