Subject Area: Electrical Engineering
This paper presents the modeling of an intelligent virtual power plant for optimal energy network management and control using machine learning techniques. The aim is to present a system that can collect data from the grid and interpret it to make control decisions that improve the quality of service on the Nigerian 36-bus, 330KV interconnected transmission network. In the study, an observation and simulation approach was used as the research methodology. The methods used were data collection from the National Control Centre (NCC) Oshogbo with Newton-Raphson load flow analysis, data processing with a shunt active filter, the Power Flow Reference Model (PFRM), the Intelligent Load Flow Control System (ILFCS), and the and the Intelligent Virtual Power Plant (IVPP). The PFRM was modeled with the Back Propagation Feed Forward Neural Network (BNN) to develop the intelligent monitoring system; the ILFCS was modeled with the PFRM and the Nigerian Electricity Regulation Commission (NERC) standard for voltage stability to develop the control algorithm. The PFRM and ILFCS were used to model the IVPP. The IVPP was implemented with MATLAB and tested. The result of the PFRM measured with Mean Square Error (MSE) was 5.623e-06Mu, which is good as it is approximately the ideal MSE value, which is 0. The regression (R) performance was 0.9999, which is good as it is approximately the ideal R of 1. The implications of these results showed that the PFRM was able to learn the patterns of the load flow correctly and was able to detect changes on the grid with respect to the voltage profile. The performance of the ILFCS showed that the load flow from the grid was intelligently monitored, and then the unstable bus was controlled, while those at their stability limit were also addressed.