Medium term stochastic load model for transformer and feeder from AMI load data spectral analysis. (October 2017)

• Record Type: Journal Article
• Authors: Zhong, Shiyin; Broadwater, Robert; Steffel, Steve
• Is Part Of: International journal of electrical power & energy systems. Volume 91(2017)
• Page Start: 1
• Page End: 9
• Publication Date: 2017-10
• Subjects: Spectral analysis -- Discrete Fourier transform -- Discrete wavelet transform -- Electricity usage patterns -- Harmonics; Electrical engineering -- Periodicals; Electric power systems -- Periodicals; Électrotechnique -- Périodiques; Réseaux électriques (Énergie) -- Périodiques; Electric power systems; Electrical engineering; Periodicals; 621.3
• DOI: 10.1016/j.ijepes.2017.02.010
• Languages: English
• ISSNs: 0142-0615

Abstract

Highlights: The paper introduces a medium term stochastic load model using the frequency domain representation of the hourly AMI load data. The load model is driven by a temperature based predictor variable, which is incorporated into the model using the local hourly weather data. The model is a stochastic model. Using the sets of load estimation values generated from the model, a transformer/feeder hourly load profile envelop can be developed. The model utilizes the frequency domain representation of the AMI data. The orthogonal relationship between frequency components gives the model its modular characteristic. The model does not depend on time domain parameters for estimating the hourly load. It only requires the daily weather condition index to generate the hourly load profile. The results demonstrate that the frequency domain load model can produce accurate medium term hourly load profiles at the transformer and feeder levels under different weather conditions. Abstract: A medium term stochastic load model using the hourly AMI load data aggregated at transformers and feeders is introduced. Load model frequency domain statistical parameters are derived from spectral analysis of AMI data. The load model is driven by a weather index based on Heating-Cooling Day Degree. This index is incorporated into the model using local, hourly weather data. Advantages of the proposed load model are discussed. Performance comparisons between the proposed model and several time-domain methods are presented.