Application of Basic Machine-Learning Classifiers for Automatic Anomaly Detection in Shewhart Control Charts

Aug 4, 2026

Upload by Levi

Modern manufacturing relies on stable process monitoring, yet traditional Shewhart control-chart interpretation can suffer from human limitations such as operator fatigue, causing delayed responses or mistaken readings of deviations. This study evaluates an AI-based approach to reduce such quality-control errors in rim production by preparing training data, training machine-learning classifiers, and assessing their real-time effectiveness on control charts. Results indicate that machine-learning classifiers can materially support quality controllers through improved supervision and more reliable pattern analysis.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Introduction

Motivation and background

Problem with manual interpretation

Purpose and research approach

Decision Making in Manufacturing anD services Vol . 18 • 2024 • pp . 83–98 Application of Basic Machine-Learning Classifiers for Automatic Anomaly Detection in Shewhart Control Charts Aleksander Woźniak* , Klaudia Krawiec** , Roger Książek*** Abstract. In today’s dynamic technological environment, innovation plays a crucial role – especially for manufacturing enterprises that constantly strive to improve the quality of their products. This article examines the quality-management issue in a company producing carrims. It was identified that real-time quality control can sometimes be unreliable due to controller fatigue, leading to erroneous data interpretation or delayed responses to deviationsin the production process. The study aimed to investigate the possibility of eliminating or significantly reducing these errors by employing a tool that is based on artificial intelligence. The article covers the preparation of training data, the training of classifiers, and the evaluation of their effectiveness in analyzing control charts in real time. The adopted hypothesis assumes that machine-learning classifiers can be effective methods of support for quality controllers. The research began with collecting measurement data from the machine and dividing it into training and test sets. The obtained results were evaluated using standard quality measures for machine-learning models. The results showed that the use of artificial intelligence can bring significant benefits in improving quality supervision in the production process of car rims. Keywords: Machine Learning, Artificial Intelligence, AI, Statistical Process Control, SPC, Quality Control, Classifiers, Quality Metrics, Python Programming, Car Wheel, Quality Issues Mathematics Subject Classification: 68T10 JEL Classification: C44 Submitted: June 5, 2024 Revised: October 28, 2024 © 2024 Authors. This is an open-access publication that can be used, distributed, and reproduced in any medium according to the Creative Commons CC-BY 4.0 license. This license requires that the original work was properly cited. * AGH University of Krakow, Faculty of Energy and Fuels, Krakow, Poland, e-mail: alekwozniak@stu[dent.agh.edu.pl](dent.agh.edu.pl) ** AGH University of Krakow, Faculty of Management, Krakow, Poland, e-mail: klakrawiec@student.agh. [edu.pl](edu.pl) *** AGH University of Krakow, Faculty of Management, Krakow, Poland, e-mail: [roger@agh.edu.pl](roger@agh.edu.pl) DOI: [https://doi.org/10.7494/dmms.2024.18.6345](https://doi.org/10.7494/dmms.2024.18.6345) 83 1. INTRODUCTION The origins of Shewhart control charts can be traced back to the 1920s, when they were first developed for quality control. These represent a statistical tool that monitors process stability by tracking sequence samples and identifying patterns that deviate from expected norms. By distinguishing between common-cause and special-cause variations, this method enables the early detection of shifts in production, thus ensuring consistent quality control over time. Shewhart control charts have been employed for qualitative analysis, becoming a key tool for collecting data during production and later analyzing it (Shewhart, 1926) . It needs to be stressed that Shewhart control charts remain in use even today; however, their roles have evolved with modern digital tools. This shift has introduced several important considerations regarding the role of AI in quality control. It remains uncertain whether AI-based systems can fully replace human operators or if they will function more effectively as supplementary tools. The potential of AI to improve the accuracy and precision of measurements in rim production is significant; however, challenges and limitations come with its implementation. The effectiveness of different machine-learning algorithms in identifying and eliminating measurement errors must also be evaluated, along with the necessary data for adequately training AI models. Addressing t

FAQs

Why can Shewhart control charts become unreliable in real-time quality control?

Real-time interpretation may be affected by controller or operator fatigue, which can lead to erroneous data interpretation or delayed responses when process deviations occur.

What is the main goal of this research?

To investigate whether AI-based tools using basic machine-learning classifiers can eliminate or significantly reduce errors in analyzing Shewhart control charts in real time.

How are the machine-learning models evaluated in the study?

Measurement data are collected from the production process, split into training and test sets, and the trained classifiers are assessed using standard quality metrics for machine-learning models.

Aug 4, 2026

Upload by Levi