Undetected short circuit faults are a significant problem in power transformers and can eventually develop into catastrophic faults. At present, frequency response analysis (FRA) is one of the well-recognized diagnostic tools for the detection of winding faults, but it has some limitations, such as a low signal-to-noise ratio (SNR) and instability caused by changes in the measuring voltage. In this paper, a novel method called sweep frequency impedance (SFI) is proposed to address the difficulties that arise from FRA. Based on the mechanism of this new method, a nondestructive testing system was established to demonstrate the advantages of SFI measurements. The SFI test system has a better stability, repeatability, and SNR by comparing it with the FRA test system. Moreover, FRA and SFI curves obtained under the same conditions was symmetrical about a specific straight line above 10 kHz, and the SFI value at 50 Hz is equivalent to the short circuit impedance (SCI) value of a transformer. These results indicate that the existing criteria of FRA and SCI methods can be used in the SFI method to detect transformer faults. Finally, the experiments on a special oil-immersed testing transformer demonstrate that the SFI detection system is feasible, sufficiently sensitive to detect short circuit faults and able to quantify the level of the fault.Index Terms -Transformer winding deformation, frequency response analysis (FRA), short circuit impedance (SCI), sweep frequency impedance (SFI), short circuit fault.
Reducing the overall cost and improving the reliability are the two primary but often conflicting objectives in power system. Preventive-maintenance schedules thus need to be optimised to trade-off among multiple objectives. An integrated methodology with three functional blocks is proposed in this study. The first block models the stochastic deterioration process of individual component with a continuous-time Markov model, of which transition rates are influenced by different maintenance extents and aging of components. The second block evaluates the reliability of a composite power system, taking into account the configuration and failure dependence of the system. Particularly, this block identifies the minimum cut sets with consideration of protection trip and operational switching. The third block employs the Pareto-based multiobjective evolutionary algorithm to find the optimal solutions in a large search space and provide a holistic view of relationships among conflicting multiple objectives. A novel representation of maintenance activities is introduced in this study specifying both the maintenance timings and extents, and is proven to outperform the authors' previous representation, specifying the maintenance frequencies only. Optimisation of the reliability, maintenance failure costs is carried out on the Roy Billinton Test System (RBTS) demonstrating the potential of this approach in handling complex systems.
Nomenclature
RBTSRoy Billiton Test System MTTF mean time to failure MTTR mean time to repair EUE expected unserved energy C o,a operation cost for component a C ins cost of inspection of component a T min,s times of minor maintenance in state s for one component T maj,s times of major maintenance in state s for one component C min,s average cost of minor maintenance for component a in state s C maj,s average cost of major maintenance for component a in state sC f,a expected failure cost of component a p i,f failure probability in interval i C avef,a average cost of failure for component a l cs and m cs failure rate and duration of one cut set l A , l B , l C and l D failure rate of events A, B, C and D r A , r B , r C and r D duration of events A, B, C, and D U cs annual duration of one cut set l top and U top failure rate and annual duration of top event in one interval n number of minimum cut sets leading to the top event C sysO overall operation cost C sysF expected failure cost of system 930
Reducing the overall cost and improving the reliability are two primary but often conflicting objectives for composite power system. Scheduling of appropriate preventive maintenance requires optimization among multiple objectives. In this paper, an integrated methodology with three functional blocks is proposed. In the first block, the stochastic deterioration process of individual components is formulated as a maintenance-dependent continuous-time Markov model. Reliability of a composite power system is evaluated in the second block. In the third block, Pareto-based multiobjective evolutionary algorithm is proposed for providing a holistic view of conflicting relationships among multiple objectives. The second block extends the authors' original minimum cut sets, by identifying the loss of energy of a load point due to not only a loss of continuity within a substation, but also a loss of continuity and a violation of transfer limit between these substations. This work also extends the representation of maintenance activities to both maintenance timings and extents. The present approach is applied to the IEEE Reliability Test System (IEEE RTS) for optimizing its reliability, maintenance, and failure costs. Results demonstrate the potential of the present approach for handling complex systems, and substantiate its improvement over the authors' previously reported work.Index Terms-Continuous-time Markov model, inter-substation loss of continuity and violation of transfer limit, minimum cut sets, Pareto-based multiobjective evolutionary algorithm, substation loss of continuity.
The maximum likelihood estimation is a widely used approach to the parameter estimation. However, the conventional algorithm makes the estimation procedure of three-parameter Weibull distribution difficult. Therefore, this paper proposes an evolutionary strategy to explore the good solutions based on the maximum likelihood method. The maximizing process of likelihood function is converted to an optimization problem. The evolutionary algorithm is employed to obtain the optimal parameters for the likelihood function. Examples are presented to demonstrate the proposed method. The results show that the proposed method is suitable for the parameter estimation of the three-parameter Weibull distribution.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.