Wednesday, December 25, 2019

Testing of the Prototype Model of the Proposed End Effector Design - Free Essay Example

Sample details Pages: 15 Words: 4630 Downloads: 5 Date added: 2017/06/26 Category Statistics Essay Did you like this example? Chapter Seven Testing of the Prototype Model of the Proposed End Effector Design 7.1 Introduction The performance of the proposed Dolly Baseboard assembly operation needs to be monitored and analyzed in order to minimize failure. As baseboard and dolly are supplied from different mould cavity in Clipsal, it is important to know the parameters such as the mould cavity number where the parts are made from or inserting speed, that might affect the consistence and cycle time of the assembly process, and to be able to vary or discard them so as to achieve the desired performance of the assembly process. The approach used in this project is an experimental technique known as Design of Experiment (DOE). Don’t waste time! Our writers will create an original "Testing of the Prototype Model of the Proposed End Effector Design" essay for you Create order DeVor, Tsong John (1992) has defined DOE as a statistical tool used in quality design and improvement. The purpose of DOE is to experiment with various combinations of parameters for the purpose of identifying the particular combinations that optimize certain design criteria or performance measures. 7.2 Mathematical representation of DOE In DOE, only the final outcomes obtained by a combination of different variables are of interest. This outcome is usually known as response, which is the performance of the dolly baseboard assembly process in this project. DeVor, Tsong John (1992) states that the response can be represented mathematically by the equation as illustrated in equation 7.1. Assuming that a system involving a mean response that is dependent on input variables x1, x2, . . . , xn. Then could be expressed as This mean that the mean response can be expressed as function with independent variables x1, x2, . . . , xn and a set of parameters 1, 2, . . . , k. The data collected during the experiment are represented by the equation as illustrated in equation 7.2. 7.2.1 Classification of variables by transfer function model The performance of a system can be described by a transfer function model as shown in 7.1. The transfer function illustrates the relationship between the inputs, defined control factors and the output of the process in the presence of noise. According to Taguchis Methods, the factors that can affect the output performance or quality can be classified into mainly four categories which are shown in Table 7.1. Table 7.1 Types of factors in an experiment S/No. Types of Factors Description 1. Signal factors These are factors that may be adjusted by user to attain target performance. 2. Control factors These are the process parameters whose values can be determined during design process. 3. Noise factors They are uncontrollable factors or controllable factors that do not want to control for the purpose of an experiment. 4. Scaling/Leveling factors Special case of control factors that may be easily adjusted to achieve a desired functional relationship between a signal factor and output response. (Source: DeVor, Tsong John, 1992) 7.3 Factors selection in the Dolly Baseboard assembly process One of the most important steps in designing an experiment for Dolly Baseboard assembly process is to select the appropriate factors to study. Basically, there are two groups of factors: The first is the experimental or input factors, those that can be manipulating and are also called the independent variables. The second is the response factors that are measured and are also called the dependent variables. 7.3.1 Input factors and their levels According to Pareto principle, there should be hundreds of factors to choose from, but only a vital few that really make significant changes to the response. It is crucial to identify these input variables and decide on their importance on the quality of output. The input factors and their levels that will have real effects on the performance of the assembly process are presented in Table 7.2. Table 7.2 Type of the input factors used in DOE S/No. Input Factors Description A Baseboard geometry with (Four Levels) Baseboard and dolly are made from different type of injection moulding machines in Clipsal, any geometrical variation might affect the performance of the assembly process, for example; if the geometry of dolly tends to be smaller and baseboard tends to be bigger, the assembly process will become harder. The dolly and baseboard can be selected from four different type of mould cavity to be used in DOE B Dolly geometry with (Four Levels) C Assembling speed with (Three Levels) This is the speed where dolly and baseboard are assembled together. The selection of this speed affects not only the performance of the assembly process, but also the assembly cycle time The setting of these three different speeds can be achieved using teach pendent of jointed-armed robot 7.3.2 Response factor It is also known as output factor, in this case, is to measure the performance of the assembly process. Taguchi mentioned that when selecting the response factors for their practicality, they should be hard, variable-scale factors (1-10) where possible. Thus, factors such as volts, shrinkage, size, and defect count make good response factors. The percent yield is an example of a poor response variable as it does not tell the quality of the rejected material. In this DOE, the performance of the assembly process is measured on a (0 to 2 variable-scale) factor which is explained as below: 0 It indicates poor assembly process; no dollies are assembled to baseboard. 1 It indicates fair assembly process; only one dolly is assembled to baseboard. 2 It indicates excellent assembly process; two dollies are assembled to baseboard 7.4 Design of three factors full factorial experiments The experimental design required is a fixed effect model of three factor full factorial designs. DeVor, Tsong John (1992) defined fixed effect model of factorial design as the levels specifically chosen by the experimenter. In hypothesis testing about level means, the conclusions will apply only to the factor level considered in the analysis. The three variables used in the experiment are assembling speed, baseboard and dolly geometry. The level condition of these variables is listed in Table 7.3 and it is noted that the levels of the assembling speed is in ascending order but not for the dolly and baseboard geometry which cannot be controlled that they might be the same or randomly different throughout the experiment. However, this is not important as it will not affect the final conclusion of the experiment. In order to adopt a standard notation, the variables are coded as A, B, and C with level coded as a, b and c and the possible test conditions will be 4 x 4 x 3 = 48. These are illustrated in Table 7.4. Table 7.3 Variable setting to determine the performance of assembly process Variables Unit Level 1 Level 2 Level 3 Level 4 Baseboard geometry (A) Cavity no.7 Cavity no.8 Cavity no.9 Cavity no.10 Dolly geometry (B) Cavity no.2 Cavity no.3 Cavity no.4 Cavity no.5 Assembling speed (C) mm/s 5 (Slow) 10 (Medium) 15 (Fast) Table 7.4 Coded and actual test conditions in standard order Coded Test Conditions Actual Test Conditions Test no. A B C Baseboard Geometry Dolly Geometry Assembling Speed (mm/s) 1 a1 b1 c1 Cavity no. 7 Cavity no. 2 5 2 a2 b1 c1 Cavity no. 8 Cavity no. 2 5 3 a3 b1 c1 Cavity no. 9 Cavity no. 2 5 4 a4 b1 c1 Cavity no. 10 Cavity no. 2 5 5 a1 b2 c1 Cavity no. 7 Cavity no. 3 5 6 a2 b2 c1 Cavity no. 8 Cavity no. 3 5 7 a3 b2 c1 Cavity no. 9 Cavity no. 3 5 8 a4 b2 c1 Cavity no. 10 Cavity no. 3 5 9 a1 b3 c1 Cavity no. 7 Cavity no. 4 5 10 a2 b3 c1 Cavity no. 8 Cavity no. 4 5 11 a3 b3 c1 Cavity no. 9 Cavity no. 4 5 12 a4 b3 c1 Cavity no. 10 Cavity no. 4 5 13 a1 b4 c1 Cavity no. 7 Cavity no. 5 5 14 a2 b4 c1 Cavity no. 8 Cavity no. 5 5 15 a3 b4 c1 Cavity no. 9 Cavity no. 5 5 16 a4 b4 c1 Cavity no. 10 Cavity no. 5 5 17 a1 b1 c2 Cavity no. 7 Cavity no. 2 10 18 a2 b1 c2 Cavity no. 8 Cavity no. 2 10 19 a3 b1 c2 Cavity no. 9 Cavity no. 2 10 20 a4 b1 c2 Cavity no. 10 Cavity no. 2 10 21 a1 b2 c2 Cavity no. 7 Cavity no. 3 10 22 a2 b2 c2 Cavity no. 8 Cavity no. 3 10 23 a3 b2 c2 Cavity no. 9 Cavity no. 3 10 24 a4 b2 c2 Cavity no. 10 Cavity no. 3 10 25 a1 b3 c2 Cavity no. 7 Cavity no. 4 10 26 a2 b3 c2 Cavity no. 8 Cavity no. 4 10 27 a3 b3 c2 Cavity no. 9 Cavity no. 4 10 28 a4 b3 c2 Cavity no. 10 Cavity no. 4 10 29 a1 b4 c2 Cavity no. 7 Cavity no. 5 10 30 a2 b4 c2 Cavity no. 8 Cavity no. 5 10 31 a3 b4 c2 Cavity no. 9 Cavity no. 5 10 32 a4 b4 c2 Cavity no. 10 Cavity no. 5 10 33 a1 b1 c3 Cavity no. 7 Cavity no. 2 15 34 a2 b1 c3 Cavity no. 8 Cavity no. 2 15 35 a3 b1 c3 Cavity no. 9 Cavity no. 2 15 36 a4 b1 c3 Cavity no. 10 Cavity no. 2 15 37 a1 b2 c3 Cavity no. 7 Cavity no. 3 15 38 a2 b2 c3 Cavity no. 8 Cavity no. 3 15 39 a3 b2 c3 Cavity no. 9 Cavity no. 3 15 40 a4 b2 c3 Cavity no. 10 Cavity no. 3 15 41 a1 b3 c3 Cavity no. 7 Cavity no. 4 15 42 a2 b3 c3 Cavity no. 8 Cavity no. 4 15 43 a3 b3 c3 Cavity no. 9 Cavity no. 4 15 44 a4 b3 c3 Cavity no. 10 Cavity no. 4 15 45 a1 b4 c3 Cavity no. 7 Cavity no. 5 15 46 a2 b4 c3 Cavity no. 8 Cavity no. 5 15 47 a3 b4 c3 Cavity no. 9 Cavity no. 5 15 48 a4 b4 c3 Cavity no. 10 Cavity no. 5 15 7.5 Method of data collection The data collection process begins by randomly selecting the test condition to be conducted. Each of the test condition is replicated 10 times and the mean average yijk is calculated. The total expected number of replicate is 48 x 10 = 480 and the estimated hours required is about 12 hours. In order to maximize the accuracy of the experiment, the test was accomplished in one full day and the same program was used throughout the entire testing, except for the three different assembling speeds mentioned earlier. Due to the insufficient parts supplied from Clipsal, dolly and baseboard must be used repeatedly. Finally, after counting all the available number of parts for the experiment, it was found that dolly and baseboard, which are made of plastics material, could be replaced with a new one in every 5 replications to avoid any inaccuracy to the final conclusion due to excessive bending. 7.2 shows the arranged baseboards and dollies that are ready for the test and a sample of the exper imental result is illustrated in 7.3. The experimental data for test 1 to 48 can be found in Appendix G. Before proceeding to the tests, it is important to make some assumptions to the experiment and they are discussed as below: a. The baseboard and fixture are secure enough to withstand the force exerted by the end effector during the assembly process. It means that their positions are fixed throughout the experiment. b. The area where baseboard is placed in the fixture, and dolly is attached to the end effector are dimensional controlled and should be constant throughout the experiment. It means that if baseboard and dolly geometry do not vary, they should be always at the same assembling position. c. The repeatability and accuracy of the motor in jointed-arm robot are kept to the minimum influence in this experiment d. During the assembly process, the areas where baseboard must be bended in order for dolly to be assembled will always resume to its original position within at least 5 replicate number of test. This means that their dimension will remain constant within 5 repetition of use. e. The geometry of baseboard and dolly made by the same mould cavity number is identical. If there is any geometrical variation, it should be small enough to have little effect in the final result. However, different mould cavity number might produce different part and this geometrical variation is allowed. Test No. Test Date Test venue Equipment used : 1 : 10 Sept 2004 : M-15a (UNISA) : Joint-arm robot Parameters settings Baseboard geometry Dolly geometry Assembling Speed : Cavity no. 7 : Cavity no. 2 : 5 mm/s Replicate no. Response Scale Factor(0-2) Replicate no. Response Scale Factor(0-2) 1 2.0 6 2.0 2 2.0 7 2.0 3 2.0 8 2.0 4 2.0 9 2.0 5 2.0 10 2.0 Mean Average, yijk = 2.0 7.6 Experimental results The graphs in 7.5 shows the results obtained from the 48 sets of experiment conducted. The graph for each test shows the average score for the 10 replicates of the assembly operation and it also indicates the consistency of the assembly process for each combination of the input factor. It can be seen that all test have a full score, except for test 2, 8, 10, 14, 20, 24, 34, 38 and 42 which score 1.9, and test 4 and 6 which score 1.8 for the consistence of the assembly process. The total number of test failed in the assembly operation is 13 and the input factor combination of these 13 tests are labelled and separated from the rest so that the root to their failure can be verified in future, and this is illustrated in 7.4 below. The assembly efficiency for the selected end effector design can then be calculated as below: Assembly Efficiency = = = 0.9729 or 97.29 % 130 Testing of the Prototype Model of the Proposed End Effector Design 7.7 Random order of test The run order of the 48 testing has been randomised as shown in Table 7.5. Randomisation of test order was exercised because it helps to lessen the effects of other factors that are not included in the study, particularly for effects that are time-dependent. Column 2 of Table 7.5 shows the test numbers which were randomly selected to be tested in order and column 6 is known as the response obtained from experiments, which is actually the mean average performance of the assembly process. The variable response scale factor is 0 to 2. Table 7.5 Test order and result obtained from each test 1 2 3 4 5 6 1 2 3 4 5 6 Test Order Test No. A B C (mm/s) Response (0-2) Test Order Test No. A B C (mm/s) Response (0-2) 1 6 a2 b2 c1 1.8 25 44 a4 b3 c3 2.0 2 2 a2 b1 c1 1.9 26 3 a3 b1 c1 2.0 3 30 a2 b4 c2 2.0 27 8 a4 b2 c1 1.9 4 42 a2 b3 c3 1.9 28 7 a3 b2 c1 2.0 5 9 a1 b3 c1 2.0 29 10 a2 b3 c1 1.9 6 12 a4 b3 c1 2.0 30 4 a4 b1 c1 1.8 7 1 a1 b1 c1 2.0 31 13 a1 b4 c1 2.0 8 16 a4 b4 c1 2.0 32 23 a3 b2 c2 2.0 9 25 a1 b3 c2 2.0 33 18 a2 b1 c2 2.0 10 15 a3 b4 c1 2.0 34 21 a1 b2 c2 2.0 11 11 a3 b3 c1 2.0 35 17 a1 b1 c2 2.0 12 19 a3 b1 c2 2.0 36 33 a1 b1 c3 2.0 13 32 a4 b4 c2 2.0 37 27 a3 b3 c2 2.0 14 37 a1 b2 c3 2.0 38 20 a4 b1 c2 1.9 15 28 a4 b3 c2 2.0 39 31 a3 b4 c2 2.0 16 40 a4 b2 c3 2.0 40 26 a2 b3 c2 2.0 17 48 a4 b4 c3 2.0 41 36 a4 b1 c3 2.0 18 22 a2 b2 c2 2.0 42 41 a1 b3 c3 2.0 19 45 a1 b4 c3 2.0 43 43 a3 b3 c3 2.0 20 24 a4 b2 c2 1.9 44 46 a2 b4 c3 2.0 21 34 a2 b1 c3 1.9 45 38 a2 b2 c3 1.9 22 47 a3 b4 c3 2.0 46 35 a3 b1 c3 2.0 23 39 a3 b2 c3 2.0 47 29 a1 b4 c2 2.0 24 14 a2 b4 c1 1.9 48 5 a1 b2 c1 2.0 Note * A = Baseboard geometry a1 = Cavity no. 7 a2 = Cavity no. 8 a3 = Cavity no. 9 a4 = Cavity no. 10 B = Dolly geometry b1 = Cavity no. 2 b2 = Cavity no. 3 b3 = Cavity no. 4 b4 = Cavity no. 5 C = Assembling speed c1 = 5 mm/s c2 = 10 mm/s c3 = 15 mm/s 130 Testing of the Prototype Model of the Proposed End Effector Design 7.8 Use of statistical software to analyze the experimental results Assembly efficiency of 97.29 % was calculated in the previous session and this shows that there will be 1 failure for the assembly operation in every 37 cycles in the actual situation. This is obviously undesirable and must be improved. Therefore, it is necessary to know what factors that actually causes the 2.71 % to affect the consistency of the assembly operation. Generally, the use of statistical software to analyse the results of a designed experiment has been a common practice in the industries where the engineer only requires a fair knowledge of basic statistics and considerable time required for the manually statistic calculation can be reduced. In the case of this project, the software use for the analysis is known as Minitab version 13, which is available in the campus computer pool. The main objective of this software is to analyse the collected data so as to improved assembly efficiency of the selected end effector design by: a. Randomising the run order of the test. b. Generating Analysis of Variance (ANOVA) table to see which factor will affect the assembly operation significantly; this is illustrated in 7.6. c. Generating main and interaction effects plot for response to see which level of the factor will affect the assembly operation significantly. 7.8.1 Analysis of variance for the consistence of the assembly process The aim of Analysis of Variance (ANOVA) is to identify the source of variance that is likely to have influence in the experimental result by comparing the P-value of the input variable to some defined confidence level such as 1 %, 5 % and 10 %. According to the interpretation of ANOVA, the source of variance will become significant if its P-value is less than the defined confidence level. Table 7.6 as shown below tabulates the result of the ANOVA generated from Minitab for the three-factor fixed effect model. From the P-value in the last column of the table, it is observed that the baseboard (A) and assembling speed (C) will significantly affect the response since their P-values are far less than 0.05, which is in fact the standard and common confidence level that is set to be 5 % in this study. Dolly (B) has the P-value that is very close to the 5 % confidence level and this shows that it might have slight effect in the assembly operation. The P-value for baseboard (A) is 0.000 and this shows that it mainly affects the consistency of the assembly process no matter what confidence level is being defined. The (A C) interaction F-ratio has the P-value of 0.005, indicating that there is interaction between baseboards and the assembling speed. Table 7.6 The ANOVA table for the consistence of the assembly process Source of variance Sum of Square (SS) Degree of Freedom (DF) Mean Square (MS) F-ratio (F) P-value (P) Baseboard (A) 0.038958 3 0.012986 11.94 0.000 Dolly (B) 0.010625 3 0.003542 3.26 0.046 Assembling Speed (C) 0.012917 2 0.006458 5.94 0.010 AB (interaction) 0.018542 9 0.002060 1.89 0.119 AC (interaction) 0.030417 6 0.005069 4.66 0.005 BC (interaction) 0.003750 6 0.000625 0.57 0.746 Error 0.019583 18 0.001088 Total 0.134792 47 * Confidence level at 5 % (for general cases) 7.8.2 Interpretation of the level of main and interaction effect Once the source of variance which will affect the assembly operation was identified, it is to focus on what are the levels of that variable input that really causes the assembly process to fail. The relative importance of the level of the three main and interaction effects of the input variables on the assembly operation are shown graphically in 7.7 and 7.8. Both the magnitude and sign of the effects mean something: a. The sign tells the direction of the effects, that is, if the response increases or decreases. b. The magnitude indicates the strength of the effect. From the graph in 7.7 above, it is observed that the sign of the baseboard level changes significantly and that is why it has the 0.000 P-value in the ANOVA table. The best combination of the variable input was found to be (c2, b4 and a1 or a3), and the worst combination of the variable input is (c1, b1 or b2 and a2) which must be discarded in the assembly operation. Moreover, all the level of the variables that fall below the dotted-line, which is the position of the 5 % confidence level used in the study, should be discarded. These levels are a2, a4, b1, b2, c1 and c3. As refer to the graph in 7.8 above, the significant interaction between baseboard and assembling speed is indicated by the lack of parallelism of the average response at each level of the variables and this is obviously why it has the P-value that is less than the 5 % confidence level in the ANOVA table. It can be seen that a2 and c1 are the main causes to this significant interaction and they must be discarded in the assembly operation. In addition, (a4 c1), (a4 c2) and (a2 c3) will also have some effect in the assembly process. The results from the main effect and interaction plot are needed to combine to identify which levels of the variable that must be discarded in order to improve the assembly efficiency. This is illustrated in Table 7.7. It can be easily concluded that a2, a4 and c1 must be discarded as their duplicated existence in the table had mainly caused the inconsistency of the assembly operation during the experiment. The generated statistical data from Minitab version 13 can be found in Appendix H. Table 7.7 Summary of the result of the main effect and interaction plot Severity of the Significance (in the assembly process) Main Effect Plot (Level of the variable input) Interaction Plot (Level of the variable input) 1 a2 a2 c1 2 c1 a2 c3 3 b1, b2 a4 c1 4 a4 a4 c2 7.9 Analysis of the root to the discarded input variable It had been found that mould cavity 8 and 10 which manufactures baseboard, and assembling speed of 5 mm/s must be discarded in order for the assembly operation to be consistent. However, it is important to understand what actually causes these two variables to be discarded so that further recommendations can be used to improve the assembly operation. 7.9.1 Cause of the discarded assembling speed Among the three assembling speeds mentioned previously, medium speed of 10 mm/s is found to be the best while slow speed of 5 mm/s is found to be the worst. It shows that the initial guess of slow assembling speed will be the best is absolutely wrong. As dolly is not really secured by the suction cup of the end effector during assembly operation, there will be a higher chance that it will move around inside the end effector and cause the assembly operation to fail if the assembling speed is too slow. However, if the assembling speed is too fast, the shaft of dolly might not have sufficient time to insert properly into the hole of baseboard and this cause the failure of the assembly operation. As a result, medium assembling speed of 10 mm/s becomes the best choice. 7.9 as shown below illustrates the failure of the assembly operation due to the inappropriate setting of the assembling speed. 7.9.2 Cause of the discarded baseboard Although mould cavity 7 and 9, which manufactures baseboard, produce excellent result during the experiment of assembly operation, it is impossible to discard all baseboard from the other two mould cavities as this will cause the number of baseboard to be insufficient for the assembly operation. In contrast, it is more important to distinguish the geometric difference of baseboard from all the mould cavities. 7.9.2.1 Inspection of the baseboard geometry All the baseboards involved in the failures of the assembly operation are deeply inspected. Since mould cavity 7 and cavity 9 have no influence in the assembly operation, one of them is also inspected to compare with the inspection result of the failed part so that any geometric difference between the measured parts can be identified. Simultaneously, it is necessary to establish some useful dimension to be measured and this is illustrated in 7.10. Basically, the reason to obtain these three dimensions of A, B and C can be briefly explained as below: A. This dimension is important as its variation will directly affect the success of the assembly process, especially when it is too large. B. This dimension will affect the position of the baseboard to be assembled. C. This dimension sits inside the fixture and any variation of it will also affect the position of baseboard to be assembled. From the graph, it can be seen that dimension B and C have the two lines that are very close to each other, while dimension A has the two lines that is quite apart from each other. Moreover, the entire dimension A values of the failed baseboard are higher than the good baseboard. Therefore, it shows that this increased dimension A values of the failed baseboard had significantly caused the assembly operation to fail. Furthermore, out of the three dimensions value, only dimension A value has increased and it indicates that the problem did not come from the shrinkage problem of the mould cavity, otherwise all the three dimension value should have together increased. It is believed that the initial assumption that baseboard will resume to its original position within 5 replicates during the experiment is not valid, and thus causes dimension A value of the baseboard to increase. This is probably due to the different setting of the temperature in the injection moulding machine which event ually affects the material property of the baseboard and causes the assembly operation to fail. 7.9.2.2 Repetitive use of baseboards in the experiment As mentioned before, each of the baseboards was used 5 times before they were replaced with a new piece in the experiment due to the limited supply of the part from Clipsal. This means that each of the new part of baseboard and dolly was used only in the first and sixth replicate of the test number, while the fifth and tenth replicate of the test number are the part that were repeatedly used for five times during the experiment. In this case, the replicate number which failed in each of the test during the experiment of the assembly operation was observed. Table 7.8 as shown below summarize this observation. It can be seen that all the test number of the failed baseboard fall between the forth and last replicate number. It shows that the poor material property of the baseboards had made baseboards sustain excessive bending and that is why they could not resume to its initial position after 4 or 5 repeatedly use, and thus caused the assembly operation to fail. As a matter of fact, baseboard will be required to be bended for only one time in the actual assembly operation in Clipsal and this shows that the actual assembly efficiency should be more than 97.29 % if baseboard or dolly was not used repeatedly during the experiment. However, it is impossible to verify this value by conducting another set of experiments due to the limited time and resource in the project. Therefore, it can be concluded that the repetitive use of baseboard from mould cavity 8 and 10 had significantly increased the baseboard geometry but this will not affect the performance of the actual assembly operation in Clipsal. Table 7.8 Summary of the replication used of the failed baseboard during experiment Failed Baseboard Test No. 2 4 4 6 6 8 10 14 20 24 34 38 42 Replicate No. in Each Test No. 10 5 9 4 10 5 5 9 5 9 5 4 10 No. of Repetition of Use 5 5 4 4 5 5 5 4 5 4 5 4 5 7.9.2.3 Observing of the assembly operation during the experiment During the experiment, not only the results of the 480 replicates of the assembly process were recorded, but their behaviours were also observed very carefully. It was found that the two shafts of dolly did not really insert into baseboard together, but they rather inserted one side and then followed by the other side into baseboard, even when dolly was pushed from its centre. This two-motion insertion situation had made dolly to be shifted very close to one side of baseboard after assembly operation and this eventually might force the inserted dolly to pounce back from baseboard, especially when the geometry of baseboard is somehow increased. This observation is illustrated in 7.12 as shown below. It can be seen that this observation is actually similar to the one that had obtained during the force analysis in chapter 4. Therefore, it can be concluded that either dolly geometry needs to be increased or baseboard geometry needs to be decreased in order to maintain the high consistenc y of the assembly operation. 7.10 Conclusion Since the response of the proposed assembly operation is either pass or fail, which is much different from those situations where specific tolerance is given. As a result, three factors full factorial experiment, which requires tremendous number of trial of 480, was conducted to analyze the performance of the assembly operation. The assembly efficiency was found to be 97.29 %, which is quite far away from the manual one of 99.69 %. The result obtained from the statistical software of Minitab shows that the optimal assembling speed is 10 mm/s and dolly geometry had very little or negligible effect in the assembly process. Baseboard mould cavity number 8 and 10 were initially found to have significant influence on the inconsistency of the assembly operation. However, it was later clarified that the repetitive use of baseboards in the experiment are then the main cause to it The automatic assembly efficiency is then believed to be much higher but this is impossible to be verified by conducting set of experiment due to the limited resource in the project. Finally, observations obtained in DOE and force analysis are used to conclude that either dolly geometry needs to be increased or baseboard geometry needs to be decreased. This is important as it will help to further increase the assembly efficiency.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.