Application of adaptive neuro fuzzy inference system for outage cost evaluation

The aim of this paper is to describe an alternative analytical method in order to evaluate customer outage cost (COC) in Thailand. The information of electrical expense, outage frequency, outage duration, and process recovery time from industrial customers is gathered. They are used to be inputs of the proposed adaptive neuro fuzzy inference system (ANFIS). In the data training by neural network, the least error of COC found in the model which adjusted the membership functions according to data ranges in customers questionnaire by 9,078 THB/event. The simulation results of 300 industrial customers show that COC is 39,600-750,000 THB/event while the COC from customer survey is 51,500-112,000 THB/event.