Machine Learning-Based Analysis of Critical Process Parameters Influencing Product Quality Defects: A Real-World Case Study in Manufacturing

TL;DR

This paper demonstrates how machine learning models can predict and prevent manufacturing defects by analyzing process parameters, significantly improving quality control efficiency in a real-world industrial setting.

Contribution

It introduces a ML-based predictive quality control approach tailored for manufacturing, utilizing real-world data to identify defect-causing parameters.

Findings

ML models achieved high accuracy in defect prediction

Proactive quality control reduced defective products

Enhanced production efficiency and product quality

Abstract

Quality control is an essential operation in manufacturing, ensuring products meet the necessary standards of quality, safety, and reliability. Traditional methods, such as visual inspections, measurements, and statistical techniques, help meet these standards but are often time-consuming, costly, and reactive. With the advent of AI/ML, manufacturers can shift from reactive to proactive approaches in quality control. This study applies ML-based models for predictive quality control in a real-world manufacturing setting. The case company produces castings for powertrain components in heavy vehicles, where poor control of core-making process parameters leads to costly defects. ML models were developed by analyzing data from two core-making machines, their processes, and maintenance logs to identify parameters associated with casting defects, enabling the prediction and prevention of potential defects before they occur. The results demonstrated good accuracy rates, helping quality and production teams identify and eliminate defective cores and thereby improving product quality and production efficiency.