Quality 4.0 & the Future of Manufacturing: It's All About the Data

1. Quality 4.0 & the Future of Manufacturing: It's All About the Data Nicole Radziwill INTELEX Email: nicole.radziwill@gmail.com - Twitter: @nicoleradziwill

2. Medvedev, G. (1991). The truth about Chernobyl. IB Tauris.

3. “You reported to Kiev on 26 April that the radiation situation at the plant in Pripyat was within normal limits?” – “Right. That’s what the instruments we had at the time showed.” p. 226 Medvedev, G. (1991). The truth about Chernobyl. IB Tauris.

4. What is being measured & how Context of the measurement Knowledge of the instrument

5. What: Exposure to ionizing radiation - 3.6 R/hour Context: 800-1500 R/hour near exposed reactor; 300 R accumulated for radiation sickness Instrument: 3.6 R/hour is the maximum measurement on the standard device used at those facilities at the time

6. Every Data Point is Important

7. Ironmaking (ore+lime+coke  pig iron) • Steelmaking (pig iron  liquid steel) • Casting • Rolling

8. Video: https://www.youtube.com/watch?v=NlAoaajyjpM (2:48)

9. Video: https://www.youtube.com/watch?v=-RYCXDUt2m8 (1:42)

10. Temperature sensors Gas pressure monitors Wind rate Water ingress detection Permeability Tuyere Kinetic Energy Flame Temperature Bosh Gas Descending Burden Hot Metal %Si Operator uses HMIs to monitor process and product quality in the SCADA system & perform interventions

12. What Can Go Wrong? Uncontrolled cooling Water leak (explosive) Incorrect amount of gas introduced to reaction Temperature too high Temperature too low Carbon monoxide release Slag solidifying From https://insights.globalspec.com/article/7809/high-performance-slag-materials-a-steel-industry-byproduct

13. Turning Off a Blast Furnace Fanning (capacity) Backdrafting (repair) Banking (short outages) Blowing out (end of life) Blowing down (inspection) “Salamander tipping” (after blow-down) Regular shutdowns are rare Unplanned shutdowns have serious consequences From https://insights.globalspec.com/article/7809/high-performance-slag-materials-a-steel-industry-byproduct

14. 1. 2. 3. Loss of Product – Loss of Assets – Loss of $ - Loss of Life Forensics Report Revealed…

15. Supply chain info “Smart Manufacturing” INTERVENTION? YES OR NO ERP Data Current State vs. Nominal State • Root Cause: • Software bug? • Hardware issue? • Process out of control? • Human error? • Cyberattack? QMS/SPC Process Data RPA data Telemetry data M2M data Data Science & Machine Learning Network logs, SIEM, cameras This is a data-intensive real-time classification problem BLOCKCHAIN

16. Unfortunately, Both Sides Can Play From https://arxiv.org/pdf/1709.06397.pdf

17. “… a leading car manufacturer… was experiencing recurrent faults, and therefore high warranty costs, on the rear tail-lights. The engineers had been looking at the rear of the vehicle for the answer (and not succeeding) however the software quickly found the root cause of the fault to be located in the roof of the vehicle, a part of the car that had not even been investigated as a possible source. Sometimes prior experience, or being too close to a problem, can inhibit a solution if an old hypothesis is applied to a new problem.” From http://www.qmtmag.com/display_eds.cfm?edno=9665074

18. Augment (or improve upon) human intelligence Increase the speed and enhance the quality of decision-making Improve transparency, traceability, and auditability Anticipate changes, reveal biases, adapt to new circumstances/ data sources Reveal opportunities for continuous improvement Learn how to learn: cultivate self-awareness and other-awareness How Do We Get There?

19. Quality through the Lens of Innovation http://asq.org/future-of-quality/ https://www.yumpu.com/en/document/read/38080390/asq-innovation-think-tank-executive-summary

20. 1: The Future of Leadership FROM EFFICIENCY to ADAPTABILITY TRADITIONAL EMERGING

21. 2: The Future of Customer Experience CONTINUOUS VOICE OF THE CUSTOMER

22. 3: The Future of Manufacturing OPERATIONS & SUPPLY CHAIN OMNISCIENCE From Wang, J., Ma, Y., Zhang, L., Gao, R. X., & Wu, D. (2018). Deep learning for smart manufacturing: Methods and applications. Journal of Manufacturing Systems.

23. 4: The Future of Education FEWER BOUNDARIES, MORE DATA CONSCIOUSNESS From Kovach, J. V., & Fredendall, L. D. (2013). The influence of continuous improvement practices on learning: An empirical study. The Quality Management Journal, 20(4), 6.

24. 5: The Future of Energy CONSUMERS BECOME PRODUCERS OLD NEW

25. Quality 4.0 = C I A Connectedness Intelligence Automation for improving performance Radziwill, N. M. (2018, October). Let’s Get Digital: The many ways the fourth industrial revolution is reshaping the way we think about quality. Quality Progress, p. 24-29. http://qualityprogress.com

26. Information is only useful when it’s valid, and you are Connected to it Automation helps bring it to you when you need it, and frees up time and effort Intelligence helps you understand and respond

27. Cyber-physical systems produce Big Data • Network infrastructure is robust; data can be shared for real-time decision support • Software libraries for advanced analytics are accessible, comprehensive, and reliable • People, machines, and data all connected in near real-time

28. Lots of it (volume) It’s coming at you fast (velocity) Different formats and sampling frequencies (variety) Huge variations in data quality (veracity) Different people/organizations produce it or own it (governance) It could easily change or disappear (control) There may be restrictions on how you use it (policy) What is Big Data? “… Big Data is anything bigger or more complex than what your organization is currently prepared to handle.” -- one of the world’s experts on cyberinfrastructure for Big Data at a National Science Foundation panel (2013)

29. Opportunity Capture Quality Maturity Profit Profit Profit External Failure Internal Failure Appraisal Error- and waste-free service Value Vision for Quality 4.0 Siloed/ Struggling Adapted from Carlson, R. O., Amirahmadi, F., & Hernandez, J. S. (2012). A primer on the cost of quality for improvement of laboratory and pathology specimen processes. American journal of clinical pathology, 138(3), 347-354.

30. Quality as: 1.0 - Inspection 2.0 - Design 3.0 - Engagement 4.0 - Discovery When people learn, they update: POLICIES PROCEDURES PRACTICES HEURISTICS When AI/ML algorithms learn, they update: PREDICTIONS KEY PREDICTORS CLASSIFICATIONS PATTERN IDENTITIES OPTIMAL PATHS AI/ML evolves data-driven decision making to be self-aware & adapt to changing environments, circumstances, and customer/stakeholder needs

31. Artificial Intelligence Machine Learning Neural Networks Deep Learning Blockchain Big Data Enabling Tech/Cybersecurity Statistics & Data Science

32. Solid Process-Driven Management System Transparent Data Architecture Cybersecurity as a Habit Strategy for Discovery Shorten Loops

33. Questions? Nicole Radziwill – Text: (703) 835-6336 Email: nicole.radziwill@gmail.com https://www.linkedin.com/in/nicoleradziwill/

35. Muffins

36. Prediction Classification Pattern Identification Data Reduction Anomaly Detection Pathfinding + = 99.8% chance of blueberry muffin 96.6% chance of not blueberry muffin

37. Deep Learning = Building a stack of components to process inputs Deep Learning Optics