Open AccessMACHINE LEARNING BASED IDENTIFICATION OF ENERGY EFFICIENCY MEASURES FOR MACHINE TOOLS USING LOAD PROFILES AND MACHINE SPECIFIC META DATA
Author(s) -
L. Petruschke,
G. Elserafi,
Borys Ioshchikhes,
Matthias Weigold
Publication year - 2021
Publication title -
mm science journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.195
H-Index - 10
eISSN - 1805-0476
pISSN - 1803-1269
DOI - 10.17973/mmsj.2021_11_2021153
Subject(s) - computer science , machine learning , identification (biology) , artificial intelligence , energy (signal processing) , machine tool , efficient energy use , prioritization , control (management) , data mining , feature selection , engineering , statistics , mathematics , mechanical engineering , botany , electrical engineering , management science , biology
Approaches to detect energy efficiency measures are associated with time consuming analysis requiring expertise. Against this background, this paper presents an expert system to identify potentials for improving the energy efficiency of metal cutting machine tools based on measurement and meta data of 35 machines. For this purpose, it is necessary to determine energy states of machine tools and control strategies of their support units. Therefore, unsupervised and supervised learning algorithms are applied and evaluated. Based on energy states, control strategies and descriptive statistics, performance indicators are developed for enabling automatic selection and prioritization of application-dependent efficiency measures.