Zhiwei Liao

An Analytic Model for Fault Diagnosis in Power Systems Considering Malfunctions of Protective Relays and Circuit Breakers

When a fault occurs on a section or a component in a given power system, if one or more protective relays (PRs) and/or circuit breakers (CBs) associated do not work properly, or in other words, a malfunction or malfunctions happen with these PRs and/or CBs, the outage area could be extended. As a result, the complexity of the fault diagnosis could be greatly increased. The existing analytic models for power system fault diagnosis do not systematically address the possible malfunctions of PRs and/or CBs, and hence may lead to incorrect diagnosis results if such malfunctions do occur. Given this background, based on the existing analytic models, an effort is made to develop a new analytic model to well take into account of the possible malfunctions of PRs and/or CBs, and further to improve the accuracy of fault diagnosis results. The developed model does not only estimate the faulted section(s), but also identify the malfunctioned PRs and/or CBs as well as the missing and/or false alarms. A software system is developed for practical applications, and realistic fault scenarios from an actual power system are served for demonstrating the correctness of the presented model and the efficiency of the developed software system.

An Online Intelligent Alarm-Processing System for Digital Substations

A flood of alarm messages in an automatic digital substation makes the monitoring task a significant challenge for the operators in a remote control center, especially under fault scenarios. An online intelligent alarm-processing system is developed based on the architecture of the digital substation. First, real-time alarms are classified according to the IEC 61850 standard in order to provide synthesized and organized alarms for the alarm-processing procedure in the next step. Then, a new and systematic alarm-processing approach for digital substations is developed. Two modules (i.e., the generation of candidate hypotheses and the truth evaluation for the hypotheses) are included in the developed approach, and these two modules are operating in parallel in online implementation. This approach could not only determine the fault/disturbance cause but also the missing or false alarms as well as the causes of the false alarms. According to actual application requirements, an online intelligent alarm-processing system is developed and applied in the Xingguo substation-the first digital substation in Jiangxi Province, China. Finally, an actual alarm-processing scenario serves to demonstrate the presented alarm-processing method as well as the developed software system.

An Analytic Model-Based Approach for Power System Alarm Processing Employing Temporal Constraint Network

The alarm-processing problem is to interpret a large number of alarms under stress conditions, such as faults or disturbances, by providing summarized and synthesized information instead of a flood of raw alarm data. Alarm timestamps represent the temporal relationship among event occurrences and consist of rich and useful information for alarm processing. However, the temporal information has not been well utilized in existing alarm-processing methods. The temporal constraint network (TCN) is a type of directed acyclic graph suitable for representing temporal logics. Based on TCN, a new analytic model is developed for alarm processing with temporal information taken into account. Three major modules are included in the developed approach or alarm processor (i.e., alarm selection, event analysis, and result evaluation). In the alarm selection module, reported alarms are divided into related groups. The function of the event analysis module is to find out what events cause the reported alarms and to estimate when these events happen. The result evaluation module is used to identify abnormal or missing alarms. Finally, two alarm-processing scenarios of an actual power system are served for demonstrating the feasibility and efficiency of the developed approach.

An analytic model and optimization technique based methods for fault diagnosis in power systems

An analytic model for fault diagnosis of power system using optimization technique is expressed as unconstrained 0-1 integer programming problem, and consequently faulty equipment identification can be solved by refined mathematical operation. Considering the configuration of automatic devices in modern power systems, such as protective relays and reclosing relays, an improved analytic model and optimization technique-based method for fault diagnosis of power system is proposed in this paper. The evaluation criteria of the presented model is improved considering the relationship of multiple main protective relays, backup protective relays, malfunctioning protective relays and reclosing relays. Improvements of analytic model for fault diagnosis of electric power system based on optimization techniques are presented firstly. A brief description about the modulars and functions of the online fault diagnosis software which is developed by the authors for Jiangsu Provincial Power Company is given. The adopted EMS data acquisition method and simulated online test results for the power system of Jiangsu Power Company are described.

A multi-objective model for transmission network restoration based on multi-agent and Tabu Search

As power system restoration (PSR) is a multi-variable, multi-objective, and multi-constrained nonlinear optimization problem, it is difficult to establish a well-defined mathematical model to reasonably describe it. Given this background, a new distributed model for PSR is proposed in this paper by combining multi-agent technology (MAT) with tabu search method (TSM). The proposed model consists of a number of Crunode intelligent agents (CIAGs) and a single intelligent control agent (ICAG). These agents are co-operative and capable of interacting with each other dynamically to gain local optimal solutions. Based on that, TSM is utilized by ICAG to search for a global optimal solution. Through flexible interaction of agents and heuristic search of TSM, the optimal solution of PSR could be obtained and this is verified by simulation results. In addition, the proposed model is applicable to different electric network topologies and has the necessary flexibility of dealing with various problems during restoration.

A Root-Cause-Analysis Based Method for Fault Diagnosis of Power System Digital Substations

Fault Diagnosis of power systems has attracted great attention in recent years. In the paper, the authors present a Cause-Effect fault diagnosis model, which takes into account the structure and technical features of a digital substation and performs root-cause-analysis so as to identify the exact reason of a power system fault occurred in the monitored district. The Dempster/Shafer evidence theory has been employed to integrate different types of fault information in the diagnosis model aiming at a hierarchical, systematic and comprehensive diagnosis based on the logic relationship between the parent fault node and the child nodes like transformers, circuit-breakers, and transmission lines, and between the root and child causes. An actual fault scenario is investigated in the case study to demonstrate the capability of the developed model in diagnosing malfunctions of protective relays and/or circuit breakers, miss or false alarms, and other faults often encountered at modern digital substations of a power system.

An on-line intelligent alarm analyzer for power systems based on temporal constraint network

The volume of alarm messages could be very large in modern large-scale power systems. Many intelligent methods have been developed for alarm processing in order to provide summarized and synthesized information instead of a flood of raw alarm data. The timestamps of alarms represent the temporal relationship among event occurrences. However, the temporal information has not yet been appropriately utilized in traditional intelligent methods. A temporal constraint network (TCN) is a kind of directed acyclic graph (DAG) which is of promise for the representation of temporal logics. Based on TCN, an on-line intelligent alarm analyzer with the temporal information of alarms taken into account is proposed for on-line operational environment. The advanced intelligent alarm analyzer is able to infer what events cause the reported alarms and to estimate when these events occurred, as well as to identify the abnormal or missing alarms. Finally, a case study is served for demonstrating the feasibility and efficiency of the proposed method.

Application of the Ward's method with TABU-search in the classification of load characteristics

Wards method is an effective method for classification of power system load characteristics. In view of the limitations of the traditional Wards method which can only reduce the number of clusters through amalgamating, an improved measure to optimize the cluster result based on TABU search is put forward along with a statistical method to guide the clustering process and test results. In this paper, the daily load curve is used as the feature vector, while the traditional Wards method is used to find a suitable number of clusters. The clustering results are optimized by means of the method of TABU search. A classification case of the load curves of 744 days in Guangzhou shows the advantage of those methods.

Substation alarm information processing based on ontology theory

Logical alarm information is the key to mining the knowledge that hidden in the power system equipment operation and maintenance information. In this paper, a knowledge mining method of association rules based on ontology is proposed in order to realize the intelligent analysis of substation alarm information. Firstly, the alarm information ontology is based on the structure of substation configuration language, and it uses the methods of short text classification and data generalization, which lays the foundation for hierarchical database of the Apri ori algorithm. Secondly, this paper uses the ontology method to construct ontology class tree that integrates the unstructured records based on three dimensions of time, frequency and range. And then we traverse data, attributes, and relationships in the tree, so as to generate the data sources that computer can use for algorithms in data mining. Finally we use the improved Apriori algorithms based on alarm ontology to calculate and draw conclusions. Experimental results show that the efficient generation of association rules and assertions rules verify the validity of the method.