Yingkai Bao

Evaluating equipment reliability function and mean residual life based on proportional hazard model and semi-Markov process

Based on Monte Carlo technique, a novel method is proposed to evaluate the reliability function (RF) and the mean residual life (MRL) of a power system component suffering from degradation. It is assumed that both the age and the health condition can affect failure rate. The Coxs proportional hazard model (PHM) is used to customize the hazard rate of component with both age and health information. PHM consists of two parts: baseline function and link function. The baseline function represents the aging process and the link function represents the influence of covariant, such as the health condition. The health condition is divided into n-stages which confirms to a semi-Markov process. Based on the Monte Carlo technique, an algorithm is developed to find out the RF and MRL of a new or aging piece of equipment. We calculate the RF and MRL in the two cases respectively: (1) a new component and (2) an aging component survived until t0. The case study shows that the proposed method can reflect the influence of both age and health condition on RF and MRL reasonably.

Comprehensive decoupled risk-limiting dispatch

Risk-limiting dispatch (RLD) is a promising approach to deal with the increasing uncertainty in both supply side and demand side of power systems. However, the conventional RLD has two drawbacks: i) it is difficult to set the value of acceptable risk level; ii) it has limitations to address the scenario with more realistic but non-Gaussian and heavy-tailed renewable output forecasting errors. To resolve these drawbacks, a conceptual framework of comprehensive decoupled risk-limiting dispatch (CDRLD) is proposed in this paper. We consider the severity level and the probability level of operating risk in a unified framework. The acceptable probability level which corresponds to the loss of load probability (LOLP) is extended to an interval number, while the acceptable severity level is consistent with the expected demand not supplied (EDNS). The whole dispatch problem is solved in a decoupling scheme, in which the severity level of operating risk is guaranteed through the severity feasibility check. Numerical simulations indicate the effectiveness of the proposed CDRLD for enhancing the economic benefits without loss of operating reliability.

Impact of human error on electrical equipment preventive maintenance policy

Preventive maintenance of electrical equipment is of great significance to maintain power system stable, prolong the service life of equipment and reduce the system power loss. The effect of maintenance is heavily affected by human factors, and human errors might have a negative influence on maintenance results. In view to this situation, this paper analyzes human factors in maintenance firstly, and then establishes a periodic maintenance model considering the impact of human errors. Numerical example analysis illustrates that the optimal policy which maximizes the availability is affected by human errors, percentage of maintenance failure and repair time. At last, this paper introduces a practical method to quantify human error probability, and several measures are suggested to manage human errors in maintenance.

Multi-objective distribution network reconfiguration based on system homogeneity

Distribution network reconfiguration is a multi- objective and discrete nonlinear combinatory optimization problem. In this paper, a novel multi-object distribution network reconfiguration method is proposed based on system homogeneity. Firstly, the relationship between power loss and reliability of systems with different homogeneity is analyzed. Then, to enhance system reliability and homogeneity, a new multi-objective model is established by coupling the reliability index with system homogeneity. Based on the adaptive settings and fast iteration method, an improved multi-objective harmony search algorithm is proposed to solve the model. The test results on the 33-bus system show that the proposed method is corrective and effective.

Power system disaster-mitigating dispatch platform based on big data

Based on the big data technique, this paper proposes an implementation scheme of the Disaster-Mitigating Dispatch Platform (DMDP), which integrates the functions of dynamic risk awareness, comprehensive analysis, decision making and multi-dimensional visualization associated with weather and natural disasters. First, the framework of the DMDP based on Hadoop 2.0 is presented. Then, the demand of mass, multisource and heterogeneous data, such as structured, semi-structured and unstructured data, for power systems disaster preventing is analyzed. The functions for disaster mitigating could be divided in three stages: prior to the disaster, DMDP performs big data mining based on history database, dispatches human resources and materials and formulates preventive measures for contingencies. During the disaster process, DMDP monitors conditions of the equipment, system and environment. Once a major contingency occurs, DMDP can alarm, identify and give corrective actions. After the disaster, the failure will be investigated and the efficiency of disaster preventing will be assessed. The proposed DMDP could serve as a useful decision-making tool for the operators.

An adaptive harmony search algorithm based on positive feedback for network reconfiguration

Distribution network reconfiguration is a complex and discrete nonlinear combinatory optimization problem. In this paper, an adaptive harmony search algorithm based on positive feedback (PFAHSA) is proposed to solve the problem for reducing power loss in distribution systems. The PFAHSA incorporates a local optimal method based on positive feedback mechanism into traditional harmony improvising criteria for better utilization of harmony memory (HM) and higher convergence speed. The algorithm parameters are updated adaptively by modified entropy, which measures the diversity of HM, to avoid prematurity and improve search success rate. The proposed method is testified on the IEEE 33- and 69-bus systems and is compared with other conventional approaches. Simulation results show that the proposed method features faster optimization speed and better global convergence performance.

Analysis of human reliability in power system switching operation considering dependency of operators

The inter dependency between actions of operators is prevalent in power system switching operation, which is necessary to be considered along with restoration when performing human reliability analysis. This paper proposes a novel method of human reliability analysis to evaluate the human error probability considering the dependency between operators. Firstly, the concept of human reliability in power system and the cognitive reliability and error analysis method (CREAM) are introduced. Then, through the analysis of power system switching operation, the factors influencing dependency level are identified and the analytical model of dependence is constructed. Based on fuzzy theory and Bayesian network, a method is proposed to quantify the restoration coefficient. Combined with CREAM extension method, the proposed method can be used to evaluate the final human reliability.

A Bayesian network approach for human reliability analysis of power system

Along with the improvement of equipment reliability, human error has become a great threat to the power system reliability and safety. However, the research of human reliability analysis in power system is still in its infancy. There is still little approach for quantitatively measuring the human reliability of power system. In this paper, the definition of human reliability of power system and a human error causal framework are given firstly. Secondly, with the suitable Performance Influencing Factors(PIFs) selected, the probability inference model based on Bayesian network for human reliability analysis is proposed. Finally, a case study shows that the proposed methodology can integrate organizational factors, situational factors, and individual factors to quantitatively measure the human reliability of power system. This approach provides forceful support for improving the human reliability of power system and has a good prospect.