Robust Load-Frequency Control of Multi-Area Smart Grid by Combining Neural Network with Real-Time Particle Swarm Optimization

The multi-area power smart grid faces the problem of maintaining stability due to load variations and power exchange through the tie-line, which leads to a significant deviation in the system's frequency.All controllers face the challenge of changing system parameters, which leads to the controller's failure to maintain the system's stability, except the adaptive control, which has the drawback of needing a mathematical system model.The novelty of the presented method lies in facing this challenge by using a combination strategy between an artificial neural network and real-time particle swarm optimization through which the controllers' gains are continuously updated according to the change in system parameters.The proposed method provides additional novelty in immediately deducing the cost function, unlike other real-time methods, which need a long time to evaluate the cost function.In this approach, the parameters of the PID controller are tuned and updated dynamically by real-time monitoring and optimization.Simultaneously, the artificial neural network was trained to predict the optimization cost function for present and next disturbances.Simulation results confirm that the proposed method outperforms the related conventional control techniques.Comparison investigations with recent works show significant enhancement in terms of dynamic performances when the system is subjected to power disturbance equal to 20% of the rated load, which gives about 58%, 45%, and 62% reduction in overshoot, undershoot, and settling time of the frequency deviation in the first area, respectively while giving about 45%, 43%, and 54% reduction in the second area respectively.Also, it exceeds the traditional methods by about 67%, 23%, and 50% in the tie-line power variations, respectively.Furthermore, results demonstrate high robustness and resilience against the system's parameters uncertainty and variations.The robustness was verified by varying the system's inertia and turbine time constant, which increases the reliability and controllability of the modern multi-area smart grid stability.