Predicting power for home appliances based on climatic conditions

Abstract

Purpose

Power management in households has become the periodic issue for electric suppliers and household occupants. The number of electronic appliances is increasing day by day in every home with upcoming technology. So, it is becoming difficult for the energy suppliers to predict the power consumption for households at the appliance level. Power consumption in households depends on various factors such as building types, demographics, weather conditions and behavioral aspect. An uncertainty related to the usage of appliances in homes makes the prediction of power difficult. Hence, there is a need to study the usage patterns of the households appliances for predicting the power effectively.

Design/methodology/approach

Principal component analysis was performed for dimensionality reduction and for finding the hidden patterns to provide data in clusters. Then, these clusters were further being integrated with climate variables such as temperature, visibility and humidity. Finally, power has been predicted according to climate using regression-based machine learning models.

Findings

Power prediction was done based on different climatic conditions for electronic appliances in the residential sector. Different machine learning algorithms were implemented, and the result was compared with the existing work.

Social implications

This will benefit the society as a whole as it will help to reduce the power consumption and the electricity bills of the house. It will also be helpful in the reduction of the greenhouse gas emission.

Originality/value

The proposed work has been compared with the existing work to validate the current work. The work will be useful to energy suppliers as it will help them to predict the next day power supply to the households. It will be useful for the occupants of the households to complete their daily activities without any hindrance.

Citation

Kaur, J. and Bala, A. (2020), "Predicting power for home appliances based on climatic conditions", International Journal of Energy Sector Management, Vol. 13 No. 3, pp. 610-629. https://doi.org/10.1108/IJESM-04-2018-0012