Development and Verification of a Machine Learning Based Motor Energy Consumption Prediction Algorithm

The popularization of smart home in people’s daily life makes motor energy consumption prediction one of the key technologies for energy saving and emission reduction. In order to improve the prediction accuracy and efficiency, this paper proposes a motor energy prediction technology based on Long Short-Term Memory (LSTM) network. This paper introduces the motor energy prediction, emphasizes its significance in the field of industrial energy efficiency and smart home, and puts forward an intelligent tool to deal with this big data problem through machine learning technology. The paper also introduces data acquisition and preprocessing, feature engineering, model design, iterative training and estimation test strategy. The experimental results show that, compared with the autoregressive moving average model (ARMA) and the autoregressive fractional integrated moving average model (ARFIMA), the proposed LSTM model has a higher determination coefficient, with a maximum of 0.98, which is beneficial to the prediction of motor energy consumption. This paper also tests the robust performance of the LSTM model and its application under different loads, and discusses the performance conclusions of the LSTM model, the intrinsic value of its types and its possible application in the creative field.