Real-Time Parameter Tracking of Power-Electronics Interfaced Composite ZIP Load Model

Modeling load and voltage dependency can enhance the accuracy of system analysis (e.g., voltage control and situational awareness). Typically ZIP load model is used to model the voltage dependency of the load demand. With the recent widespread deployment of power electronics-based loads, traditional ZIP models are not suitable to model the voltage dependency of loads efficiently in smart distribution systems and microgrids. Power-electronics interfaced load demands the new voltage-dependent behavior models at the point of interconnection or at the aggregated level. This work develops a novel real-time parameter estimation algorithm for composite electronic-ZIP load models (CEZLM) using synchrophasor data. An adaptive window selection algorithm is proposed to extract estimation windows with almost constant load to voltage sensitivity. Estimation window extraction is followed by mode-sensitive parameter estimation for the CEZLM. A data pre-processing is developed using ensemble-based anomalous-data filtering and mitigation in the measurement stream. A cascade filter with a Kalman filter followed by a median filter is used to reduce the impact of noise. Two optimization algorithms: Constrained Least SQuares (CLSQ) and Sensitivity-Driven Constrained Optimization (SDCO) are proposed for tracking CEZLM parameters.CLSQsolves the constrained least-squares problem for parameter estimation andSDCOminimizesL₂norm of a sensitivity-driven objective function withL₁norm regularization to estimate CEZLM parameters. A T&D co-simulation framework is considered to demonstrate the performance of CEZLM parameter estimation for representative loads in a smart distribution system. Proposed algorithms have been validated using IEEE 30 bus system and 24-Bus 25kV distribution feeder with detailed CEZLM simulated usingPSS/EandHypersim real-time.