Load Uncertainty Modeling And Risk Assessment Of Static Voltage Stability

In recent years, each part in modern power system has experienced rapid development and change, such as topology, generation and transmission mode, control and management methods and load types, etc. The already existing research findings in the static voltage stability risk assessment domain mainly have the following shortcomings:(1) The load uncertainty modeling is always oversimplified so that few papers have discussed the uncertainty effect composed by nominal load consumption, random variation of load increasing direction and power factor and bus groups on which load growing simultaneously;(2) Constant power load model used in continuation power flow(CPF) algorithm is unrealistic, CPF is limited only to solve deterministic subjects, failing to take probability character into account;(3) Most risk assessment indices are still based on mechanism derivation instead of being constructed from view of probability and different risk assessment aspect composed by bus, line and whole system. Certain evaluation by maximum transmission volume calculation, Jacobi matrix eigenvalues and sensibility analysis are not enough.Based on load uncertainty modeling, CPF algorithm modification and the risk assessment framework of static voltage stability establishment, this paper presents a risk assessment method of static voltage stability.Load uncertainty modeling is accomplished by three steps: Modified fuzzy C mean cluster method(MFCCM), Cholesky decomposition and multi-dimension normal distribution sampling method are adopted to get initial load profile in the load uncertainty modeling; Then, one bus group division system consist of correlation coefficient and mutual information is established to classify buses into different groups according to the similarity degree of their load curve; Finally, to simulate stochastic fluctuation existing in load power factor and growing direction, a probability hyper-cone load growth model is proposed based on Monte Carlo method. Load uncertainty model introduced above is easy to be inserted into the power flow algorithm and more practical to simulate load fluctuation.To consider uncertainty elements and load characteristic in simulation, this paper propose a Monte Carlo- modified CPF algorithm incorporating load uncertain model. Comprehensive load characteristic can be incorporated in the load flow algorithm by deducing equivalent load admittance and modifying admittance matrix; Binary search technology is adopted to improve computation efficiency and overcome the prematurity problem of continuation method simultaneously; Modified CPF algorithm incorporating load uncertain model is combined with Monte Carlo method to simulate probability distribution of every component, especially load uncertainty in this paper. The validity and the outstanding performance of proposed modified algorithm is proved by example analysis and algorithm comparison, it is showed at the end of this chapter.The proposed risk assessment method of static voltage stability in power grid mainly consists of uncertainty modeling of load consumption, improvement of load flow algorithm and construction of risk indicators. The first step is to establish load uncertain model on the basis of historical load data. Next, perform load flow by the MCMCPF algorithm, this algorithm can avoid equations reaching ill-conditioned state and incorporate the composite load characteristic when performing load flow. Finally, calculate the risk indexes and recognize high risk operation region. Risk indexes including system expected loaded risk index, bus low voltage loading risk index, line transmission power ratio risk index and line extreme transmission margin risk index are proposed to describe respectively the static voltage stability risk of buses, lines and system.Taiyuan 110 kV system is studied and analyzed in term of risk ranking and comparison of probability density distribution, and the validity and effectiveness of proposed risk assessment method is verified. High risk area is recognized, the influence of load characteristic model, load growing scenarios and simulation parameters on risk assessment results are also illustrated.