Development of Soft Computing and Full Load Forcasting of Electrical Output Power of A Combined Cycle Power Plant

soft computing is a result of new scientific efforts that makes modeling, analysis and finally controlling of complex systems possible with more ease and success. Full load forcasting of electrical output power of a power plant is important to maximize benefit. Load operation of a power plant is affected by four main parameters as input including ambient temperature, ambient humidity, ambient pressure and exhaust vapor pressure. Soft computing methods are assessed in this paper in summary then using a body of data’s which have been gathered more than 6 years a definite learning machine was designed using descend gradient method to predict load of output power then a probability machine was designed using bayes algorithm in order to increase this prediction. Conducted comparison between definite learning machine and probability learning machine shows that probability learning machine can predict load of output power of this power plant more accurate with 96.89 % efficiency.

کلیدواژه ها:

soft computing ، load prediction ، definite learning machine ، probability learning machine