DWT-Based Aggregated Load Modeling and Evaluation for Quasi-Static Time-Series Simulation on Distribution Feeders

This paper presents an extension of a previously reported discrete wavelet transform (DWT) based load modeling methodology which targets at modeling time-series load profiles to enable the more effective and realistic quasi-static time-series simulation of distribution feeders. The time-series load model is composed of two major parts: 1) The low-resolution field-measured load data which usually have a resolution of 30 or 15 min., 2) The high-resolution load variability model data extracted from the established variability database which can have resolution up to 1 sec. A load aggregation methodology is developed to aggregate the load profiles so that load profiles at different transformer ratings can be effectively modeled. Validation, evaluation, and analysis of the developed load modeling approach has been performed on the IEEE-123 feeder and an actual utility feeder from California. The analysis completed on the two feeders have demonstrated the effectiveness and revealed the value of the developed model for distribution feeder quasi-static time-series simulation at high temporal resolution.

I. Introduction

Load modeling has always played a critical role in distribution system simulation, especially in recent years, load models which could enable quasi-static time-series (QSTS) simulation are in high demand. This is because utilities and grid operators desire effective load models, at a high temporal-resolution, to develop and test new control algorithms and applications which are used to inform the challenges brought by the increasing growth of roof-top PV systems and other distributed energy resources (DERs) which are being integrated into the power grid [1]-[2]. Traditional load modeling and load modeling aggregation methods are discussed in [3] but such methods typically characterize loads based on assumed load classes and dominant behaviors. This work focuses on an empirical, data-driven method that leverages utility data.