Prediction of Household Electricity Consumption Using Convolutional Neural Network

Electrical energy is an energy source that is needed in various fields of life. It is important to utilize electrical energy usage data to determine usage patterns, to provide consumption insights, and to contribute to energy conservation. This energy consumption may not necessarily support increasing productivity due to wastage. The loss of electrical energy is one of the aspects that escalate the energy consumption, thus it is very important for consumers to be aware of the amount of energy lost. The objective of this research is to identify the effectiveness of the Convolutional Neural Network (CNN) method to predict the level of electrical energy consumption in buildings. The data used is ECO (Energy Consumption and Occupancy) which was collected over a period of 8 months in 6 households. Electrical appliances used are refrigerators, hair dryers, coffee machines, water heaters, washing machines, and computers. The analysis is in the form of predicting electrical energy consumption. The analysis steps are identifying problems, understanding the work of the existing system, and analyzing the system. The analysis is a prediction of electrical energy consumption. This research combines event detection algorithm with CNN method. The event detection algorithm identifies the occurrence of increased power that indicates the use of appliance. After the prediction of electrical energy consumption with the CNN method was tested, the RMSE value is 2.2056. This RMSE value is considered acceptable as a reference in building a real application and being able to predict electrical energy consumption.