Research On Probabilistic Evaluation Methods For Static Voltage Stability In Power System

It is a vital guarantee to maintain secure and stable operation of power system for the sound development of national economy and people's normal life. Voltage instability is widely recognized as a significant threat of power system blackout. Hence, voltage security assessment is one of the most important tasks in control centers of power systems. Power system is in nature a large one with random characteristics, and power industry deregulation results in more highly stressed and unpredictable operating conditions. Therefore utilizing conventional deterministic methods of voltage stability analysis, system dispatchers and operators can not fully identify the voltage security level. It is of theoretical and practical significance to research on probabilistic evaluation of voltage stability so as to improve power system security and economy.The problem of voltage stability evaluation becomes more complex when random factors are taken into account. This thesis is devoted to probabilistic assessment algorithm of static voltage stability considering uncertainty of the load level and load parameters, bus load correlation and load forecasting uncertainty as well as random branch outage which play an important role in voltage stability. Main research works and contributions are summarized as follows:(1) Emphasizing on uncertainty of the load level and load parameters, a probabilistic assessment algorithm of voltage stability based on Monte-Carlo simulation and modal analysis is proposed. On the basis of synthetic load model that involves induction motor and ZIP load, probabilistic sampling techniques are employed to cope with these uncertainties mentioned above, and voltage stability is evaluated in the form of a set of probabilistic indices. Case studies stress on investigating the effects of load uncertainty and the proportion of motors on probabilistic assessment of voltage stability.(2) Aiming at bus load correlation and load level uncertainty existing in practical system operation, a method for probabilistic voltage stability evaluation based on the Fuzzy C-Means (FCM) clustering algorithm is presented. The proposed method employs FCM algorithm to classify and synthesize bus load levels over the research period of time, and multidimensional normal distribution sampling technique is adopted to deal with bus load correlation and load level uncertainty synthetically. Tests on IEEE standard system demonstrate the feasibility and validity of the proposed method, and the effects of the number of fuzzy clustering, bus load correlation and uncertainty on voltage stability are investigated.(3) How to take random branch outages into account is a tough problem of voltage stability evaluation. In this thesis an efficient point estimate method based is proposed, and a nonlinear model based on load margin optimization model is adopted. While dealing with the effects of random branch outages and injection power uncertainty on the distribution of load margin harmoniously, the proposed method overcomes the shortcoming of the linearized method which has difficulties in coping with reactive power generation constraints. The proposed method avoids linearizing power flow equations and has characteristics of simpliness, accuracy and low time-consuming. Tests on several systems demonstrate the feasibility of the method, and the assumptions in calculation steps along with the errors caused are investigated. Several conclusions are presented.(4) Aiming at random direction of bus load increase, a method for determinating the closest voltage collapse point based on an improved particle swarm optimization algorithm is proposed. An optimization model is built to search for the random direction of bus load increase, and to obtain the shortest distance to voltage instability from the global boundary of voltage stability domain to base operating point considering various system constrains. By adding a constraint to the proposed optizimation model, the calculation of an on-scenario critical point may be realized expediently. During the search, several measures are presented to improve the efficiency. In the numerical experiments the feasibility and good performance of the proposed algorithm is demonstrated.(5) Introduce modern intelligence algorithm into calculation of voltage collapse critical point. A genetic bacterial colony chemotaxis algorithm (GBCC) based on bacterial chemotaxis (BC) algorithm and genetic algorithm (GA) is proposed and applied to calculating voltage collapse critical point. The proposed algorithm searches for optimum by bacterial colony instead of individual bacterial and guides mutation direction instead of stochastic direction. Convergence speed and the ability of global search are improved due to information exchange among individuals in the group. An optimization model for calculating voltage collapse critical point based on GBCC algorithm is built. The advantage of using the new method to find the maximum loading point is the ease of formulating and solving the problem with various constraints without the need to form any kind of matrices or get the inverse of these matrices, which may be in some cases difficult. Several numerical tests are given to demonstrate the validity and efficiency of the proposed method.