Power System Unit Commitment Problem With Wind Farms

As the problems of energy shortage and environmental crisis have becomeincreasingly prominent, the development of clean renewable energy sources to replacetraditional fossil fuel for generation is becoming more and more urgent and important.Wind power is clean and safe, and it has low cost. It also has the most obvious economicadvantages. Wind power has been developed rapidly in our country and even around theworld. However, as the randomness, intermittence and uncontrollability of wind power, theintegration of large-scale wind power brings new challenges to the power systemgeneration scheduling and dispatching. The article carries out the research about modelingand solution methods of the power system unit commitment with wind farms.(1) As wind power is a random variable, and wind power forecasting accuracy is stilllow, directly using wind power prediction curve for unit commitment modeling may havesimple process, but is difficult to simulate wind power output ranges, thus too rough to getthe optimal solution. The article proposes both the unit commitment model based on thewind power interval prediction information and the stochastic unit commitment modelbased on scenarios. The former uses wind power interval prediction information to considerthe system reserve capacity demand caused by the wind power uncertainty, and the lattermodels the possible output range of random variables through a certain number ofrepresentative scenarios and the corresponding probabilities, and it makes the optimal uniton/off decision satisfying the minimal operating cost. Then, it works with rolling plan toachieve the coordination dispatch of day-ahead decision and real-time scheduling. Newscenario generation and reduction technologies are proposed. Example simulation resultsshow the validity and usability of the models.(2) Unit commitment problem belongs to the two-level programming problem, and is ahigh-dimensional, non-convex, nonlinear and mixed integer optimization problem, which isdifficult to obtain the optimal solution. So, the article proposes the inner and outer layeroptimization method. The outer layer uses the advanced Quantum-inspired Binary PSO forthe regular unit on/off problem, and the inner layer uses the primal-dual interior pointmethod for load economic dispatch problem. In addition, for improving the efficiency ofthe algorithm and the accuracy of solutions, based on the basic Quantum-inspired BinaryPSO, the article proposes two improvements. The one is that this paper adopts partially greedy mutation strategy to make particles easily jump out of the local optimal solutions,the other is that it deploys heuristic adjusted regulations to correct cross-border individuals.(3) For the proposition of the energy conservation policy, two models of unitcommitment with wind farms considering air pollutant emissions are introduced and a newmulti-objective Quantum-inspired Binary PSO algorithm is proposed. The one model usesweighted coefficient method to transform the multi-objective scheduling problem intosingle-objective optimization problem, and the other model considers a double-objectivefunction considering both costs and emissions and uses the multi-objectiveQuantum-inspired Binary PSO algorithm based on the Pareto optimization to solve it.Example simulation results indicate that, considering both costs and emissions can get moreappropriate solutions and make compromise in the environmental and economic benefits,and multi-objective Quantum-inspired Binary PSO algorithm can get better solutions andbe more practical and usable compared to weighted coefficient method.