Intelligent identification of manufacturing operations using in-situ energy measurement in industrial injection moulding machines

Abstract

In energy-efficient and sustainable manufacturing systems, a good understanding of energy consumption in context of manufacturing operations is required. In this paper, we propose a generic framework for online identification of manufacturing operation states based on real-time energy data. Using Discrete Wavelet Transform (DWT) and modified universal threshold filter, time-series data is segmented by detecting stepwise changes. A two-stage Fuzzy C-Means (FCM) is then used to cluster extracted segments according to manufacturing operation states. As such, the online identification is carried out based on Euclidean distance of the incoming segments to cluster centroids. An implementation of our proposed framework on industrial injection moulding machines is presented with intensive analysis and discussion.