An efficient scenario generation technique for short-term wind power production

Power system operators and electricity market participants are facing many challenges due to the rapid expansion of uncertain energy resources, such as wind power. An efficient way of handling these uncertain resources in optimization is through the use of stochastic programming (SP). An essential step in SP is scenario generation, at which multiple plausible realizations of each stochastic variable (e.g. wind power output) are generated. The generated scenarios for each stochastic variable need to reflect the statistical properties of that variable. In this work, a new heuristic technique to generate wind power scenarios is presented. The scenarios generated using this technique meet reasonably well the statistical characteristics of the hourly wind power forecast and the hour-to-hour coupling of wind power outputs. Simulation results show that the presented technique, despite its simplicity, is effective in presenting wind forecast statistical characteristics. The use of the proposed technique in SP is illustrated via the application of SP-based wind thermal energy trading in electricity markets.