Large-scale integration of wind power into grid leads to uncertainty of generalized load flow

making new requirements on generalized load modeling. How to accurately build a generalized load model considering the stochastic variation and spatial correlation of uncertainties has become an urgent problem to solve. A novel method considering spatial correlation based on probability statistics is proposed to learn and abstract steady-state generalized load characteristics. According to the power direction of bus nodes connecting wind farms

the node characteristics are divided into source characteristics and load characteristics; then

the measured active power is used to represent characteristics and segment. The range of the segments is determined adaptively and the probability distribution is got by probability statistics. The radial basis function (RBF) neural network is used to abstract the node characteristics considering relevant characteristic parameters between segments divided by different nodes based on the spatial correlation method. Then the generalized model is applied to the actual case about the risk analysis. It is verified by simulation results that the new method can not only build a model accurately and comprehensively

but also refine system space risk forming scenarios by introducing spatial correlation to generalized load modeling

which is useful for dispatching and controlling with wind power integration taken into account. This work is supported by National Natural Science Foundation of China (No. 51177091).