Additional Study Notes Flashcards
6M
Variation occurs because of 6Ms:
- Manpower - operator error, e.g. lack of training
- Machine - worn parts or tools, e.g. improper maintenance
- Materials - variation in raw materials, e.g., wood grain
- Methods - differences in procedures or setups.
- Measurement - measurement- induced error
- Mother Nature- outside influences
Taguchi approach to experimental design
- Planning the experiment 2. Designing the experiment 3. Analyzing the data 4. confirming the solution 5. evaluating the results
Taguchi approach to experimental design: planning
- define the quality characteristics
- translate into measurable quantities
- identify the factors which influence the quality characteristics
- determine the number of levels for each factor
- recognize if there are interactions between factors.
Taguchi approach to experimental design: Analyzing
Hypothesis test to assess whether the effect on quality characteristics is likely to be real or just due to chance.
Taguchi approach to experimental design: Designing (I)
- “full factorial” experimental design
- run each level of each factor together with each level of every other factor: all possible experiments; e.g. 5 factors at 2 levels each requires 2 to the power of 5, which is 32 experiments.
- measure the quality characteristics
- provides info on interactions.
Taguchi approach to experimental design: Designing (II)
- “Fractional factorial” experimental design
- in the real world, not all factors may interact with each other factor
- For independent factors, an experiment can be designed that requires only a fraction of all possible combinations.
- Taguchi revolutionized experimental design by introducing Taguchi Tables
Taguchi approach to experimental design: Confirming the solution
- Determine the best treatment combination
- Run a number of experiments using that treatment combination to confirm solution
- Check the assumption that some (or even all) interactions are 0 or negligible.
- Confirmation of the “optimal” combination is critical.
Taguchi approach to experimental design: Evaluating the results
- Cost benefit analysis to determine the financial implications of the chosen treatment combination
- Product and process design are critical for quality and minimal societal loss
- However, the firm cannot adopt a process that is so expensive that viability is compromised.
DOE good practices
- Use midpoints in your intervals because optimal settings you get in DOE often do not occur at the high or low value, but in between
- Randomize the experimental run order because you can reduce the chance that differences in experimental materials or conditions strongly bias results.
Six Sigma
- 6σ Goal declared by Motorola in 1987
- Core Principles
- Everything is a process
- Every process has variation
- Every process can be measured
- Every process can be improved and variation reduced
- The target is 3.4 defects per million “opportunities” in customer output
- Six Sigma - Management as a System
- Is a philosophy of doing business
- Focuses on eliminating defects through reducing variation
- Uses teams for maximum effectiveness
- Driven by data
- Focuses on results
The objectives of Six Sigma
- Six Sigma aims to define the causes of defects, measure those defects, and analyze them so that they can be reduced.
- A Six Sigma defect is defined as anything outside of customer specifications
- A Six Sigma opportunity exists when there is a chance for a defect to occur
Two Effects of a Six Sigma Project
Two Limitations of Six Sigma
- Risk overlooking VOC
- Not applicable to non-repetitive process
Short-term vs. Long-term Sigma
This relates to two factors:
(1) whether you have short-term or long-term data about the process;
(2) whether you are interested in the short-term or long-term performance (“capability”) of the process.
Depending on these you would apply the shift as follows to the table below:
DMAIC and DMADV
- DMAIC: Define, measure, analyse, improve, control: used for established processes
- DMADV: Define, measure, analyse, design, verify: used for new processes
Sandcone Model
- Process improvement needs to follow a certain logic. You start with quality (meeting customer needs, capability). Next you seek dependability (in control), in other words, to consistently achieve that quality.
- Once that is in place, and only then, do you seek to reduce lead-time by focussing on speed. The outcome of these combined is cost efficiency.
- Important here is that cost efficiency is an outcome, not the focus of your effort. If you make cost efficiency the focus of your process improvement, you are likely to make decisions on a short-term basis that will be unsustainable (as the process is neither in control nor capable).
