Study On The Algorithm Of Distribution Network Reconfiguration And Service Restoration With Distributed Generations

In recent years, along with the gradually depletion of the conventional energy such as oil, coal and natural gas particularly on the earth, and the increasing global concerns for environmental protection, energy saving issues the development of the power science technology, the distributed generation (DG) technology based on wind power, photovoltaic power generation and other renewable energy technologies is becoming more and more mature and has become a hot research topic. However, after a great of DGs being connected into the distribution network, the distribution network structure, operation and control mode will be changed tremendously. The domestic and foreign researchers have done much research in the grid-connected DG impacts on the power system, but few of them studied on the distribution network reconstruction and service restoration after fault with large-scale DGs connected, to say nothing of the research on how to use the planned DG islanding resonablly to restore the service of the outage area with no fault under premise of ensuring the reliability of power supply of the important loads. So the study on this area has an important theoretical and practical significance. In this dissertation,the three-phase unbalanced power flow algorithm for the distribution network with multiple DGs and distribution network fault diagnosis, reconstruction, power supply recovery with DG have been studied intensively. The contributions of this dissertation are summarized as follows:(1) The distribution network reconfiguration, fault location and service restoration are all based on power flow calculation of distribution system. To solve the problem of power flow calculation of distribution network connected with various DGs, DG modes are constructed respectively according to the different features of different types of DGs, thus the nodes with DG are classified into PQ node, PV node, PI node and P-Q (V) node in this dissertation. In view that back/forward algorithm cannot cope with weakly meshed distribution network and PV nodes, this dissertation presents an improved layer-by-layer current back/forward sweep power flow algorithm based on traditional back/forward sweep algorithm by using superposition principle and iterative, compensation method to calculate the unbalanced current of three-phase loads and puts forward the method of coping with weakly meshed distribution network and PV nodes. The method simplifies power flow calculation of distribution system with DGs.The examples of IEEE33-bus test system and a practical distribution network has proved the efficiency and feasibility of the algorithm.(2) That different types and multiple DGs are connected into the distribution network will have great influence on the reconfiguration and optimization of the distribution network. Due to the frequently changes of the loads and the time uncertainty of DGs in the grid, The distribution network real-time reconstruction is very necessary, the calculation speed and efficiency are very critical in real-time reconfiguration, the previous distribution network reconfiguration were either slow in calculation or couldn’t find the optimal solution. The dissertation presents an improved mixed-integer differential evolution algorithm based on traditional differential evolution algorithm that can deal with discrete variables to solve the reconstruction optimization problems of distribution network with distributed generation. The method takes the minimum network loss as the objective function, through the accelerated and migration operations to overcome premature convergence of the traditional differential algorithm due to a small population size and improve the global search capability. The integer encoding method based on loop circuit is proposed to reduce the code length, and greatly improves the proportion of feasible solutions, and further increases computational efficiency, laying the foundation of the real-time reconfiguration.(3) It makes the topology of distribution network more complex when a large number of distributed generations are connected into it, distribution network protection and fault diagnosis are more difficult to make. However, the fault diagnosis and location is the base of distribution power system service restoration after fault occurred; its response speed is the key to increase the reliability of power supply reliability. For this reason, a fault location approach suitable to the distribution network containing large-scale DGs is proposed. Firstly, according to the results of power flow calculation, a database of voltage sag is established; then power quality information such as voltages, currents and power is acquired by power quality monitors installed in the critical nodes of the distribution network and coordination analysis on collected nodal voltage data and voltage sag data of the established database is performed; and then the node with the matching extent value most close to1is determined as the fault point., The method is accurate and fast for fault location, and even if the distribution network structure changes, after modifying the corresponding node voltage sag database slightly,it is still suitable to locate the fault occurred effectively.(4)A new challenge is proposed for the service restoration of distribution network when a large number of distributed generations are connected into the distribution network, in view that the traditional distribution network service restoration algorithm did not consider the islanding of the grid-connected DG. A distribution network service restoration algorithm considering the DG islanding is proposed. First, according to the importance of the loads for DG Islanding division, called the planned islanding. When serious failures result in large area power blackout, DGs are transferred into the islanding operation mode based on islanding division method to maintain the power supply for the important load. Then using the binary improved mixed integer differential evolution algorithm to restore the service of the blackout area with the non-fault area, taking the minimum switch operations and the minimum network loss as the objective function, trying best to recover more loads, for those who can not be restored, just cut them.By using the DG islanding operation scientifically, not only the power supply reliability of the important loads are ensured, but also more common loads service are restored. Finally, the measures are presented after the faults have been eliminated and the network needs restoring the initial operation if there are unplanned DG islanding operation.