Yarn Breakage Prediction using ML in Spinning Mills

Yarn breakage during textile manufacturing processes such as weaving and spinning represents a critical challenge that significantly impacts production efficiency, product quality, and operational costs. This paper provides a comprehensive review of yarn breakage prediction methodologies, tracing the evolution from empirical and statistical approaches to contemporary machine learning techniques. We examine the fundamental mechanisms of yarn breakage, including the complex stresses that warp and weft yarns undergo during weaving processes. The review synthesizes research on traditional prediction methods, artificial neural network applications, and recent advances in automated machine learning and explainable artificial intelligence for yarn breakage forecasting. A detailed analysis of a pioneering 2022 study demonstrates how Industry 4.0 concepts, Internet of Things data, and Auto-ML tools can achieve predictive performance with R² values up to 0.73 for weft break prediction. The paper also discusses practical implications for textile manufacturers, including preventive measures enabled by accurate breakage predictions, and identifies future research directions in this rapidly evolving field.