Question Set Five Quality Policy and Statistical Process ...



Question Set Four Quality Policy and Statistical Process Control (SPC)

1) Why Quality Control: QM provides resources to develop correct processes. But what is considered correct changes and processes suffer entropy thus, the need for constant monitoring. With QC operators can recognize process drift, variance drift, know what a process is doing relative to specifications, and can tell if process improvements worked relative to specifications and the original process. If a process is robust, why have QC? Cost of failure vs cost of inspection. Kp rule or Ci or k1 is cost of inspection per unit, Cf or k2 is cost of failure per unit, K = k1/k2 and when p is greater than k ( 100% inspection if p is less than k ( no inspection; p is probability of defective part. Acceptance sampling ???

2) What is a Quality Control Policy (QCP)? A description of the process used to track a production process and why that quality process is being used. The QCP also includes information relative to the production process.

• Determine context of the production process relative to best of practice and why that specific context exists. Determine what portion of the context should be changed and if it can be changed, and what it would take to create that change? Evans and Lindsay (E&L) Ch 1-9

o Context is organizational structure, organizational culture, approach to human resource/capital, and attitude towards and use of quantitative and qualitative tools to make sound decisions. How effective is the firm relative to the three main ideas of QM (stakeholder centric, goal congruency (employee involvement), continuous improvement) and their skill in executing management/leadership, assessment, design, conformance, and control (MADCC ) in all levels of the organization. Production process means you are making something: a good, a service, a decision, a strategy, a report, a … These processes encompasses both support and primary processes (Michel Porter’s value chain). Context will constrain process management, that is why it is necessary to understand context if you are going to do a good job of process management.

• Determine what is the product the production process is currently creating and why is it that product, is it the correct product, what is the delta between what is produced and what is wanted by the customer? E&L CH5

▪ Needs analysis: characteristics of product (Garvin 8, Kano 3, Juran 3). Correct product: one that provides greater value to customers than competitor’s at a price that satisfies all stakeholders to the extent that cash flows are not negatively affected. Remember that a process can be a product.

• Determine the current production process for producing the product in question (process flow diagram E&L CH7). This step in conjunction with the next step is a ‘lean’ analysis.

What steps of the production process determines appropriate levels of the characteristics delineated in number two above?

What is done at each step? Why is it done that way?

What inputs, resources, business rules, and information are utilized at each of the above steps?

What are the duration times?

What are the bottleneck, Takt time, throughput time, and capacity?

How well is the bottleneck managed?

How are they controlling and assessing the current process?

What is wrong with the current production process and control process, if anything (Pareto Analysis and other mgt tools)?

What ‘waste’ (lean) is apparent?

▪ What are the pain points?

▪ This process is the current production process, not the proposed production process or its associated control process.

• Determine what level of performance (actual product) is currently achieved relative to expected (correct product, specification and/or DPMO) and the degree to which the process addresses those expectations. What does the current level of performance cost them in lost revenue, rework, inspection, and/or additional costs relative to possible/desirable outputs? E&L CH 10-13 Current and expected level of performance of the production process, one should also delineate what level of capability the control policy should have based on what is possible with current technology. This step is a lean/waste analysis.

• Develop a set of controls, assessment and/or process changes you think would achieve expected performance and an over arching, general statement of the purpose for the various control methods and process changes you propose. E&L CH 10-13 For each step of the production process, determine if it is needed and/or what can be done to increase performance of the process. Determine how you are going to control the activities of each of your revised steps. Really, you are creating a better production process, then a control process for that production process. The control portion of your manuscript should address how you propose to maintain real time control and should contain such statements as, ‘To determine overall effectiveness of the production process and control process we will perform end of production process evaluation by …..’

o To get here, the information from 3 above or your new improved process steps and characteristic thereof are absolutely necessary. Do you control the inputs and/or the resources used to alter/shape/evaluate the inputs, how do you know the information received and sent is correct (step 18 below)…..

• Referencing the production process, how good does this production process have to be to satisfy customers? Where will you set the control limits for those steps that you are using SPC? Will you use the standard +/- 3 standard deviations, more than 3, or less than 3? E&L CH 10-13 The answer to this question will determine how ‘good’ the control process has to be and help you calculate the costs of Type I and Type II errors, which determines how many standard deviations you set your control limits at, how often you measure, and the sample size you take when you measure.

• For each of the controls and assessments in number five above, determine what to measure (inputs, outputs, durations…) that will insure ‘correct product’ status. E&L CH 10-13 referring to the new production process

• Express why the above measures are necessary and why others are not. E&L CH 10-13

• Determine how each of the different measures will be made. E&L CH 10-13 This step helps to delineate your control process and includes how the control or measure will be done. For instance, will a person take a sample and then analyze that sample with something, or will a machine take the sample and analyze that sample, or will a machine measure continuously and a software program analyze the results, compare to a standard and then direct another machine to adjust the process, or…

• Determine where on the product, input, and/or process-time-line each measure will be performed. E&L CH 10-13 this step helps to delineate your control process

• Determine where in the production process the product, input, and/or process time line each measure will be performed. E&L CH 10-13 this step helps to delineate your control process (think “The Goal” by Goldratt)

• Determine who and/or what should perform each of the measures. E&L CH 1-13 This step helps to delineate your control process

• Determine sample size for each of the measures. E&L CH 10-13 This step helps to delineate your control process (think rational sub-group and probability of catching a given shift in the mean, measured in standard deviation, and cost of measuring versus cost of mistakes happening.)

• Determine how often each of the measures should be performed. E&L CH 10-13 This step helps to delineate your control process. Robustness of process, cost of mistakes, and cost of measuring and analyzing should all be considered.

• Determine how process operators should react to the data generated from each of the measures. E&L CH 1-13 This step helps to delineate your control process (is it common or special cause of error, is the measure a real time feedback mechanism or an end of production step or process evaluation).

• Determine how (and why) the data from each of the measures should be displayed (most likely a control chart of some type or large board in the production/office area). E&L CH 10-13 This step helps to delineate your control process.

• Determine who (and why) should have access to the displayed data. E&L CH 1-13This step helps to delineate your control process and stakeholder relationships.

• Determine how the effectiveness of the QCP should be determined E&L CH 12 how are you going to ensure that your control process is working (calibration of measuring tools and measurer, is the measure being done ch 12 pp 623-630).

• Conduct a cost to benefit analysis to ensure that the NPV/IRR/Pay Back Period of your proposed quality control process and/or process improvement is positive. In addition, conduct a sensitivity analysis that demonstrates breakeven and the range of possible results depending on variation of input variables. Attachments and E&L CH12 Area under the curve, Taguchi cost curves, Solver.

3) Why should there be a process improvement effort and a QCP? Consistency. So everyone knows what to measure and why they are measuring it. Management will always get the same reports and will know what is going on.

4) What is SPC? Statistical method of monitoring processes to determine state of control. (changes in degree of variation, and mean drift)

5) What are control charts? Line chart with data plotted overtime relative to a mean and upper and lower control limits of variation, usually +/- 3 SD/sqrt(n) , depending on cost of detecting a Type I and Type II error relative to the cost of not detecting those errors

6) What is SPC good for?

a) Display variation of process to determine if process is in control

b) Should process be adjusted

c) Determine capability (given there are specifications and have a continuous variable type of data)

d) Benchmarking changes and to other like processes

e) Displaying process information (capability and history) to stakeholders

|SPC and Capability |Process in Control |Process out of Control |

|Capability >= 1.33 |Like this situation |Adjust process |

|Capability < 1.33 or unknown |Reduce Variation by changing process (if |Get process under control and then reduce |

| |cost effective), but in the meantime |variation (if cost effective) all the while|

| |inspect defectives out if cost effective |inspecting defectives out if cost effective|

7) What state does the variance and mean have to be in for control charts to be used? In other words, what does statistical control mean? Statistical control means that the variance and mean of a process are constant over time and occurrences that have a low probability of happening, happen with a low probability.

a) Can control charts be used when there is no variance? No

b) Can control charts be used when there is only assignable cause? No

c) Can control charts be used when there are trends in the mean? Only if the trends are steps or the trend is constant. Control charts depend on inference.

8) How do you measure a new process

Pre-Control pp707. Divide the specification range into green yellow and red zones. Red beyond the Tolerance Limits, yellow is from the one half point of the range from the nominal (center) to the Tolerance, green is the first half from the nominal point to a Tolerance Limit. If 5 consecutive parts do not fall within the green zone, stop the process and adjust. If all are in green, continue. If had to stop, adjust, measure first product, if yellow measure second, if second is yellow stop and adjust if green keep going, if any are red stop and adjust, if first is green keep going without measuring second and then collect enough to determine control limits using appropriate control charts and sampling plan.

9)What is acceptance sampling and should it be done, if so, when?

Acceptance Sampling plans, rectifying inspection (100% sampling to remove defectives), and kp

k = k1/k2 p > k inspect all items p < k, do not inspect p is proportion defective, k1 = cost of inspecting an unit, k2 = cost of allowing a defective unit (into the production system, Deming assumed that all finished product would be tested( expand to system and put a high cost on failure as customers would be discovering the failure) Assumption is that all lots come from the same population.

AQL (accepted quality level)

The acceptable quality level or AQL of a sampling plan is a measure of the level of quality routinely accepted by that sampling plan. It is defined as the percent defective (defectives per hundred units X 100 ) that the sampling plan will accept 95% of the time. This means lots at or better than the AQL are accepted at least 95% of the time and rejected at most 5% of the time. 

Table 2. AQL Sampling Table based on the Mil-STD-105D

|  |Acceptable Quality Levels for Normal Inspection (% defective) |

|Lot Size/ Sample |.01 |

|Size | |

LS=2 to 8

SS=2 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 | |LS=9 to 15

SS=3 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 | |LS=16 to 25

SS=5 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 | |LS=26 to 50

SS=8 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |1 |1 | |LS=51 to 90

SS=13 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |1 |1 | |91 to 150

SS=20 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |1 |2 | |151 to 280

SS=32 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |1 |2 |3 | |281 to 500

SS=50 |0 |0 |0 |0 |0 |0 |0 |0 |0 |0 |1 |2 |3 |4 | |501 to 1200

SS=80 |0 |0 |0 |0 |0 |0 |0 |0 |0 |1 |2 |3 |5 |7 | |1201 to 3.2K

SS=125 |0 |0 |0 |0 |0 |0 |0 |0 |1 |2 |3 |5 |7 |10 | |3201 to 10K

SS=200 |0 |0 |0 |0 |0 |0 |1 |1 |2 |3 |5 |7 |10 |14 | |10001-35K

SS=315 |0 |0 |0 |0 |0 |0 |1 |2 |3 |5 |7 |10 |14 |21 | |35001-150K

SS=500 |0 |0 |0 |1 |1 |1 |2 |3 |5 |7 |10 |14 |21 |- | |150001-500K

SS=800 |0 |0 |0 |1 |1 |2 |3 |5 |7 |10 |14 |21 |21 |- | |  |.01 |.015 |.025 |.04 |.065 |.10 |.15 |.25 |.40 |.65 |1.0 |1.5 |2.5 |4.0 | |                   

Copyright © 2001-2005 . All Rights Reserved.

1st and last product of a run

Different heuristics

9) What types of data are there? Continuous/Variable or Attribute/Discrete/Integer (+)

10) What charts should be used for what type of data? Answer: do tree pp 702

11) Know how to create each type of control chart. Examples

12) Know how to use each type of control chart. What does it all mean when to use each chart and how could you use multiple charts for one process.

13) What are proper control limits? Cost of type one vs. type two errors

14) What is a proper sample size? Depends on robustness of process, speed of process, cost of failure vs. inspection, cost of type one and two errors

15) How often should samples be taken? Depends on robustness of process, speed of process, cost of failure vs. inspection, cost of type one and two errors

16) Can, or should different types of charts be used for the same product? Yes and probably should be

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download