Uncertainty Theory Based Power Supply Equipment Maintenance Studies

Power supply equipment maintenance is one of the most important assignments for power utilities, and it is the main cause of power outages and reducing of power grid reliability. However, at present, maintenance arrangement generally depending on the personal experience neither guarantees efficiency nor meets ever-increasing requirements of information and automation. Although the research on power supply equipment maintenance is at its preliminary stage, it exposes attracting attention of researchers in recent years. With the electricity market-oriented revolution, more uncertainties concerning power supply equipment maintenance issues are requested to be handled. In this dissertation, the author introduces the uncertainty theory into maintenance issue, and thus making innovation in the several aspects.First of all, various types of maintenance concepts are reclassified and compared, including maintenance strategy, maintenance mode, maintenance optimization, maintenance implementation and management, etc. By summarizing the principle of maintenance strategy making, a maintenance strategy optimization model is set up based on credibility theory. Taking modeling fuzzy comprehensive evaluation of various influencing factors as its core, this dissertation considers the fuzzy expected value and variance of equipment importance as important parameters in formulating maintenance strategy. The optimized results give memberships corresponding to various maintenance modes.Secondly, this dissertation analyzes the stochastic characteristics such as load, generator output, line status, maintenance resources, etc. of the transmission line maintenance optimization problem, and establishes two types of transmission line maintenance optimization models based on stochastic expected value bi-level programming and expected chance constraint bi-level programming separately. The former evaluates the probability and consequences of insecurity and quantifies the risk item in form of static safety controlling cost. The latter considers system reliability index during maintenance period as one of the constraints into the model.Third, this dissertation introduces credibility theory to describe and process various fuzzy factors in distribution network maintenance optimization. Fuzzy bi-level programming models aiming at the minimization of the total maintenance fuzzy cost are established, and the objective function takes maintenance expense and load transfer during each period into consideration. A hybrid intelligent optimization algorithm integrating fuzzy simulation, heuristic load transfer, particle swarm optimization, and tabu list is adopted to solve the model. Optimized results not only validate the fuzzy expected value and variance of objectives, but also compare optimization values according to different confidence levels.Finally, a new method based on dependent-chance programming, which is the trade-off of the economy and reliability in the process of maintenance, is proposed. Simultaneously, the stochastic factor and fuzzy factors are taken into account in the distribution network maintenance optimization, and the probability theory is applied to describe the randomness of equipment status and system operation mode, and in addition the credibility theory is used to analyze and model equipment failure rate, load prediction value. The credibility concepts of economic goal and reliability index meeting certain expectations are introduced. An optimal maintenance arrangement scheme can be obtained, via maximizing credibility of uncertainty event's happening.In this dissertation, power supply equipment maintenance optimization model based on uncertainty bi-level programming is approximately the same as the actual maintenance scheduling making processing. Hence, it possesses instructive significance and thus practical to be acknowledged by the power utility.