| This dissertation is supported by National Natural Science Foundation of China(50977094) and Chinese Universities Scientific Fund(CDJZR11150012).Expected reliability indices are used in traditional reliability evaluation, but theycan not reveal the distribution rule and structure characteristic of reliability indicesbecause reliability indices are random variables indeed. The reliability parameters arebasic input data no matter which evaluation models and algorithms are applied, so theiruncertainty have great effect on the evaluation results. Reliability evaluation resultsespecially the probability density function for bulk power systems affected by reliabilityparameter uncertainty is an important topic which needs to solve urgently.From the viewpoint of random law, the reliability indices of bulk power systemsare characterized by their probability density distributions in this dissertation. Inaddition, their probability distribution features affected by parameters uncertainty arefurther researched and calculations of the marginal probability distributions of reliabilityindices are realized.â‘ Mathematical model of probability distribution of bulk power systemconsidering parameter uncertainty is built and the two-loop Monte-Carlo simulation isused to calculate it. The convergence criterion of the two-loop Monte-Carlo simulationis defined. Then the marginal probability distributions of reliability indices arecalculated considering partial reliability parameters by this method. This method is areference method due to its simplicity.â‘¡The probability distribution of bulk power system considering parameteruncertainty is calculated through mixing the point estimate method and bulk powersystem reliability evaluation. Compared with the two-loop Monte-Carlo simulation, thepoint estimate method has higher efficiency. The point estimate method is verified to bevalidity through comparing to the two-loop Monte-Carlo simulation. Then the marginalprobability distributions of reliability indices are calculated considering all parametersuncertainty by2m+1point estimate method.â‘¢The marginal probability distribution functions for component time to failureand time to repair are deduced, and expressed discretely using numerical integration forsimplicity. Based on these marginal distributions for components failure and repair, thesequential Monte Carlo simulation is adapted to calculate the marginal probabilitydensity functions of system reliability indices, which explores a new way to research probability distribution characteristics of system reliability incorporating parametersuncertainty. It can be seen from the results obtained for the RBTS and IEEE-RTS79power systems that the two-loop Monte-Carlo simulation could only be a referencemethod because its low efficiency, and the point estimate method is not suitable forlarge scale power system, however, the developed Monte-Carlo simulation is notrestricted to the scale of the power system, so it could take into account the effect whichparameters uncertainty have on probability distribution for bulk power systemefficiently. |
PDF Full Text Request