Probability Load Flow Calculation Method Research Considering Power Relativity Of Wind Farms

In recent years, the rapid development of wind power makes its own random characteristics impacting to grid growing interested by researchers. Coupled with a large number of uncertainties inherent factors (such as the random change of power and load; the failure of power, transmission and substation equipment, etc.) bring difficulties of large computation and hardly reflect situation comprehensively to the basic power flow calculation in analysis of power system operation. Traditional deterministic analysis ideas and tools may have been unable to meet the needs technically; the analysis methods of considering random factor are gradually paid attention. To address this issue this paper combines theoretical knowledge of semi-invariants,Gram-Charlier series expansion and Von Mises together and gets the node voltage or power flow cumulative distribution function through the concept of probability in linear AC model of power flow equations.Random variables are often assumed to be independent of each other in probability load flow. But the absorption of reactive power is the function of active power and terminal voltage towards to the wind farms composed by a large number of asynchronous generators in our country combining the simplified equivalent circuit. Without considering the relevance of active and reactive power will result in some extreme combinations such as active efforts to zero but reactive power efforts to maximum or active efforts to maximum but reactive power efforts to zero, it may bring great error to the calculation of probability load flow. If the error results are applied in the planning, operation or analysis fields of power system, wrong results may be got.This paper proposes a probabilistic load flow method considering active and reactive power relativity of wind farms to solve the problem that active and reactive power of wind farms are not independent to each other in probability load flow calculation. This method combines active and reactive power of wind farms with corresponding weighting values of sensitivity matrix and converts them to an independent random variable. It makes probability load flow calculation by cumulant to get the node voltage probabilistic distributions of power system with wind farms integrated on the basis of full consideration about randomicity of wind speed and other stochastic disturbances, such as varying loads and generation unit. IEEE-14 system integrated with wind farms is been analyzed. The test which uses Monte Carlo method as a standard shows that probability load flow method which active and reactive power relativity of wind farms is taken into account is more accurate than general probability load flow method.