Evaluating Outage Cost of Food Industries in Navanakorn Industrial Estate by Adaptive-Network-Based Fuzzy Inference System (ANFIS)

Abstract

This paper presents the impact and outage cost data of food industries in Navanakorn industrial estate in order to apply on planning, design and maintenance in electric power distribution system. The customer data and outage costs of food industries in Navanakorn industrial estate provided from survey approach are collected. Frist, the surveyed data which is significantly different or inconsistent with the remaining data is performed the hypotheses with chi-square method to identify the factors which relate to the outage costs at significant level of 0.05. The results of hypotheses show that four factors of average monthly of electrical energy
consumption, time range of outage, outage duration and process recovery time are significantly related to the industrial outage cost. After that, these factors are assigned to be the inputs of customer outage cost model in ANFIS for evaluating outage cost of food industries in Navanakorn industrial estate. The analytical results show that the unplanned outage cost is higher than planned outage cost. In additional, the outage cost of small scale, medium scale and large scale of food industries evaluated from ANFIS model is lower than the outage cost from industrial customer survey. The learning algorithm by neuron network to adjust the
membership function parameters of fuzzy inference system provides more appropriate information than using survey outage cost directly.