Temporal and Spatial Distribution of Power System Voltage based on Generalized Regression Neural Network

In order to deeply mine the information contained in power system operational big data, and to strengthen the system situation awareness, a data-driven system-wide voltage prediction model based on generalized regression neural network(GRNN) is proposed. To get well understanding to the system operation environment, the probability model of wind power, photovoltaic, load power and unit start-stop are established based on historical data, and the extreme operation data is supplemented by Monte Carlo sampling, which is used as robust training data, in order to improve the generalization ability of the prediction model. GRNN is trained by huge practical data plus the simulated extreme data to demonstrate voltage distribution from temporal and spatial perspective. The simulation is tested on the modified RTS-79 system, the results show that the GRNN method works well in predicting the node voltage of the power system with the average prediction error of 0.22%. The voltage temporal distribution and spatial distribution is visually depicted, respectively. And the voltage distribution feature around the system can be figured out.