Recognition and Segmentation of Heat Treatment Processes Based on PLR

In industrial processing, heat treatment encompasses various techniques such as annealing, tempering, quenching, tempering, surface hardening, and more. These diverse processes impart different characteristics to metallic materials, catering to the specific requirements of subsequent production stages. As each technique corresponds to a distinct temperature range, temperature serves as a pivotal process parameter, exerting a decisive influence on the quality of manufacturing equipment. This study focuses on the analysis of temperature time series data in the context of heat treatment. We propose an PLR algorithm based on key Point and Bottom-Up merging (KP-BU PLR) to partition different process stages and extract corresponding features. Simultaneously, we address the issue of anomalous fluctuations in temperature time series data encountered in practical production processes. To tackle this challenge, we introduce a sliding window-based approach for anomaly detection and handling, enhancing the information density of production data, while ensuring the stability of the KP-BU algorithm. Finally, the feasibility and effectiveness of the proposed methods are validated through experimental verification using real production data.