Research On Transient Stability Constrained Optimal Power Flow In Power Systems

Optimal power flow with transient stability constraints in power systems is an important research subject of power system transient stability preventive control. At present, the research hotspots focus on the disposal of transient stability constraints and the introduction of highly efficient optimization algorithms to solve this problem. According to different approaches to deal with transient stability constraints, two main categories of approaches are used: the indirect approach based on time domain numerical simulation and the direct approach based on Lyapunov transient energy function, and the former one can be further divided into the numerical discretization method and constraint transcription method.Viewing from its mathematical model, transient stability constrained optimal power flow (TSOPF) is a large scale nonlinear programming problem. As a mature and efficient optimization algorithm, the interior point method and its improvements are widely utilized in solving this problem. This dissertation mainly studies the numerical discretization and interior point method based differenced TSOPF problem. TSOPF research has currently proceeded from single credible contingency to multiple contingencies and from small scale test cases to medium and large scale practical power systems. The interior point method will easily suffer from 'the curse of dimensionality' such as enormous time consumptions, excessive memory occupation and even insolvabilities in solving such kind of large scale optimization problems as differenced TSOPF. Aiming at the above deficiencies of existing interior point TSOPF algorithm, this dissertation attempts to decrease its dimension from two aspects to reduce the problem scale and enhance computational efficiency.First, an iterative contingency filtering strategy is developed for multiple credible contingencies. In TSOPF, the number of contingencies that will have an effect on the objective function is not too much, so it is neither necessary nor realistic to involve all the contingencies in TSOPF formulations. Therefore, an iterative multi-contingency filtering strategy based on the concept of active set and time domain simulation is proposed. Active contingencies are firstly identified with time domain rotor angle curves information at the current optimal operational point, then critical contingencies that dominate other active ones are filtered out, and interior point TSOPF program is implemented once with only the critical contingencies to obtain the next optimal operational point. This process iterates until no new active contingencies are found.Second, an improved TSOPF formulation is proposed to overcome the drawback of oversize dimensions of correction equations in interior point method. One cause why the dimension of interior point TSOPF is so large lies in the fact that power system differential algebraic equations are generally discretized to equality constraints. In fact, any numerical discretization method has truncation error and there's no need to treat the difference equations as equality constraint. This dissertation proposes an order-reduced interior point TSOPF algorithm, in which the differential algebraic equations are converted into differenced inequality constraints related with numerical integrator precision. Hence TSOPF problem scale is reduced nearly by half, moreover, the overall time consumptions and memory occupations are also reduced, and the computational efficiency of interior point method is greatly enhanced.The proposed contingency filtering strategy and order-reduced interior point algorithm are both tested on several cases of different scales. Test results indicate that the proposed contingency filtering strategy is feasible and reliable, and order-reduced interior point algorithm is more computationally efficient than conventional approach. These two enhancements have pointed out a new way to solve large scale TSOPF problems and shown a good application prospect.