A Taguchi Experiment within a DMAIC Improvement Cycle
Section A: Taguchi Experiment and DMAIC
Introduction
Process improvement has in the recent past received significant attention and in this PMA a Taguchi experiment will be performed (within a DMAIC improvement cycle) particularly on the standard design for the optimization of any suitable quality characteristic (Rao et al., 2008). The PMA will also involve analysis of the data from the experiment which will help in providing conclusion on the design as well as providing prediction of the results from the optimized design eventually confirming whether the experiment was successful. The PMA will also involve providing comments on the confirmation results by explaining the selected choice of quality characteristic; definition of control and noise factors included together with an explanation of the selected choices; preparing an experimental design to be carried out together with detailed information on how it was performed; and finally indicating how the results were analyzed as well as outlining the drawn conclusions.
Taguchi Experiment adopts the Taguchi philosophy which reiterates that, if the product didn’t work properly under some of the different conditions, parameters (design constants) are changed for its appropriate functioning (Rao et al., 2008). The Taguchi Experiment is in the old sense considered parameter design. Under the Taguchi method/experiment, functional improvement using tuning should be carried out under standard conditions subsequent to conducting stability design mainly due to the fact that tuning is improvement done on the basis of response analysis. In other words, quality engineering (QE) is focused not on response but on interaction between designs and signals or noises (Atkinson, Donev and Tobias, 2007). Hence, quality engineering maintains that we should do tuning only under standard conditions after completing robust design. Therefore, a two-dimensional tensor does not necessarily represent noises in an SN ratio and in quality engineering; noise effects should have a continual monotonic tendency (Rao et al., 2004). This implies that Taguchi experiment through quality engineering should adjust a function to a target value under standard conditions only after the variability is reduced. The idea is that, functional improvement or robust design for certain conditions of use needs to be completed, following which tuning is done (Pyzdek and Keller, 2009).
Taguchi experiment also helps to design and optimize the quality characteristic, especially the system design which forms the approach to Off-Line Quality Control as defined by Taguchi. System design is the creative part of the process and can be viewed as the action of creating a feasible design, or getting into the design ‘space’. This involves both the parameter design and tolerance design. However, Taguchi experiment also optimizes the quality Characteristic representing the output measure(s) which the experimentation is designed to optimize. They must closely support the key customer satisfaction criteria and should reflect as nearly as possible the energy transformation which is taking place within the manufacturing process.
DMAIC (which is an abbreviation for Define, Measure, Analyze, Improve and Control) is termed as an improvement cycle that is data-driven which is used to improve, optimize and stabilize designs as well as business processes. This implies that in Six Sigma projects the DMAIC improvement cycle is usually the core process improvement tool even though it is not exclusive to Six Sigma meaning that it can also be used as the framework for the facilitation of other design improvement applications. The DMAIC helps to improve performance sustainability through the DMAIC Improvement Cycle whereby the DMAIC consists of a structured problem-solving methodology widely used in design development to improve speed, quality and reduce cost (Bagchi and Kumar, 1992).
DMAIC is an abbreviation of the five improvement steps it comprises such as: Define; Measure; analyze; Improve and Control. All of the DMAIC process steps are required and always proceed in the given order. In particular, each of the steps involved in the process of DMAIC to help improve performance sustainability consists of its respective five improvement steps. For instance, in order to improve performance sustainability through DMAIC, the design problem must be defined through clear articulation of the design problem, its goal, project scope, potential resources as well as high-level project timeline (Bagchi and Kumar, 1992). The project charter document is used to capture all the information concerning the definition of the problem. The second improvement step of DMAIC that helps to improve performance sustainability is the measure in order to objectively establish present baselines which should be used as the process improvement basis. This involves data collection for the establishment of process performance baselines (Ravella et al., 2008).
The third step of the DMAIC that is used to improve performance sustainability is the analysis of the collected data for the identification, validation and selection of root cause for elimination. This allows that the most potential root causes for performance variability are eliminated in order for the process to be improved eventually making sure that improvement of performance is sustainable (Rao et al., 2008). Moreover, the fourth step of DMAIC that is used for ensuring performance sustainability is the actual improvement which involves the identification, testing and implementation of the solution to the design problem, partly or wholly. This is done through identification of creative solutions for the elimination of key design problem root cause or causes. Furthermore, control is the last step of DMAIC that is used to make sure there is improvement in performance sustainability is control which aims to sustain the achieved gains. This involves monitoring of the achieved improvements for making sure the success is continuous and sustainable, especially through creation of a control plan (Ben-Gal, 2005). Moreover, for improved performance sustainability there should be consistent updating of documents, design processes as well as training records. A control chart is also very useful for assessing performance improvements stability over time by acting as a guideline for continuation of the process of monitoring and providing a response plan for each measure considered in the performance improvement (Logothetis and Wynn, 1989).
Why Taguchi Experiment
As noted in the introduction Taguchi experiment helps to design and optimize the quality characteristic, especially the system design applying Taguchi experimentation approaching with an understanding of the key concepts such as robustness, noise and control factors as well as signal to noise ratio (Rao et al., 2004). Taguchi experiments use two approaches such as the one factor at a time approach and the full factorial experimentation approach, and both of them have distinct advantages and disadvantages. However, in both the advantages are usually more compared to disadvantages making them necessary to apply in process improvement. For example, the advantages for the one factor at a time approach of Taguchi experiment are: each run includes only one factor which varies from the one included in the previous run; at first glance it looks scientific; often represents the natural response; and lastly it seems to be efficient (Ghosh and Rao, 1996). The disadvantages of the one factor at a time approach of Taguchi experiment are: the basis for comparisons between factors depends on only two results; with larger numbers of levels and factors they are less realistic; in practice reproducibility of results cannot be guaranteed; and also there is no consideration of the potential inter-dependence of factors (Rao et al., 2008).
The full factorial experimentation approach also has advantages and disadvantages, where the advantages are: it is the most rigorous; the scientific community mostly prefers it; there is investigation of every possible combination of factor levels; and lastly all possible interactions are investigated and its conclusions are highly reproducible. Furthermore, the disadvantages of this approach are: it is very expensive and time consuming; has contributed to avoidance of experimentation in the industry, there is need for large numbers of experiments; and lastly it often use is done in the conditions that are ‘laboratory’ type (Logothetis and Wynn, 1989).
Planning the Experiment
The scope and goals of project
This will mainly consider the quality characteristic that are included as well as the goals of the project which are necessary for the optimization of the quality characteristic of the aeroplane. In this Taguchi Experiment the quality characteristics that shall be considered are the control and noise factor because they are regarded as the key points (Ravella et al., 2008). However, the representative of the failure modes present and the good additivity will also be determined. This will be followed by careful selection of the control and noise factors together with a keen consideration of the possible significant interactions which will require the use of the smallest array which will support the chosen requirements. Finally, in the determination of the scope and goals of the project, the selection of the analysis for the response data will be done in order to give maximum ‘robustness’.
Selection of the right quality characteristic
This involves a careful selection of the quality characteristic which is appropriate to enable the Taguchi experiment to be effectively carried out where in the case of the considered aeroplane the selected quality characteristic is distant where wings length and width of the aeroplane are investigated in this Taguchi experiment since the distant is a crucial quality characteristic of aeroplane that must be carefully analyzed for effective performance of the aeroplane.
Selecting the correct control and noise factors
The control and noise factors considered in this Taguchi experiment are the wings length and the width of the aeroplane mainly because they usually determine the distant of an aeroplane considering that distant is a crucial determinant of the quality of an aeroplane.
Moreover, considering that the selected quality characteristic selected for this Taguchi experiment is distant while the selected control and noise factors are the wings length and width of the aeroplane; then the appropriate experimental regime to be selected for the experiment is bigger- the better. This is due to the fact that when an aeroplane is more distant, a quality characteristic that can be determine by the wings length and width of the aeroplane, the aeroplane will tend have higher quality than the less distant ones. Furthermore, level average analysis is carried out to estimate the optimum combination of factors.
Level Average Analysis
In this Taguchi experiment the level average analysis determines the average response for each factor level of the aeroplane to facilitate the selection of the optimal factor settings through comparison of the average response data. This allows prediction of the process average for the optimal levels which then allows the comparison of the magnitude of the prediction with the confirmation run for the purpose of proving the result (Bagchi and Kumar, 1992). The measurement of the selected quality characteristic as well as control and noise factors using a table is carried out as shown below:
Table 1: Level average analysis
|
TEST |
WEIGHT | ANGLE | WINGS LENGTH | WIDTH | P/C | WING TIPS | C x D |
R |
| A | B | C | D | E | F | G | ||
| 1 | + | + | + | + | + | + | + | 2.26 |
| 2 | + | + | + | – | – | – | – | 2.52 |
| 3 | + | – | – | + | + | – | – | 1.93 |
| 4 | + | – | – | – | – | + | + | 2.81 |
| 5 | – | + | – | + | – | + | – | 3.19 |
| 6 | – | + | – | – | + | – | + | 1.61 |
| 7 | – | – | + | + | – | – | + | 1.65 |
| 8 | – | – | + | – | + | + | – | 2 |
Table 2: Response table
| A | B | C | D | E | F | G | |
| (+) C1 | 2.4 | 2.4 | 2.1 | 2.3 | 2.0 | 2.6 | 2.1 |
| (-) C2 | 2.1 | 2.1 | 2.4 | 2.2 | 2.5 | 1.9 | 2.4 |
| Range | 0.3 | 0.3 | 0.3 | 0.1 | 0.5 | 0.7 | 0.3 |
The two tables above present the level average results for all factors, but the average for the control and noise factors selected for this experiment such as the wings length and the width of the aeroplane are shown below. Therefore, the averages for the length and the width of the aeroplane represented in the table as factors C and D respectively are shown below:
C1D1 = 2.26 + 1.65/2 = 1.96 = 2
C1D2 = 2.52 + 2.0/2 = 2.26 = 2.7
C2D1 = 1.93 + 3.19/2 = 2.56 = 2.6
C2D2 = 2.81 +1.61/2 = 2.21 = 2.2
Estimating the Optimum
Estimating the optimum combination of the level of the selected factors is crucial in the prediction of the response at the levels that are selected for the purpose of making use of the additivity of the effects of each of the selected control and noise factor. However, in the estimation of the optimum only the strong effects are usually used to avoid optimistic predictions. Therefore, the estimating of the optimum was carried out as shown below where the result obtained was 3.20.
(A1B1)C2D1E2F1
M = Ť + (C2D1– Ť) + (E2– Ť) + (F1– Ť)
M = 2.25 + (2.6-2.25) + (2.5-2.25) + (2.6-2.25)
= 2.25 + 0.35 + 0.25 + 0.35
= 3.20
This implies that the expected results were confirmed.
Conclusion
Through the utilization of the Taguchi experiment as the experimental approach, it was possible to carry out a prediction of the outcome of the control and noise factor combination in order to effectively evaluate the selected quality characteristic. Confirmation of the predictions was done by running the Taguchi experiment at the selected levels followed by comparison of the results with the prediction. The predictions in the level average analysis were confirmed through estimation of the optimum combination. The obtained results clearly indicate that the selected control and noise factors directly influence the quality characteristic hence their careful investigation would improve the design process leading to an overall improvement in the quality of the aeroplane.
The Taguchi experiment has numerous benefits as observed in this project because it results to reduction of design defects as a result of reducing the effect of noise variation without necessarily incurring the expense of removing the noise. Also the Taguchi experiment may result to significant reduction of cost because the chosen levels may be less expensive compared to those that operated previously, a factor may also significantly improve the quality of an aeroplane. Moreover, the Taguchi experiment resulted in a robust process for the design manufacture of the aeroplane.
Section B: Reflective writing
It is undoubtedly evident that through the PIUSS module I have managed to learn a lot of things concerning the improvement of processes. In particular, I have gained significant knowledge and information concerning quality engineering which is essential in improving processes that are involved in operations of machines both mobile and immobile. The PIUSS module has been crucial in making sure that I become more knowledgeable in terms of quality issues whenever process improvement is concerned. For example, there are various methods of process improvement which are learned in the PIUSS module which go a long way in making sure that there are significant improvements on the processes concerned with quality improvement.
For example, the widely considered process improvement method covered in the PIUSS module is the Six Sigma method which covers a wide range of factors concerned improvement of quality as well as allowing significant improvements on the production and operational processes. Six Sigma methods is the widely used method of process improvement and utilize various tools and techniques. This set of tools and techniques is very powerful strategy for making sure that production is significantly improved. I have also learned that Six Sigma methods seeks to improve the quality of process outputs through identification and removal of defects (errors) causes as well as minimization of variability in business processes and manufacturing.
Throughout the PIUSS module I have also learned that there is various quality management methods used in Six Sigma including statistical methods as well as creation of a special infrastructure of everyone within the organization who offer expertise in the respective quality management methods. For instance, in every Six Sigma project that is conducted within an organization there are several defined sequence of steps that should be followed, and they should have value targets that can be quantified including reduced pollution, reduced costs, reduced process cycle time, increased profits as well as increased customer satisfaction.
As a result of the scope of the PIUSS module, I can certainly state that I have gained immense knowledge on quality characteristic, system design including control and noise factors all of which are fundamental to process improvement and establishment of effective quality management processes. The PMA project included in the PIUSS module also has offered me an opportunity to refine my knowledge application by giving me hands on experience when it comes to application of knowledge, information and skills obtained throughout the PIUSS module. In particular, through the PIUSS module I have been able to understand that if appropriately applied, methods used in process improvement there would be considerable improvement in the way manufacturing systems and business operations are carried out.
In conclusion, I have to admit that the PIUSS module has significantly changed my approach to process improvement by making me more quality conscious. This will ensure that I give defects (errors) reduction and variability elimination in manufacturing processes the first priority. This would go far in ensuring that process improvement is given the necessary attention.
References
Rao, R.S., Kumar, C.G., Prakasham, R.S. and Hobbs, P.J. (2008). “The Taguchi methodology as a statistical tool for biotechnological applications: A critical appraisal”. Biotechnology Journal, 3 (4): 510–523. Cite uses deprecated parameters (help)
Rao, R.S., Prakasham, R.S., Prasad, K.K., Rajesham, S., Sarma, P.N. and Venkateswar, R.L. (2004). “Xylitol production by Candida species: Parameter optimization using Taguchi approach”. Process Biochemistry, 39 (8): pp. 951–956. Cite uses deprecated parameters (help)
Logothetis, N. and Wynn, H. P. (1989). Quality Through Design: Experimental Design, Off-line Quality Control, and Taguchi’s Contributions. Oxford University Press, Oxford Science Publications. pp. 464.
Atkinson, A. C., Donev, A. N. and Tobias, R. D. (2007). Optimum Experimental Designs, with SAS. Oxford University Press. pp. 511
Ghosh, S. and Rao, C. R. (1996). Design and Analysis of Experiments. Handbook of Statistics 13. North-Holland.
Gaffke, N. and Heiligers, B. (2004). “Approximate Designs for Polynomial Regression: Invariance, Admissibility, and Optimality”. pp. 1149–1199.
Ben-Gal, I. (2005), “On the Use of Data Compression Measures to Assess Robust Designs”, IEEE Trans. on Reliability, Vol. 54, no. 3, 381-388. Available at: http://www.eng.tau.ac.il/~bengal/Journal%20Paper.pdf
Ben-Gal, I., Katz, R. and Bukchin, J., “Robust Eco-Design: A New Application for Quality Engineering”, IIE Transactions, Vol. 40 (10), p. 907 – 918. Available at: http://www.eng.tau.ac.il/~bengal/Eco_Design.pdf
Box, G. E. P. and Draper, N. (2007). Response Surfaces, Mixtures, and Ridge Analyses, Second Edition [of Empirical Model-Building and Response Surfaces, 1987], Wiley.
León, R.V., Shoemaker, A.C. and Kacker, R.N. (1987). Performance measures independent of adjustment: an explanation and extension of Taguchi’s signal-to-noise ratios (with discussion), Technometrics, vol. 29, pp. 253–285
Bagchi, T.P and Kumar, M. (1992) Multiple Criteria Robust Design of Electronic Devices, Journal of Electronic Manufacturing, vol. 3(1), pp. 31–38
Ravella S. R., C. Ganesh K.R. Prakasham, S. and Hobbs, P.J. (2008) The Taguchi methodology as a statistical tool for biotechnological applications: A critical appraisal Biotechnology Journal Vol. 3: pp. 510–523.
Adams, C.W., Praveen, G., and Wilson, C.E. (2003). Six Sigma Deployment. Burlington, MA: Butterworth-Heinemann. Cite uses deprecated parameters (help)
Breyfogle, F.W. (1999). Implementing Six Sigma: Smarter Solutions Using Statistical Methods. New York, NY: John Wiley & Sons.
Keller, P.A. (2001). Six Sigma Deployment: A Guide for Implementing Six Sigma in Your Organization. Tucson, AZ: Quality Publishing.
Pande, P.S., Neuman, R. P. and Cavanagh, R.R. (2001). The Six Sigma Way: How GE, Motorola, and Other Top Companies are Honing Their Performance. New York, NY: McGraw-Hill Professional. Cite uses deprecated parameters (help)
Pyzdek, T. and Keller, P.A. (2009). The Six Sigma Handbook, Third Edition. New York, NY: McGraw-Hill.
Snee, R. D. and Hoerl, R.W. (2002). Leading Six Sigma: A Step-by-Step Guide Based on Experience with GE and Other Six Sigma Companies. Upper Saddle River, NJ: FT Press.
Tennant, G. (2001). SIX SIGMA: SPC and TQM in Manufacturing and Services. Aldershot, UK: Gower Publishing, Ltd.
Use the order calculator below and get started! Contact our live support team for any assistance or inquiry.
[order_calculator]