Research On Distributed And Parallel Voltage/Var Optimization Of Large-Scale Interconnected Power Systems

Voltage and reactive power optimization is an important means of improving the security and economic benefit of the interconnected power systems. The computing speed of conventional sequential algorithms can't meet the need for on-line analysis. The distributed and parallel algorithms transform the optimization problem of the whole networks into some parallel sub-problems of sub-areas. They reduce the size and complexity of the optimization problem, and decrease the amount of data communication. They can efficiently solve the real-time computation of the interconnected power systems. The distributed and parallel voltage and reactive power optimization problem of the interconnected power systems is discussed in detail in the dissertation. Primary work and achievements are presented below.1. The algorithm, named distributed and parallel voltage and reactive power optimization algorithm based on auxiliary problem principle (DPVROA-APP), and the parameter evaluation are studied. According to the DPVROA-APP, after the interconnected power systems are partitioned into some sub-areas, the decomposition and coordination model is built by replicating the variables of frontier nodes. Then the auxiliary problem principle is used to decompose the augmented Lagrange function of the decomposition and coordination model. Finally the distributed and parallel multi-area optimization problem is formed. Each sub-area can independently choose its own optimization algorithm. The DPVROA-APP is efficient for coarse-grained large-scale interconnected power systems with faster convergence and less data communication than other algorithms. The convergent performance of DPVROA-APP is dependent on the power systems and the parameters. The theoretical convergent conditions are studied. The relationships of the convergent performance with the parameters, which guide the parameter evaluation, are outlined through large number of tests.2. The power network partition affects the convergent performance ofDPVROA-APP. A good method of partitioning power networks makes the numbers of buses in all sub-areas similar and the numbers of frontier buses as small as possible. To improve the computing speed of DPVROA-APP, the maximum-cardinality matching based fast sub-area division algorithm is presented in this dissertation. It reduces the frontier buses of each sub-area in turn on the basis of random initial partition and at the same time balances the numbers of buses in all sub-areas. In this dissertation, the multiple-way graph partitioning algorithm is first introduced to divide power networks into several sub-areas. Besides, several important deductions are educed for the first time to improve the multiple-way graph partitioning algorithm. The maximum-cardinality matching based fast sub-area division algorithm can reduce not only frontier nodes as many as possible at each iteration but also total iteration number. It is applicable for the voltage and reactive power optimization of the large-scale interconnected power systems.3. The allocation of the shunt reactive compensatory equipment is a premise of the voltage and reactive power optimization. In order to obtain the optimal security and economic benefit, some load buses should be selected as the allocation sites of the shunt reactive compensatory equipment. The models of some conventional methods are complicated and hard to solve. And some methods can't obtain satisfying compensatory effect. To overcome these shortcomings, an approach for the allocation of reactive compensatory equipment using voltage deviation with small-disturbance is put forward. With taking into account different operating conditions, the load buses with big average voltage deviation after random small reactive power disturbance are selected as compensation sites. In this method, greedy algorithm, which is a classical heuristic algorithm, is adopted to solve the objective function. The model of this method is simple and easy to solve. This method is applicable for on-line calculation of large-scale power systems. Simulation results show that the selected compensation buses, which locate near load centers or far away from generators, are robust and can reduce network power loss and improve voltage quality effectively.4. There exists regional difference in the power networks in our country. Torealize the DPVROA-APP on the basis of the existing voltage and reactive power optimization software in every power network, the multi-agent technique is utilized to set up a distributed open software platform. A novel distributed optimal voltage and reactive power control system based on the multi-agent technique is proposed. There are three layers in this system, with each layer containing one kind of agent. The three kinds of agents are Management Agent (MA), Area Optimization Agent (AOA), Voltage and Reactive Power Control Agent (VRA). MA allocates the shunt reactive compensatory equipment, and divides power networks into some sub-areas, as well as starts and coordinates the distributed and parallel optimization of AOAs. AOA locates in sub-area. After receiving the order from MA all AOAs adopt the DPVROA-APP to compute in parallel. VRA controls shunt capacitor, shunt inductor, on-load tap changer, or generator. Whenever voltage or reactive power violation of the local bus is detected, VRA regulates voltage and reactive power control equipment actively and rapidly, thereby making the bus voltage and reactive power run close to the optimal set values. Simulation results show that this system has the advantage of fast speed, and can reduce network power loss and improve voltage quality obviously. Besides, benefiting from its reliability, flexibility and adaptability, it is convenient to be realized on the basis of the existing voltage and reactive power optimization software in our country.