Application of neural networks to automatic load frequency control

The first theoretical analysis on PI controllers dates back to 1910. Even though the speed of response and overall stability of the system is slow, PI controllers are still used today. On the contrary, twenty first century customers, aware of power quality standards, demand faster response and very small settling. This paper reports the implementation of neural control to reduce load frequency fluctuations. A power system model has been simulated within a MATLAB environment. A comparative study between the frequency response of the system using a PI controller and a neural controller trained with Levenberg Marquardt algorithm has been done. The neural controller boasts of its superiority over a PI controller in terms of its settling time and peak overshoot and its simplicity of realization.