Distribution Network Reliability Evaluation Model Study Incorporating Distributed Generators

This paper is funded by national natural science fundation of China project(probabilistic risk assessment and optimimal configuration model research for flexiblegeneration and transmission system, project number:50977094), the basic scientificresearch business expenses in special scientific research project of central university"power grid parameter uncertainty analysis of probabilistic risk assessment andprobability density analytical calculation research"(CDJZR11150012). As the greatconsumptions on fossil fuels bring environmental problems, exploiting clean, flexibleand renewable distributed generation has become a heat debated issue over the world.However, the high penetration of distributed generation converts conventional radialdistribution network with single power supply to a new one where the customers areconnected and multiple power resources exist. Thus when considering feeder failures,part of the distribution system is probably separated into islands. Moreover, with theprobabilistic and intermittent nature of renewable distributed generators, the reliabilitymodel and techniques have changed. This paper focuses on the reliability assessmentmodel of distribution system incorporating distributed generators, the key points arelisted below:Firstly, this thesis presented the research’s purpose and significance, summarized thedomestic and abroad reliability assessment research of conventional distribution network anddistribution network with distributed generators, introduced the frequently-used distributionnetwork reliability index, reliability model of component and series-parallel systems reliabilityindices formula, as well as the type and model of distributed generators, the access way, thegrid-connected mode, the mode of operation and island grid operation.Then, based on hierarchical region, the reliability model and algorithm forcomplicated distribution system is proposed.The traditional distribution systemreliability assessment model is mostly based on the failure mode and effects analysis,which is proceeding from a known cause to a necessarily related effect, i.e.enumerating fault events and then analyzing their effects on load points. Throughconverse thinking this paper adopts the idea which proceeds from results to the cause,i.e. enumerating load points, searching the distribution components set which cancause load point interruption with different outage duration. According to the differentoutage duration of load points, a breadth-first-search technique is used in this paper which taking malfunction of the circuit breakers and fuses into account. The conceptsof load point repair zone, isolation zone, isolation switching zone and zero zone aredefined and a hierachical algorithm is also proposed.Once more, the reliability model of grid-connected photovoltaic system (PVS) isbuilt.the effect of Photovoltaic-diesel hybrid generation system on the reliability ofdistribution network is deeply researched in this paper. Firstly, by analyzing PVScomponent failure modes and their impacts on PVS output power, the reliability modelof PVS is built. Then the probability model of PVS output power was presented byconsidering both illumination uncertainty and PVS component random faults. Adoptingdiscrete convolution method to obtain probability for successful operation of intentionalisland, and seeing intentional island as a load point,with the proposed method, the islandis regarded as a load point and the formula to compute load point reliability indexes isdeduced.Lastly, considering the uncertainty of distributed generation output and loadvolatility, Discrete convolution method’s computation is complex and simplifiedmodel exists error.This paper further presents a distribution network reliabilityevaluation model incorporating distributed generation based on an efficient pointestimate method. The proposed method can transform the probabilistic problem into adeterministic mathematical one. The estimate points are conducted on the basis of theprobability distribution of distributed generation output and load, and then thedistribution network reliability index is estimated by the functional relationshipbetween the obtained estimate points and distribution network reliability.The reliability assessment of distribution network with and without distributedgeneration is conducted on IEEE-RBTS Bus6.The study results confirm that:theproposed reliability algorithm based on hierarchical region is valid. It can also be seenthat incorporating the probabilities of stuck breakers as well as fuses isnecessary.Reliability of load points and system closely related to installation positionof circuit breaker and isolating switch in network, and installing appropriate breakersand fuses for some load points and system can considerably improve reliabilityperformance. Another conclusion is that the penetration of photovoltaic generation canenhance the reliability of island load points as well as the system, and the extent of thisimpact depends on the capacity of diesel engine. The photovoltaic-diesel mixedgeneration system installed at the end of the distribution network can better improvethe island load point and system reliability performance. When evaluating the reliability of distribution system incorporating distributed generation, the result usingpoint estimate is very close to that using discrete convolution, with adequate accuracy.