Research On Transient Energy Constrained Optimal Power Flow

As a grand strategy to form a strong and smart grid including the ultra-high voltage backbone and the coordinated network at all voltage levels, how to ensure the power system transient security and economy simultaneously is urgent needed. Hence, the optimal power flow with transient stability constraints (OTS) problem, which could consider both the security and economy, is highly important for theoretical research and practical application. The previous OTS model was non-adaptive due to the heuristic generator angle transient stability constraints, and its optimization process was also very complicated and time consuming, therefore a adaptive transient energy based OTS model and the corresponding efficient optimal method have been researched in this thesis to solve these problems.To overcome the heuristic feature of generator angle constraints, a transient energy based OTS model is primarily studied. Based on the Integrated Extended Equal Area Criteria, the transient energy constrained optimal power flow model is proposed and eliminates the heuristics threshold of the traditional generator angle constraint. The proposed transient energy based OTS model is adaptive for different power systems and various fault conditions, and it reduces the number of inequalities in OTS model as well. Furthermore, the derivative-free particle swarm optimization algorithm is utilized to solve the non-convex non-differentiable OTS model with the considerations of the valve point effects and the carbon emission penalty. As the proposed transient energy based model is adaptive and the particle swarm optimization algorithm is reliable and effective, their combination provides a basis for solving the OTS problem.As the OTS problem consisting of the transient stability and optimal power flow issues simultaneously is very complicated and time consuming, the following three enhancements are conducted further to improve the efficiency of solving the OTS problem.In order to enhance the transient stability analysis efficiency in OTS, the parallel varied-order multi-step Taylor series method is researched. According to the numerical analysis theory, a dynamically controlled varied-order multi-step Taylor method is formed. Theoretical analysis indicates that the proposed Taylor method has a good numerical stable region and can effectively control the derivative order in transient stability analysis according to the accuracy, thus eliminates the calculation redundancy of the fixed-order Taylor series method. Based on modern computer technologies, a parallel varied-order multi-step Taylor series method is further put forward for transient stability analysis by which the computation of higher order derivatives are parallelized. The proposed parallel method greatly improves the efficiency of transient stability analysis from the viewpoints of mathematical solver and the parallel computation, thus it speeds up the transient stability analysis in OTS optimization.To further directly determine the transient stability in OTS optimization, an method for efficiently calculating the practical dynamic security region is also researched. Either the analytical method or the point-fitting method for practical dynamic security region calculation is required to search for at least one power vector as critical transient stable; however, the critical active power vector calculated by the dichotomy in time domain simulations in previous work is inefficient. Based on the transient energy margin sensitivity, the critical power vector is obtained efficiently in an iterative manner; then the practical dynamic security region is derived from these energy margin sensitivities of the critical power vector. The proposed transient energy margin sensitivity method for the practical dynamic security region calculation is efficient and effective, and it retrenches the process of transient stability analysis in OTS optimization and speeds up the optimization.Finally, the particle swarm optimization algorithm is improved to enhance the optimization efficiency, and a synthesized approach is proposed to solve the OTS problem quickly. By dynamically nonlinear adjusting the inertia weight, cognitive weight and social learning weight, the improved particle swarm optimization (IPSO) algorithm is formed. The results of mathematic benchmarks approves that the proposed IPSO method has a fast convergence and good optimization performance. Based on the key points researched above, namely the variable-order multi-step Taylor series strategy, the practical dynamic security region and the IPSO method, a parallel synthesized approach is established to solve the OTS problem. The simulation results validate that the proposed synthesized approach can obtain the satisfied optimal solutions with reduced time consuming, thus it can solve the OTS problem efficiently and effectively.