Online Reliability Prediction of Energy Systems with Wind Generation

Reliability assessment of energy systems integrated with renewable sources, e.g. wind farms, has been studied by several researchers recently. To deal with uncertainties in the energy sources and the load, a number of methods have been proposed to model and predict their behavior. Different from the existing approaches which are mainly off-line, in this paper we propose two efficient methods for optimal online prediction. The first method is founded upon optimal linear estimation in conjunction with the Levinson algorithm and linear regression. The second method utilizes stochastic analysis of continuous-time Markov chains. The key feature of the chain is identified and the optimal online prediction problem reduces to a quiet simpler one by means of the Markov property. Simulations using real data show that our strategy for reliability prediction, while simple to implement, is efficacious and promising.