Clustering-based probability distribution model for monthly residential building electricity consumption analysis

Electricity is now the major form of energy used in residential buildings and has seen a significant increase in usage over the past decades. One of the main features of electricity use in residential buildings is the diversity of total electricity consumption and use patterns among households. Current models may not be able to simulate and generate electricity use curves or reflect the variations accurately. To fill this gap, this research simulates electricity use curves in residential buildings with a clustering-based probability distribution model. The model extracts feature parameters to represent the electricity use level and patterns and then conducts a two-step cluster analysis to identify the distinctions of both electricity use levels and patterns. Based on the clustering results, probability distributions are fitted for all feature parameters within each sub-cluster. The model is then validated with three validation approaches. Monthly electricity consumption in households of the Jiangsu Province, China, was studied to test the performance of the model. Lastly, this paper discusses the application of this model under different spatial resolutions and analyzes the temporal-relevant model features.