A Review of Neural Networking Methodology to Different Aspects of Electrical Power Systems

There is an ever increasing consumer base in the electrical power industry and with the rapid advancement in the field of electricity generation, electricity can now be supplied to many underdeveloped areas in impoverished locations, previously considered difficult or inaccessible. Along with this expansion, there is also the growing use of semiconductor electronic and solid state switching devices in everyday life. These devices are nonlinear loads and therefore produce severe distortions in the power flow. Providing a solution to the variety of problems that result from these applications requires the estimation and accurate analysis of the multitude of data available. The most pressing of these problems, as has been found surveyed in recent decades, are load forecasting, estimation of harmonics, fault detection, economic dispatch, security assessment and voltage stability analysis. A great many methods of analysis of these problems have been proposed, mainly statistical ones. However, in recent times, artificial intelligence techniques have gained popularity in this field, most notably the method of neural networking (NN).The popularity of neural networks lie in their speed, accuracy and learning ability in the case where a mathematical relation cannot be established. In lieu of the number of applications of neural networks in power systems, this paper provides an outline and brief descriptions of the proposed methods of analysis of the problems mentioned above.