Intelligent Optimization Techniques–Based Wheeling Pricing in Competitive Power Market

In this paper, Bialek tracing has been employed to track the contribution of each generator and load in the wheeling flow of each transmission line. For computing wheeling prices, the mega volt ampere-mile method among various embedded cost based pricing techniques has been used, in which optimized flows have been determined by employing metaphor-less Rao-3 algorithm-based AC-optimal power flow (OPF). Embedded cost based pricing techniques are more reliable for wheeling pricing because they are based on actual flows and recover fixed costs of wheeling facilities. In the spot power market, due to continuously varying load conditions, the flows over transmission lines have also varied drastically. These varied flows are determined by running the OPF program for each and every load condition. Consequently, the computation of wheeling prices in real time for these varied flows has become quite a time-taking process. Instead, employing the artificial neural network (ANN) approach for estimating wheeling prices for any loading scenario, instantly and accurately, is quite an attractive, efficient alternative. Various radial basis function neural network have been developed for each cluster in which wheeling prices lie in a closed range. All these neural networks were trained and tested under parallel computing environment to acquire speed gain. The proposed neural network and parallel computing-based intelligent approach for estimating wheeling prices in the real-time competitive power market has been demonstrated and examined on IEEE 30-bus system and Indian utility 146-bus system. The results are found well within acceptable accuracy limits.