The concept of Quality 4.0 is a new approach to manufacturing that leverages Industrial big data (IBD), Industrial Internet of Things (IIoT), and artificial intelligence (AI) to monitor, control, and improve manufacturing quality. It is the fourth wave in the quality movement and is based on empirical learning, empirical knowledge discovery, and real-time data generation, collection, and analysis.
Various authors have provided their own interpretations of what Quality 4.0 means, but the emphasis is generally on technological innovations and connectivity. Successful implementation is one of the most pressing challenges confronting Industry 4.0, with surveys showing that 80%-87% of data science projects never reach production. Quality 4.0 requires a problem-solving strategy to combine the supporting technologies in a meaningful way.
To address this challenge, a scientific paper proposes a seven-step problem-solving strategy that guides the implementation of the PMQ initiative - Identify, Acsensorize, Discover, Learn, Predict, Redesign and Relearn (IADLPR). The paper discusses the desirable characteristics of machine learning projects and the importance of selecting the right project to drive the company's success in deploying AI.
The paper also highlights the significance of sensorization as a driver of Industry 4.0 and the use of smart sensors to achieve quality standards. The paper further explains the role of machine learning algorithms (MLA) in the structure of AI and the importance of selecting the appropriate MLA for solving complicated AI problems.
In this scientific paper, the focus is on validating machine learning algorithms (MLA) for effective pattern recognition. The three basic types of MLAs are supervised, unsupervised, and reinforcement learning. Each MLA has inherent biases, so it is important to choose the appropriate algorithm for a specific dataset.
The paper demonstrates that there is a trade-off between minimizing type-I and type-II errors in classification, and reducing one type of error increases the other. Edge analytics is also an important tool for processing data close to the source, enabling predictive systems to be implemented in the field. However, there are challenges in terms of the V's (volume, velocity, variety, veracity, and variability) of manufacturing systems, and vigilance must be applied in the learning and forgetting process.
In complex manufacturing processes, traditional inspection approaches are not 100% reliable and can result in false positive and false negative errors. To overcome this, a data-driven method called PMQ can be used to develop a quality control system. This eliminates the need for manual inspections and creates an empirical-based QC system.
The paper provides a seven-step problem-solving strategy for PMQ, which includes identifying potential projects, observing the process, creating training data, designing the classifier, developing a prediction system, redesigning the process based on data mining results, and developing a relearning strategy for the classifier. The paper includes a case study of the Ultrasonic Welding of Battery Tabs (UWBT) where PMQ was used to develop a quality control system for the process.
The article discusses the challenges that companies face when implementing AI and data science projects in manufacturing, as these projects can often fail to deliver value despite significant investments. To address this issue, the article proposes a seven-step problem-solving strategy called IADLPR2 that can help managers identify when to drop a project and assign a new one to the data science team.
The strategy involves identifying learning targets, deploying relevant sensors, generating signals with discriminative information, applying machine learning algorithms, optimizing predictions, conducting statistical analyses, and relearning. The article also discusses the Quality 4.0 initiative, which focuses on real-time defect detection and emphasizes the importance of having key performance indicators, high-quality data, a universal data standard, a reporting system, technological and analytical tools, and predictive analytical tools.
The successful implementation of Quality 4.0 can lead to improved product quality, reduced waste and cost, increased customer satisfaction, and improved financial results. The article concludes with a case study from General Motors and suggests that future research could focus on adapting the seven-step problem-solving strategy for deep learning applications.
In conclusion, Quality 4.0 is a new approach to manufacturing that leverages the latest technological innovations to improve manufacturing quality. The successful implementation of Quality 4.0 requires a problem-solving strategy to combine the supporting technologies in a meaningful way.
The seven-step problem-solving strategies proposed in the scientific papers and articles mentioned above offer a practical roadmap for companies looking to implement Quality 4.0. By leveraging Industrial big data (IBD), Industrial Internet of Things (IIoT), and artificial intelligence (AI), companies can monitor, control, and improve manufacturing quality, leading to improved product quality, reduced waste and cost, increased customer satisfaction, and improved financial results.
Authors
Escobar, C. A., Macias, D., McGovern, M., Hernandez-de-Menendez, M., & Morales-Menendez, R. (2022). Quality 4.0–an evolution of Six Sigma DMAIC. International Journal of Lean Six Sigma, (ahead-of-print). https://doi.org/10.1108/IJLSS-05-2021-0091