Study on Search and Correction of Error Dominant Regions in Simulation Based on Measured Trajectories

With the proposal of the carbon peak and carbon neutrality goal, a significant influx of new energy sources with random and fluctuating characteristics is being integrated into the power grid. Consequently, the adaptability of the traditional four-parameter model, as well as new energy models, to the actual power grid is becoming increasingly prominent. Conversely, simulations often rely solely on the accuracy of models and algorithms and are based on predefined conditions. To address the aforementioned challenges, this paper utilizes PMU-measured data to identify error-dominant regions. Initially, it employs a hybrid dynamic simulation method to partition the power grid simulation system into blocks. Concurrently, it establishes a simulation credibility evaluation system and employs comprehensive simulation credibility to identify error-dominant regions. Subsequently, it iteratively decouples sub-regions to systematically identify error elements. Subsequently, the perturbation method was employed to analyze the trajectory sensitivity of each parameter of the element. Parameters with low sensitivity were assigned typical values, while those with high trajectory sensitivity were chosen for identification. The genetic algorithm was then utilized for online parameter verification and optimization. Finally, the proposed method was validated using RTDS real-time simulation software. Experimental results demonstrate that the calibrated model simulation results exhibit better alignment with measured data, highlighting the accuracy and reliability of the proposed approach.