| As the traditional fossil energy sources exhaust, and the environmental pressure and the load demands increase, it is difficult to adapt to the human sustainable development for traditional power generation methods. Solar energy is a kind of clean, environmentally friendly, relatively uniform distributed and huge potential renewable energy. For human beings, it is inexhaustible source of energy. And it is significant important for energy supply and energy security to develop the Photovoltaic(short for PV) generation in grid-connected DG form. The output power of grid-connected PV is directly determined by light intensity received. However, the light intensity is random, intermittent and non-scheduling. With the increase of PV capacity, it is necessary to make research on the grid-connected PV’s impact on power system.The probability of PV output make it difficult for traditional power flow calculation to reflect the grid-connected PV’s impact on power system. So the probabilistic theory should be considered into power flow calculation. In the probabilistic power flow calculation, Probability and statistics methods are used to solve the probabilistic elements. Combined with the derivation between cumulant and Gram-Charier series expansion, the probabilistic distributions of voltage and power flow are obtained, and grid-connected PV’s impact on power system can be reflected comprehensively. This method is programmed in Delphi programming language and applied in25nodes distributed networks in Jiangxi to get the curves of PDF or CDF of each node. Node voltage safety probability can be obtained through the statistic analysis of different time durations. In this method, calculation and statistic under different weather types, the result can be referenced by distributed network planning and operation. At the end of thesis, PV output prediction based on SVM is introduced, and programmed into software. This method is proven effective by practical example. |
PDF Full Text Request