Generating Industrial and Residential Load Profile Scenarios with a Multivariate Elliptical Copula Approach
Year: 2024
Type: conference-paper
Abstract: Power congestion is becoming a significant concern within the Dutch electrical grid, drawing increasing attention. This challenge is aggravated by the existence of industrial and residential sites that are already experiencing local congestion in the Netherlands. To enhance the network's capacity, it is imperative to gain a comprehensive understanding of the current load profile (LP) and their statistical properties, setting the stage for the generation of accurate scenarios that can be later used as benchmark in power system studies. In response to it, this paper presents an approach to generate synthetic LPs of mixed-use areas with industries and households using Multivariate Elliptical Copula (MEC), which preserves the dependence structure between Probability Distribution Functions (PDF). The study focuses on data obtained from a Dutch industrial and residential area where 195 electrical consumption measurements were recorded throughout the year with a time resolution of one hour. By using this data as input, three synthetic datasets are generated with different shapes, and all of them effectively replicate the statistical characteristics and temporal correlation structure of the original data. To achieve this, Spearman's rank correlation coefficient was employed, together with analysis of average power behavior and histograms. The model evaluation was quantified using Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE). The model consistency further emphasize the effectiveness of MEC approach in capturing the key statistical properties of LPs from sites that include both residences and industries, along with their dependency patterns.
Institution: Eindhoven University of Technology
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4407938533