Study On Wind Power-integrated Power System Planning And Operation Considering The Uncertainty Of Probability Distribution

Nowadays, due to the environmental pollution and the conventional energy crisis, wind power as a clean and sustainable energy gains worldwide popularity and applications. However, because of the intermittency and volatility of wind power, the structure and operating features of power systems are affected by the large-scale integration of wind power which also raises the requirement of the transmission and peak load regulation abilities of power systems. This dissertation takes the uncertainty on the probability distribution of the random variables (such as wind power and its forecast error) in the power systems into consideration, and is devoted to addressing power system planning and operation problems with large-scale wind power integrated. The main work of this dissertation can be summarized as follows:1. A maximum entropy principle-based method is presented for grid-connected wind power capacity optimization. Firstly, the cumulant method is used to calculate the information of several orders of moments for the stochastic power flow variables in the power system. Secondly, the maximum entropy principle is applied to solve the most possibly realized probability distribution of the stochastic power flow, using the partial information of the power flow variables. Based on this, the wind power capacity optimization problem is formulated as a chance-constrained programming model, which takes the system security requirements into consideration and maximizes the wind power capacity.2. A generation dispatching model based on the maximum entropy principle is presented for power systems with high penetration of wind power. First of all, the generation dispatching strategy is formulated by linearizing the power flow equation at the operating point. Secondly, the generation dispatching optimization model is presented, aiming to find an optimal dispatching strategy that minimizes the generation cost and satisfies the security constraints of the power system. Furthermore, the maximum entropy principle is used to find the most possibly realized probability distribution of the power flow, thus provide an accurate probabilistic circumstance to solve the generation dispatching model. At last, the proposed method is compared with the methodolodies based on Monte Carlo simulation and Gram-Charlier expansion in the numerical studies.3. A transmission system planning method considering the uncertainty on the probability distribution of wind power is proposed. The method guarantees that the security requirements of system operation can be satisfied under all the possible probability distribution scenarios of wind power, while minimizing the investment costs. Firstly, the distributionally robust chance-constrained optimization model is applied to the transmission system planning problem. Secondly, the S-lemma and Schur complement are used to eliminate the random variables, so as to convert the probabilistic model into a deterministic model with matrix inequalities. At last, the genetic algorithm based on the linear matrix inequality optimization is adopted to solve the deterministic model. Simulations are performed on a real power system to verify the effectiveness of the proposed method.4. A distributionally robust optimization method is proposed for the reserve schedule decision-making problem with partial information of wind power. Considering the uncertainty on the probability distribution of wind power, a distributionally robust joint chance-constrained programming model is used to formulate the reserve scheduling problem, and simultaneously solves the generation dispatch strategy for the conventional units while determining the minimum cost reserve capacities and the optimal way to deploy the reserves at the real time. The solution of this model guarantees that the system security requirements can be satisfied over any possible probability distribution of wind power. To achieve the tractability of this model, the S-lemma and Schur complement is applied to convert the probabilistic problem into a deterministic bilinear matrix inequality-constrained problem which can be solved by the sequential convex optimization algorithm.5. A distributionally robust coordinated reserve scheduling model is presented, considering the conditional value-at-risk-based wind power reserve requirements. The object of the model is to minimize the total cost of conventional generation and reserve procurement, while satisfying the security requirement over all possible probability distributions of wind power forecast error. In this model, a distributionally robust formulation based on the concept of conditional value-at-risk is presented to obtain the wind power reserve requirement, and the coordination of various reserves and the constraints of generation ramping rates are considered. In the case studies, the proposed method is compared with the normal distribution-based coordinated reserve scheduling method which assumes the wind power forecast error is of normal distribution.The proposed planning methodologies and operation strategies provide technique support for wind power integration, and enrich the existing power system planning and operation studies.