Experiment Description
Statistical process control, traditionally used for quality control in various industries, was effectively implemented in an experiment focusing on both post-process and in-process data analysis. This control method, crucial for detecting process shifts and proposing corrective actions, typically relies on limited sampled data, potentially causing delays in identifying problems and additional costs. However, with advancements in ICT, the experiment leveraged the increased availability of in-process data, enabling immediate detection of issues.
Traditionally, solutions in this area were based on univariate methods with certain assumptions, but the increasing complexity of data rendered this approach less effective. The experiment successfully addressed this by applying AI and machine learning principles, utilizing multivariate approaches made feasible by the larger datasets available. A specific process was defined for control, and AI monitoring methods were tailored to the situation, implemented, and preliminarily assessed using historical data.
This new approach was then presented to representatives of selected SMEs. Potential overlaps were discussed, and proposals for potential use-case projects were defined. The experiment not only highlighted the benefits of AI in-process monitoring and control but also addressed the challenges, potential pitfalls, and barriers, offering a comprehensive view of AI application in statistical process control.
List Title
Impact
The experiment successfully guided Czech companies towards innovative AI-based methods for monitoring and control of their processes, thereby enhancing the competitiveness of not only the Czech industry but also other sectors. Initially aimed at improving industrial manufacturing, the application of Statistical Process Control (SPC) extended to diverse areas such as healthcare, environmental sectors, and general services. In instances where AI deployment was challenging due to inadequate data infrastructure, the experiment played a pivotal role in identifying and describing these barriers, motivating SMEs to address them, for instance, by developing sophisticated data platforms and establishing related processes.
As a result of this experiment, a notable number of SMEs began considering and discussing the potential use of AI, machine learning, and data science in process control, monitoring, and data-driven quality management systems. The initiative successfully motivated at least two SMEs to start preparing their data infrastructure for future AI-based quality control and monitoring. Furthermore, it supported at least one SME in developing a proof of concept for AI-based process control, marking a significant stride in the practical application of AI in various business operations.