Application of genetic algorithms for control design in power systems

Proposed is a new method of designing control systems which relies on a combination of advanced system simulators and genetic computation. In contrast with standard approaches applied to control systems, which often rely on a simplified model of a system, this framework provides realistic full-scale modelling abilities, accomplished via system simulators along with the optimisation versatility of genetic algorithms. In particular, genetic algorithms are capable of handling high multivariable and multimodal optimisation problems which usually arise in the setting of complex systems. The method is implemented as an extension to the electromagnetic transients (EMT) program EMTDC™. The combination of EMT program and genetic algorithms resulted in a tool for optimal control design in the area of power systems. A method of using this tool and the results of its application to the design of optimal controls for the CIGRE benchmark model and a multi-terminal DC system are presented.