Prediction of energy consumption of machine tools using multi-gene genetic programming

In the past, researchers have applied different analytical, numerical, and empirical modelling techniques to analyze energy consumption. In the present study, computational artificial intelligence-based Multi-Gene Genetic Programming is used to model the energy consumption of machine tool. The experiments were performed on a heavy-duty HMT lathe machine tool under a dry environment in the interest of sus-tainable machining. The Taguchi full factorial orthogonal array L27 was used to develop the experimental plan. The power consumption of the machine tool was measured using a Fluke 435 power analyzer. The dataset was split into training and testing data based on the 80???20 ratio. Further, 99.77% goodness of fit was achieved in training and 98.60% for testing the model. The adequacy of the model was tested by determining four error indices i.e. root means square error, mean absolute error, sum of squared error, and mean square error. The model is validated by conducting two hypothesis tests, t-test and f-test on predicted data. The hypothesis results confirm the model's goodness of fit statistically, indicating that the proposed model can be easily applied in the manufacturing industry to predict energy consumption. Copyright (C) 2022 Elsevier Ltd. All rights reserved. Selection and peer-review under responsibility of the scientific committee of the International Confer-ence on Artificial Intelligence & Energy Systems.