This document summarizes a presentation on using 7Epsilon methodology for in-process quality improvement and risk-based process optimization. It discusses how 7Epsilon incorporates risk-based thinking, as outlined in ISO 9001:2015, to prevent issues and enhance process performance. Key points include using sensor data and penalty matrices to determine optimal input ranges, embedding organizational process knowledge, and satisfying Clause 6.1 of ISO 9001:2015 to address risks and opportunities for improvement. The 7Epsilon approach analyzes multiple factors together to discover relationships and create new knowledge for continuous quality management.

1. 7Epsilon for ISO9001:2015
A Risk Based In-line Process Optimization Case Study for a
Connected Enterprise
Dr. Rajesh S. Ransing, Swansea University, UK
Dr. Meghana R. Ransing, p-matrix Ltd., UK
Presentation Id: TO5; Date: 15th November 2016; Time: 11-noon.
ASQ International Conference on Quality Standards,
November 14-15, 2016 Pittsburgh, USA

2. Learning Objectives
 Use in-process data and product specific process
knowledge for hypothesis generation (clause 6.1).
 Embed risk based thinking and organizational knowledge
within your organization.
 Learn 7Epsilon Vs Six Sigma.
 Understand 7Epsilon in ISO 9001:2015 context Process Inputs:

3. In-line Process Optimization: Case Study
 Objective:
Discover an optimal range for sensor values – if one exists.
 Process Output(s):
%Downtime for a line calculated hourly (or other defined time limit), OR
%Defective components produced in a batch.
 Process Inputs (Categorized as Sensors):
Data for all sensors and tags collected for an hour (or a defined time limit).
An ‘on-off’ switch - the number of times it is ‘on’ or ‘off’ in an hour.
Continuous or ordinal variables - a median, 5th and 95th centile values per
hour.
Categorical variables - one value for every process output value.

4. Product Specific Process Knowledge
Organisational process knowledge for a given product is
i. the actionable information;
ii. in form of optimal list of measurable factors and their
ranges
(Sensor 1 value 0.026 – 0.05; Sensor 2 value: 0.007 –
0.009);
iii. in order to meet desired business goals (process
outputs)
(e.g. minimize deviation from expected results, reduce %
downtime or rework time etc).

5. Risk Based Thinking
‘Risk Based Thinking’ is
the DNA of ISO9001:2015 & AS9100:2016
It’s not contained or isolated in one section or clause …

6. Risk and Opportunities Based Thinking
Clause 4.4: Quality management system and its processes
4.4.1 f: address the risks and opportunities as
determined in accordance with the requirements of 6.1
Clause 9 Performance Evaluation (and 9.3 Management Review)
9.3.2.e: the effectiveness of actions taken to address
risks and opportunities (see clause 6.1)
Clause 6.1 Actions to address risks and opportunities in order to
(i) give assurance that the QMS can achieve its intended
results,
(ii) enhance desirable effects,
(iii) prevent, or reduce undesired effects,
Clause 10 is Improvement.

7. What is Risk in Risk Based Thinking?
‘Risk’ in Risk Based Thinking should
• go beyond its usual interpretation in
Risk Assessment Forms,
• help trigger innovation,
• create new organisational knowledge.

8. What is Risk in Risk Based Thinking?
To understand ‘Risk’ in
ISO9001:2015 and
AS9100:2016 context,
we need to understand
‘Uncertainty’.

9. Uncertainty or Deficiency of Knowledge

10. Uncertainty or Deficiency of Knowledge

11. Uncertainty or Deficiency of Knowledge

12. What are your options for:
 Undertaking in-process quality improvement projects
 Implementing plant wide process optimization strategies
 Connected Enterprise, Industry 4.0, Big Data
 Satisfying requirements of the Clause 6.1of ISO9001:2015
• Prevent or Reduce Undesired (Negative) Effects
• Enhance Desired (Positive) Effects
OR
• Avoid BAD days
• Repeat GOOD days

13. Option 1: Plot Process Output Vs Process Input Graphs

14. Option 2: Embed Risk Based Thinking via Penalty Matrices

15. Option 2: Embed Risk Based Thinking via Penalty Matrices

16. Option 2: Embed Risk Based Thinking via Penalty Matrices

17. Option 2: Embed Risk Based Thinking via Penalty Matrices

18. Option 2: Embed Risk Based Thinking via Penalty Matrices

19. Option 2: Embed Risk Based Thinking via Penalty Matrices

20. Option 2: Embed Risk Based Thinking via Penalty Matrices

21. Penalty matrices using p-matrix software
 p-matrix software analyses hundreds and thousands of penalty
matrices among factors.
 Classifies factor settings as Optimal, Avoid or No Effect.
 Ranks them in order of importance.
 At end of analysis, a process engineer would get a list of top 15-20
matrices he/she needs to look at.
 Analyses discrete and continuous parameters together in one
analysis.
 Up to 200 factors and 10 responses can be analysed at once.
 The algorithm is scalable for Big Data and In-line Analysis.

22. Lets make a Solemn Promise …

23. Confirmation Trial Plan
Process
Inputs
Minimum
Value
Maximum
Value
Optimal Range Optimal Values
Sensor 1 0.019 0.05 Top 50% > 0.026 & < 0.05
Sensor 2 0.007 0.012 Bottom 50% > 0.007 & < 0.009
Sensor 3 6.204 6.527 Top 75% < 6.299 & < 6.527
Sensor 4 7.714 8.028 Bottom 50% < 7.714 & < 7.847

24. In-process Quality Improvement using 7Epsilon steps
What’s New?
Organisational knowledge management (Clause 7.1.6)
Actions to address risks and opportunities (Clause 6.1)
7Epsilon = Six Sigma + Industrial Internet of Things (IIoT)

25. Conclusions
 Risk - An effect of uncertainty (on an expected result)
 Uncertainty == deficiency of knowledge.
 Deficiency of knowledge are opportunities for creating
additional knowledge (e.g. Risk Based Tolerance Synthesis).
 The effect of uncertainty manifests itself as deviation(s) from
expected results.
 7Epsilon’s penalty matrix approach
 quantifies the effect of uncertainty by penalizing deviation
from desired response (Clause 6.1), and
 embeds risk based tolerance synthesis and creates new
organizational knowledge (Clause 7.1.6).

26. Further Reading
Ransing R.S., Batbooti R.S., Giannetti C, and Ransing M.R., “A quality correlation
algorithm for tolerance synthesis in manufacturing operations”, Computers and
Industrial Engineering, Volume 93, Pages 1–11, March 2016.
Ransing R. S., Giannetti, C., Ransing, M. R., & James, M. W. (2013). A coupled
penalty matrix approach and principal component based co-linearity index
technique to discover product specific foundry process knowledge from in-
process data in order to reduce defects. Computers in Industry, 64(5), 514-523.
Giannetti, C., Ransing, M. R., Ransing, R. S. et.al (2015) “Organisational
knowledge management for defect reduction and sustainable development in
the foundry industry”, International Journal of Knowledge and Systems Science
(IJKSS), 6(3), 18-37, July-September 2015.
http://www.slideshare.net/7Epsilon/as-91002016-and-iso-90012015-clause-
93-management-review-whats-new
http://www.slideshare.net/7Epsilon/key-changes-to-iso-9001-2015-and-7-
steps-of-7-epsilon

27. www.7epsilon.org
7Epsilon: In-process Optimization in
a Connected Enterprise Environment