Demand response ability evaluation based on seasonal and trend decomposition using LOESS and S–G filtering algorithms

Abstract

With the more and more frequent short-term peak loads appear in power systems in recent years, it is challenging to keep load and frequency steady for power systems. Demand response is launched to shave the load peak and fill the load valley. Daily load and regularity of electricity consumption behavior for special transformer users are studied in this paper. Demand response ability evaluation method based on seasonal and trend decomposition using locally weighted regression (LOESS) and Savitzky–Golay (S–G) filtering algorithms is proposed, which can optimize the effectiveness of demand response, and make the total power of demand response as close to expectations as possible. First, according to the arrangements of startup and shutdown for various devices of special transformer users in the actual production and business, seasonal load curve platforms that can reflect the habit of devices put into production for special transformer users in a day are obtained by seasonal and trend decomposition using LOESS (STL). Second, the S–G filtering algorithm is used to determine the load power for each load curve platform and evaluate the demand response ability of special transformer users. To show the effectiveness of the proposed evaluation method, a peak-shaving demand response for special transformer users in Zhejiang province is utilized for case studies. The simulation results of users with different demand response abilities show that the evaluation method proposed can accurately evaluate the demand response ability of special transformer users and provide scientific guidance for power companies to implement demand-side management.

Keywords

• Demand response ability
• Daily load regularity
• STL algorithm
• S–G filtering algorithm