Comparative Assessment to Predict and Forecast Water-Cooled Chiller Power Consumption Using Machine Learning and Deep Learning Algorithms

Over the last few decades, total energy consumption has increased while resources remain limited. Energy demand management is crucial for this reason. To solve problem, predicting and forecasting water-cooled chiller power using machine learning deep are presented. The prediction models adopted thermodynamic model multi-layer perceptron (MLP), time-series MLP, one-dimensional convolutional neural network (1D-CNN), long short-term memory (LSTM). Each group of compared. best in each then selected implementation. data were collected every minute from an academic building at one universities Taiwan. experimental result demonstrates that MLP with 0.971 determination (R2), 0.743 kW mean absolute error (MAE), 1.157 root square (RMSE). trained day three consecutive days new to forecast next consumption. LSTM 0.994 R2, 0.233 MAE, 1.415 RMSE. both indicated very close predictive values actual value.