Research On Distribution Network Optimal Planning With Distributed Generation

Scientific and rational planning for power system is the premise of system safe,reliable and economic operation. Distributed generation provides users withconvenient and environmentally friendly energy. Combining distributed generationwith main distribution network is the developing trend of smart grid. The connectionof distributed generation with distribution network will bring profound revolution tothe traditional distribution network planning and operation. In that background, thisarticle about optimization of power system planning studies consists of two parts:load forecasting method with distributed generation and distribution networkexpansion planning with distributed generation.Load forecasting is the precondition and foundation of network pla nning. Sincethe installation location and capacity of distributed generation is randomness anduncertainty, Its output will also be affected by climate conditions and many otherfactors, which will increase the difficulty for load forecasting with distributedgeneration. Summarizing the considering load forecasting with distributed generationbased on analysis traditional load forecasting methods. Support vector machinetechnology with strong learning ability can deal with small samples and with goodaccuracy, but there’s some difficulty to choose the appropriate parameters. For thisreason, the improved particle swarm optimization algorithm (FAPSO) to optimize therelevant parameters of support vector machine which is found to overcome theshortcoming of the traditional way is proposed. The disturbance for extreme valuealso added in the algorithm to avoid the algorithm to fall into local optima. A modelbased adaptive particle swarm optimization algorithm and support vector machine forload forecast with distributed generation is established. The results show the proposedmethod is better than the traditional support vector machine.This paper based on the impact of distributed generation networking to thedistribution network planning and operation, established a bi-level model fordistribution network expansion planning with distributed generation, the upper issources planning, and the lower is distribution network optimization. To solve theproblem of traditional planning without considering the types and run time of DGs, amethod based on annual duration load curve and power cost characteristic curve todetermine DG type and working time of each power is proposed. We choose the candidate locations for distributed generation with considering the impact ofdistributed generation connect with distribution network about network losses andVoltage quality. This paper adopted an improved genetic algorithm for distributionnetwork programming model to optimize the solution, used a simple binary encodedmethod for the chromosome and improved the traditional back-forward sweep powerflow calculation based on level correlation matrix. Graph theory is also used inrestoration the probable infeasible problem in distribution network planning. At last,the modified IEEE33nodes simulation example verifies that the proposed method isreasonable and effective.