Machine Learning in Manufacturing: Quality 4.0 and the Zero Defects Vision
()
About this ebook
- Provides an understanding of the most relevant challenges posed to the application of Artificial Intelligence (AI) in manufacturing
- Includes analytical developments and applications and merges a quality vision with machine learning algorithms
- Features structured and unstructured data case studies to illustrate how to develop intelligent monitoring systems with the capacity to replace manual and visual tasks
Carlos A. Escobar
Dr. Carlos Alberto Escobar worked as a research scientist at the Amazon Last Mile Delivery and Technology organization and as a senior researcher at the Manufacturing Systems Research Lab of General Motors, Global Research and Development. He also worked as Faculty Aide at Harvard Extension School. Dr. Escobar obtained his Ph.D. in engineering sciences with a concentration in artificial intelligence (2019) and a master’s degree in quality engineering (2005) from Tecnológico de Monterrey. He also obtained a master’s in industrial engineering (2016) from New Mexico State University. He is an industrial engineer (2001) from Instituto Tecnológico de Ciudad Juarez. Currently, he studies a master’s in management (2024) at Harvard Extension School. He has published over 30 scientific articles in top journals. His research topic has been presented in top conferences, including the American Society of Quality. According to a published bibliometric study, he is considered one of the most cited and fruitful authors in Quality 4.0 (2022). The interest in his publications (2023) lies in the 99% at the Research Gate platform compared to his cohort of researcher registered in 2015. Dr. Escobar was recognized as the SHPE STAR of Today (2021) by the Society of Hispanic Professional Engineers, the largest association of Hispanic in STEM in the U.S. Dr. Escobar was in the Mexican national team of martial arts, he was inducted into the Hall of Fame of Ciudad Juarez (2015) after his retirement. Today, he enjoys teaching his colleagues this sport.
Related to Machine Learning in Manufacturing
Related ebooks
Artificial Intelligence for Process & Product Innovation Rating: 0 out of 5 stars0 ratingsApplications of Human-Computer Interaction and Robotics based on Artificial Intelligence Rating: 0 out of 5 stars0 ratingsIoT for Smart Operations in the Oil and Gas Industry: From Upstream to Downstream Rating: 0 out of 5 stars0 ratingsMoving To The Cloud: Developing Apps in the New World of Cloud Computing Rating: 5 out of 5 stars5/5Machine Learning with Noisy Labels: Definitions, Theory, Techniques and Solutions Rating: 0 out of 5 stars0 ratingsDesign of Experiments for Engineers and Scientists Rating: 0 out of 5 stars0 ratingsArtificial Intelligence and Software Testing: Building systems you can trust Rating: 0 out of 5 stars0 ratingsDeveloping Structured Procedural and Methodological Engineering Designs: Applied Industrial Engineering Tools Rating: 0 out of 5 stars0 ratingsDesign for Additive Manufacturing Rating: 0 out of 5 stars0 ratingsPractical Machine Learning for Data Analysis Using Python Rating: 0 out of 5 stars0 ratingsComprehensive Guide to Implementing Data Science and Analytics: Tips, Recommendations, and Strategies for Success Rating: 0 out of 5 stars0 ratingsPractical Machine Learning with Python: A Problem-Solver's Guide to Building Real-World Intelligent Systems Rating: 0 out of 5 stars0 ratingsMethodology for the development of systems based on learning objects Rating: 0 out of 5 stars0 ratingsBuilding Big Data Applications Rating: 0 out of 5 stars0 ratingsXAI Based Intelligent Systems for Society 5.0 Rating: 0 out of 5 stars0 ratingsAI for Healthcare with Keras and Tensorflow 2.0: Design, Develop, and Deploy Machine Learning Models Using Healthcare Data Rating: 0 out of 5 stars0 ratingsDeep Learning: Convergence to Big Data Analytics Rating: 0 out of 5 stars0 ratingsUnlock Lucrative Opportunities with Modern Technology Skills: A Comprehensive Guide to Earning Big Money Rating: 0 out of 5 stars0 ratingsUncertainty in Data Envelopment Analysis: Fuzzy and Belief Degree-Based Uncertainties Rating: 0 out of 5 stars0 ratingsNavigating the Future: Ten Essential Considerations for Thriving in 2023 and Beyond Rating: 0 out of 5 stars0 ratingsMachine Learning for Finance Rating: 0 out of 5 stars0 ratingsOverview of Industrial Process Automation Rating: 4 out of 5 stars4/5Input-Output Models for Sustainable Industrial Systems: Implementation Using LINGO Rating: 0 out of 5 stars0 ratingsAutomated Network Technology: The Changing Boundaries of Expert Systems Rating: 0 out of 5 stars0 ratingsContextual Design: Design for Life Rating: 4 out of 5 stars4/5Decision-Making Techniques for Autonomous Vehicles Rating: 0 out of 5 stars0 ratingsSchaum's Outline of Principles of Computer Science Rating: 0 out of 5 stars0 ratingsProgramming with STM32: Getting Started with the Nucleo Board and C/C++ Rating: 3 out of 5 stars3/5
Automotive For You
How to Drive: Real World Instruction and Advice from Hollywood's Top Driver Rating: 4 out of 5 stars4/5Car Basics Rating: 5 out of 5 stars5/5California Driver’s Practice Tests: DMV Practice Tests Rating: 0 out of 5 stars0 ratingsIt Came from the Garage! Rating: 4 out of 5 stars4/5The Official Highway Code: DVSA Safe Driving for Life Series Rating: 3 out of 5 stars3/5How to Rebuild GM LS-Series Engines Rating: 0 out of 5 stars0 ratingsHow to Build a Car: The Autobiography of the World’s Greatest Formula 1 Designer Rating: 4 out of 5 stars4/5Girls Auto Clinic Glove Box Guide Rating: 4 out of 5 stars4/5Automotive Wiring and Electrical Systems Rating: 4 out of 5 stars4/5How to Rebuild VW Air-Cooled Engines: 1961-2003 Rating: 5 out of 5 stars5/5Let's Build A Camper Rating: 4 out of 5 stars4/5GM 4L80E Transmissions: How to Rebuild & Modify: How to Rebuild & Modify Rating: 5 out of 5 stars5/5Auto Repair For Dummies Rating: 5 out of 5 stars5/5GM Turbo 350 Transmissions: How to Rebuild and Modify Rating: 0 out of 5 stars0 ratingsWelding for Beginners in Fabrication Rating: 4 out of 5 stars4/5How to Rebuild Any Automotive Engine Rating: 0 out of 5 stars0 ratingsHow to Rebuild the Small-Block Mopar Rating: 5 out of 5 stars5/5The Nuts and Bolts of NASCAR: The Definitive Viewers' Guide to Big-Time Stock Car Auto Racing Rating: 5 out of 5 stars5/5David Vizard's How to Build Max Performance Chevy Small Blocks on a Budget Rating: 0 out of 5 stars0 ratingsCarTalks: Car Basics Rating: 4 out of 5 stars4/5The Reckoning Rating: 4 out of 5 stars4/5Lemon-Aid New and Used Cars and Trucks 2007–2018 Rating: 0 out of 5 stars0 ratingsLearn To Ride A Motorcycle In Five Minutes Rating: 0 out of 5 stars0 ratingsRVs & Campers For Dummies Rating: 0 out of 5 stars0 ratings
Reviews for Machine Learning in Manufacturing
0 ratings0 reviews
Book preview
Machine Learning in Manufacturing - Carlos A. Escobar
Preface
Suppose you have been drawn to this book. In that case, you are most likely aware of the extraordinary potential of using Artificial Intelligence (AI) in manufacturing. Still, you must also be well aware of the challenges and concerns posed by this technology. I, Carlos, worked in the Manufacturing Systems Research Lab of General Motors for more than 6 years. My coauthor and mentor, Ruben Morales-Menendez, PhD, is a Full Professor at Tecnológico de Monterrey. We have developed, published, and implemented novel technologies to solve real problems and create value. In this book, we describe our experiences using AI in the manufacturing sector; we discuss the challenges we have faced and share our vision. Before moving forward, we would like to thank José Antonio Cantoral Ceballos, PhD, for his contributions to chapter seven, nine, and overall feedback to our book, which led to a significant improvement. Dr. Cantoral is a Research Professor at the Tecnológico de Monterrey, Computer Science Department; his research focuses on Deep Learning solutions to different problems, particularly the study of neurological signals. We also thank the Tecnológico de Monterrey for supporting this book.
Today, the industrialization of artificial intelligence is a significant megatrend and its application for quality control is one of the most cited priorities. Therefore, manufacturing companies can competitively position themselves among the most advanced and influential companies by successfully implementing Quality 4.0, providing opportunities for qualified professionals to become a leading force in the industry. However, this is not an easy task because quality management leaders often need help developing a vision for Quality 4.0; they need to learn new technologies and paradigms to keep innovating and achieving professional growth. In this book, we focus on applying machine learning techniques to process-derived data for monitoring, controlling, predicting, and improving the quality of discrete manufacturing systems. We have made a conscious effort to keep the math and coding aspects at easily understandable levels for most engineers. Thus, it is possible to focus more on the practical aspects of selecting and solving intractable engineering problems through machine learning. However, for an in-depth perspective of the mathematics and coding details of machine learning, we recommend the following books: Machine Learning: A Probabilistic Perspective (2012) by Kevin P. Murphy, An Introduction to Statistical Learning: with Applications in R (2021) by Gareth James, Daniela Witten, et al., and Introduction to Machine Learning with Python: A Guide for Data Scientists (2016) by Andreas C. Müller and Sarah Guido.
This book presents a Quality 4.0 initiative developed in the Manufacturing Systems Research Lab of General Motors and supported by a novel problem-solving strategy, which evolved from the traditional Six Sigma cycle of design, measure, analyze, improve, and control. We thank Dr. Jeffrey A. Abell, Dr. Jorge Arinez, Dr. Debejyo Chakraborty, Dr. Megan McGovern, and Dr. Michael A. Wincek for supporting the development of this vision; our papers are cited across the book.
We review the theoretical background of Quality 4.0 and describe its concrete applications and new quality control paradigms. We also present several case studies to illustrate the main concepts. This book enables engineers to develop intelligent predictive systems, i.e., Learning Quality Control (LQC) systems. Also, managers without any AI background can learn to identify valuable business projects and directors can successfully deploy a business vision for Quality 4.0.
Carlos A. Escobar, PhD
Chapter 1: Introduction
Abstract
The manufacturing industry plays a predominant role in boosting a country's economy. Smart manufacturing has ushered in a new era of using technological innovations in the manufacturing process; the processes involved exhibit rapidly increasing complexities. Many of the founding techniques and paradigms of traditional quality control (QC) methods are not able to handle all these dynamics. Therefore, in the last decade, traditional quality management philosophies have plateaued, and QC professionals started stagnating with little innovation to offer. Today, the industrialization of artificial intelligence (AI) is a megatrend that dominates the business landscape, and it requires the attention of industry managers. The application of AI to manufacturing systems for QC and improvement drastically improves upon the zero-defects vision proposed by Crosby. Consequently, the industrialization of AI offers an excellent opportunity for quality professionals to return to their lead roles in the manufacturing sector.
Keywords
Evolution of quality control; Importance of manufacturing; Limitations of traditional methods; Quality 4.0; Smart manufacturing
This chapter is organized as follows (Fig. 1.1). The manufacturing industry has a key role in the economy of a country (Section 1.1); smart manufacturing (SM) has been characterized using technological innovations in manufacturing processes, increasing the level of complexity. Manufacturing companies are leaders in innovation, productivity, and research and development; many technological advances have originated in the manufacturing industry. Producing high-quality products is a top priority for the manufacturing industry because it improves customer retention, builds brand trust, boosts return-on-investments, enabling business growth, etc (Section 1.2).
The introduction of AI into the manufacturing industry improves the zero-defects vision promoted more than decades ago. The technologies of the fourth industrial revolution (I4.0) are moving forward the frontiers of manufacturing sciences. Innovation comes from the cognitive computing capabilities that drive technologies such as AI, industrial internet of things, cloud storage and computing, etc; their combination enables a smart and connected manufacturing environment (Subsection 1.2.1).
Figure 1.1 Contents of the Chapter 1.
(Subsection 1.2.2) The most important building blocks of SM are the cyber-physical systems, data-driven approaches, real-time iterations, self-learning adaptations, and executions. Data-driven approaches are also part of the building blocks of SM; the most relevant applications based on data-driven are: detection, prediction, automation, optimization, and augment human intelligence. There are eight relevant SM characteristics with respect to business and tactics, namely, flat, sustainable, agile, people-oriented, profitable, innovative, current, and competitive (Subsection 1.2.3). SM has emerged as a compelling topic (i.e., cybersecurity, maintenance 4.0, hyperautomation, smart scheduling, smart design, and mass customization) for research (Subsection 1.2.4), preserve their competitiveness in competitive markets.
Lately, most mature organizations have merged the traditional quality philosophies and statistical techniques to create high conformance production environments as a part of an evolution of the problem-solving strategy in modern quality movement in manufacturing (Section 1.3). Traditional quality philosophies based on statistics have raised manufacturing standards to very high conformance levels, but they are limited in addressing the challenges posed by SM (Section 1.4). Quality 4.0 (Q4.0) is the next natural step in the evolution of quality (Section 1.5); this book presents a Q4.0 initiative aimed at driving AI-based innovation, the application of AI to manufacturing systems for QC, and improvement drastically improves upon the zero-defects vision.
1.1. Motivation
The manufacturing industry plays a predominant role in boosting a country's economy. The added value and employment contribution to the global Gross Domestic Product (GDP) have not changed significantly since 1970 in developing countries [124]. At present, manufacturing is the driving economic force of the most advanced countries [250]. The global market for manufacturing is forecasted to grow from US $649.8 billion in 2020 to US $732.2 billion in 2027 [108]. The nations with high manufacturing output [310] also have the highest GDP [142]. Manufacturing increases the living standards by increasing the purchasing power of the people living in industrialized societies. Advanced manufacturing companies are leaders in innovation, productivity, exports, and research and development. Most technological advances have originated in these companies [321].
In the United States, for every dollar of domestic manufacturing value-added, another $3.60 of economic activity is generated elsewhere across the economy and for every manufacturing job; there are 3.4 jobs created in nonmanufacturing industries. No other sector comes close to these numbers, [111]. Today, China is the world's largest manufacturing economy, and it is considered one of the most competitive nations in the world; this sector helped China to rise as a global economic superpower [97, 170]. In today's global competitive market, delivering high-quality products is a top priority for most manufacturing companies. High quality improves customer retention, builds brand trust, and boosts return-on-investments, enabling business growth [47].
Quality has been defined variously by researchers and experts. The American Society of Quality (ASQ) defines a quality item as a product or service that is free of deficiencies [9]. Joseph Juran defines quality as fitness for use [199].
According to Joseph Juran, quality begins by understanding who the customers are and how and why they would use a product. This information is then used to drive all improvement activities and develop a customer-focused business strategy. A more extensive definition of quality considers eight dimensions: (1) performance, (2) features, (3) reliability, (4) conformance, (5) durability, (6) serviceability, (7) aesthetics, and (8) perceived quality [104]. Philip Crosby states that good, bad, high, and low quality are relative concepts, and the meaning of quality is conformance to requirements. For the ideas conveyed in this book, Crosby's definition of quality was the most appropriate.
Crosby also introduced the concept of zero defects in manufacturing [60]; it was one of the four absolutes of quality management [61]. However, when Crosby proposed the idea of zero defects in 1980, the manufacturing process was labor-intensive and depended largely on the skills of human operators. Because of the technological limitations of the time, the zero-defects concept remained mainly a managerial tool that acknowledged the importance of quality and motivated employees to do their best to reach this goal. The creation of unrealistic standards was criticized by E. Deming and J. Juran [25, 148], Eliminate slogans, exhortations, and targets for the work force asking for zero defects and new levels of productivity.
This is clearly aimed at zero defects. In this context, Crosby stated, It is merely setting performance standards that no one can misunderstand [60].
Broadly speaking, quality systems are divided into three categories: quality control (QC), quality assurance (QA), and quality improvement (QI). QC is the process of applying statistical and analytical techniques to determine if a manufactured item conforms to the design specifications. According to ASQ, QC involves operational techniques and activities used to fulfill quality requirements. QA refers to all the planned and systematic activities implemented within the quality system that can be demonstrated to provide confidence that a product or service will fulfill the quality requirements [9]. QI is the process of constantly identifying projects aimed at creating breakthrough levels of performance by eliminating defects. QI requires a good managerial team that can identify relevant projects, create a successful team, and allocate the required resources. The three synergized systems (QC, QA, and QI) have helped create a high conformance production environment. Therefore, in today's manufacturing world, most mature companies operate their processes at very low defects per million opportunities (DPMOs).
Modern manufacturing systems have ensured that defects are rarely generated; however, quality standards need to be further improved because customers expect perfect quality. Intense global competition has led to low profit margins [150, 77]; therefore, warranties can make the difference between profit and loss [38]. Moreover, customers now use the Internet and social media tools (e.g., Google product review, YouTube, Facebook, and Instagram) to share their product experiences, which can potentially go viral in a couple of days or even hours; this leaves organizations with little flexibility to recover from their mistakes [264]. Thus, a single negative customer experience can immediately affect a company's reputation and can influence loyal customers [140]. Today, the introduction of AI into the manufacturing process vastly improves upon the zero-defects vision promoted by Crosby more than four decades ago. This book presents a Q4.0 initiative aimed at driving AI-based innovation. Merriam-Webster defines initiative as the power or opportunity to do something before others do it. They also define it as a plan or program that is intended to solve a problem. This initiative improves upon the zero-defects vision from a technological