A Novel Strategy to Estimate and Manage the Power Consumption of Household Appliances

Residential and commercial buildings account for approximately 26.6% and 21.2% of worldwide electrical power consumption, respectively. Because of the increasing need for energy and the lack of present energy sources, there will be a significant challenge in the future. As a result, with the rise of Smart Cities and Smart Grids, it is becoming increasingly vital to have effective strategies for controlling and reducing the consumption of electricity. One method of achieving consumption management is to predict the power usage of household equipment. This study focuses on power usage prediction using data from the UCI repository. In the preprocessing step Checking for null values, eliminating outliers, and modifying data are performed. Following that, the essential features are selected for dimension reduction using the Principal Component Analysis (PCA) technique. The Deep Sequential Prediction Model (DSPM) is the suggested deep learning method for predicting power consumption. Finally the significance of feature extraction and the effectiveness of the proposed model is evaluated using the R2 score and MSE. The proposed model is compared with the standard Artificial Neural Network (ANN) model. To examine the importance of the PCA method, feature-extracted data (from a correlation plot) and pre-processed data (without feature extraction) are used. The simulation shows that power consumption prediction utilizing the PCA feature extraction module along with the proposed DSPM model provides an excellent R2 score.