Research On Optimization Algorithms Of Transient Stability Control In Power System

Transient stability emergency control and preventive control are two main measures of stability control in power system. Preventive control obtains a steady-state operating point to ensure the transient stability of the system when common expected faults occur, and emergency control develops appropriate control strategies when serious faults occur to ensure the transient stability of the system. They improve the transient stability of the power system from two different sides, and have strong complementarities with each other. At the same time, both have high real-time requirements, and efficient algorithms must be introduced in the face of power systems with large scale and complex models.In this paper, optimization algorithms of power system stability control are intensively studied from these two aspects. The main works of this paper are as follows:(1) An interior-point method based on constraints conversion technology is proposed to solve the optimal load-shedding and generator-tripping problem. The transient stability criterion is that the deviation between the power angle of all generators and the center of inertia does not exceed the maximum allowable deviation during the study period. This process constraint can fully guarantee the system's transient stability, and its gradient is derived subsequently in this paper. The predictor-corrector primal-dual interior point method is used to solve the transformed optimal load-shedding and generator-tripping problem.(2) As some numerical optimization algorithms have convergence problems and are difficult to deal with the complex models in power systems, a parallel pattern search algorithm is proposed for optimal load-shedding problem. This method is a derivative-free optimization algorithm which can avoid the problem of calculating the gradient of the transient stability constraint with complicated models, and it's easy to realize without the convergence problem. To the repeated calculation of the transient stability, PSS/E is called to process dynamic simulation, which combines the existing simulation software and the algorithm, raises the flexibility of simulation models and methods, and increases the reliability of the results. Meanwhile, an appropriate frame of parallel search is used to increase the search efficiency.(3) An interior-point method based on constraints conversion technology for TSCOPF is proposed. Constraints conversion technology is applied to convert the DAE in TSCOPF into a transient stability constraint whose gradient is derived subsequently. In the conversion process, the power flow equations remain in the optimization problem, which can avoid the unsolvable power flow in the iterative process. Approximate Hessian matrix of the transient stability constraint is obtained by using BFGS, and sparse property is involved in the approximation process to accelerate the computing speed. Meanwhile, the steady-state constraints use exact Hessian matrix to improve the convergence property. As for TSCOPF with multiple credible contingencies, a general transient stability constraint is formed to further enhance the computing speed as well as a framework of parallel computing.Systems with different scales are tested by using the proposed optimization algorithms in this paper, and the results show the effectiveness and rapidity of the algorithms.