Petri Nets for Fault Diagnosis of Large Power
Generation Station Abstract –In this paper, a simplified fault diagnosis method based on Petri nets is proposed to estimate the faulty item/section(s) of a large power generation station.The Petri nets are used as a modeling tool to build fault diagnosis models of item/section(s) of power station which aim to diagnose accurately the faults when a large amount information of SCADA system are detected in the control room.It can diagnose and estimate the faulty item/section(s) correctly for multiple faults as well as simple faults.In order to testify the validity and feasibility of that method, a computer simulation of High Dam power generation station is used.It is shown from three study cases that Petri nets fault diagnosis method has many merits such as: accurate fault diagnosis results, easy and flexible correctne of Petri net fault diagnosis models for each item/section(s).Keyword –Petri nets, fault diagnosis, power station.
1.INTRODUCTION Fault diagnosis of a large power generation station can be a proce of discriminating faulted power station item/section(s) by tripping of their protective relays and circuit breakers.Therefore, it requires information from SCADA system.When the information arrives at the control room, the operators analyze the data and diagnose the faulted item/section(s).The accuracy and speed of the diagnosis proce depend entirely on the experience of the operators.However, as the complexity of power station increases, especially in the case of multiple faults, a lot of alarm information are transmitted to the power station control room.Under such situations, the operator should diagnose the faulty section rapidly and accurately.For this reason, the fault diagnosis systems have to be developed in the control rooms to aist, support and help the operators to carry out their tasks in diagnosis procees.Resent researches have been made toward developing fault diagnosis system.Most of these efforts are based on Expert Systems (ES) [1–4].Artificial intelligence approaches, such as, artificial neural networks [5–7], genetic algorithm (GA) [8], family eugenics based evolution theory [9], immune algorithm [10] are developed.Two corresponding fault diagnosis researches for power generation station based on fuzzy relations and Bayesian networks respectively are given in reference [11, 12].Petri nets have characteristic of the parallel information proceing, concurrent operating function and considered as a very suitable and useful modeling tool.Some methodologies of modeling and analysis for the fault diagnosis of power system with Petri nets are proposed [13 – 16].The fault diagnosis systems are used widely in power systems and substations.In this paper, a simplified fault diagnosis method based on Petri nets for a large power generation station is proposed.This power generation station includes: generation units, step up power transformers, station service transformers, station buses and autotransformers.The proposed fault diagnosis method utilizes the information of the protective relays and circuit breakers to build Petri net model for each faulty item/section(s) of a large power generation station.The faulty item/section(s) can be diagnosed and estimated from the final state of the fired Petri net.Moreover, a comparison of effectivene and performance of the proposed Petri nets, fuzzy relations and Bayesian networks is presented.The proposed method is tested on 15.75/500 kV High Dam power generation station which affiliates to Hydro Plants Generation Company (HPGC) in Egypt.The testing results demonstrated that proposed method is easy reasoning, strong practicability of fault diagnosis models and finally, it aists and supports the operator in control room of the power station to make the right decision.2.MODELING METHOD OF PETRI NETS 2.1 Petri Net Definition A Petri net is a one of several mathematical and graphical representations of discrete distributed systems [17, 18].As a modeling language, it graphically depicts the structure of a distributed system as a directed bipartite graph decision.2.2 Petri Net Modeling Power The typical characteristics exhibited by the activates in a dynamic event-driven system, such as concurrency, decision making, synchronization and priorities, can be modeled effectively by Petri nets [20]: Sequential Execution; In Fig.2 (a), transition 2 t can fire only after the firing of 1 t .this imposes the procedure constrain “ 2 t after 1 t ”.Such procedure constrains are typical of the execution of the parts in a dynamic system.Also, this Petri construct models the casual relationship among activates.Conflict; Transitions 1 t and 2 t are in conflict in Fig.2 (b).Both are enabled but the firing of any transition leads to the disabling of the other transition.Such a situation will arise, for example when a machine has to choose among part types or a part has to choose among several machines.The resulting conflict may be resolved in purely non-deterministic way or in a probabilistic way, by aigning appropriate probabilities to the conflicting transitions together.Concurrency; In Fig.2 (c), the transitions 1 t and 2 t are concurrent.Concurrency is an important attribute of system interactions.This is a neceary condition for a transition to be concurrent is the existence of a forking transition that deposits a token in two or more output places.Synchronization; It is quite in a dynamic system that an event requires multiple resources which related to circuit breakers and protective relays in this paper.The resulting synchronization of resources can be captured by transitions of the type shown in Fig.2 (d).Here, 1 t is enabled only when each of 1 p and 2 p receives a token.The arrival of a token into each of the two places could be the result a poibly complex sequence of operations elsewhere in the rest of the Petri net model.Eentially, transition 1 t models the joining operation.Mutual Exclusive; Two procees are mutually exclusive if they cannot be performed at the same time due to constraints on the usage of shared resources.Figure 2 (e) shows this structure.For example, a robot may be shared by two machines for loading and unloading.Two such structures are parallel mutual exclusion and sequential mutual exclusion. Priorities; Such a modeling power can be achieved by introducing an inhibitor arc.The inhibitor arc connects an input place to transition, and is pictorially represented by an arc terminated with a small circle.The presence of an inhibitor arc connecting an input place to a transition changes the transition enabling conditions.In the presence of the inhibitor arc, a transition is regarded as enabled if each input place connected to the transition by a normal arc (an arc terminated with an arrow).Contains at least the number of tokens equal to the weight of the arc, and no tokens are present on each input place connected to the transition by the inhibitor arc.The transition firing rule is the same for normally connected places.The firing, however, does not change the marking in the inhibitor arc connected places.A Petri net with an inhibitor arc is shown in Fig.2 (f).1 t is enabled if 1 p contains a token, while 2 t is enabled if 2 p contains a token and 1 p has no token.This gives priority to 1 t over 2 t .
Fig.2 Petri net primitives to represent system features.
本文节选自《基于Petri网大型发电站故障诊断》（《艾因夏姆斯工程学报》）中的部分章节
专业及生僻词汇：
Fault Diagnosis 故障诊断；Generatoin Station 发电站； Petri nets Petri网，一种建模方法；SCADA在线监控系统；
multiple faults 多重故障；discriminate 区别，辨别；relay 继电器； circuit breakers 继电器；omplexit 复杂性；Expert Systems 专家系统； Artificial intelligence 人工智能；genetic algorithm遗传算法； artificial neural networks人工神经网络；Bayesian贝叶斯定理的； genetic algorithm家族优生学；fuzzy 模糊的；genetic algorithm 免疫算法； affiliates 附属公司；practicability实用性；discrete 离散的； distributed 分布的、分散的；bipartite 双边的、双向的；
concurrency 同时发生的；synchronization 同步；Sequential 连续的； non-deterministic 不确定性的；Mutual 共同的；constraint 约束、局促； inhibitor arc 抑制弧；inhibitor 抑制剂；High Dam高坝；
基于Petri网的大型发电站故障诊断
摘要：本文提出了一种基于Petri网的简化故障诊断方法用来判断大型发电站的故障元件或区域。用Petri网作为一种建模工具来建立发电站内元件或区域的故障诊断模型，其目的是当控制室内的监测控制和数据采集系统检测到大量信息时能够准确的诊断所出现的故障。它可以正确的诊断并估计故障元件或区域内出现的多重故障或单一故障。为了证明该方法的有效性和可行性，对高坝发电站利用该方法进行计算机仿真。结果从三个研究案例中显示出Petri网故障诊断方法具有许多优点，比如，精确的故障诊断结果,Petri网故障诊断模型内的每个元件或区域都简单灵活，并具有较高准确性。
关键词：Petri网，故障诊断，发电站
1.前言
大型发电站故障诊断能够作为识别发电站内元件或区域的一个过程，并通过保护继电器和断路器跳闸实现。因此，它需要来自监测控制和数据采集系统的数据。当该数据达到控制室时，运行人员对该数据进行分析并判断故障元件或区域。该诊断过程的准确性和快速性完全依赖与运行人员的经验。然而，随着发电站复杂程度的增加，尤其是在多重故障出现的情况下，大量报警信息会被传送到发电站控制室内。在这种情况下，运行人员需要迅速而准确的判断判断出现故障的部分。因此，研究人员开发出了用于控制室内的故障诊断系统，从而促进，支持并帮助运行人员完成他们在故障诊断过程中的任务。
目前研究工作都倾向于对故障诊断系统的开发。大多数研究都建立在专家系统（ES）的基础上[1–4]。人工智能的方法，比如，为人工神经网络[5–7]，遗传算法(GA) [8]，基于家族优生学进化理论[9]，免疫算法开发[10]。分别给出了基于模糊关系和贝叶斯网络两种相应的发电站故障诊断研作为参考[11, 12]。Petri网具有并行信息处理的特点，并发操作功能，并被认为是一种非常适合有效的建模工具。提出了一些基于Petri网的电力系统故障诊断的建模和分析的方法[13 – 16]。
故障诊断系统被广泛应用在电力系统和变电站。本文提出了一种基于Petri网的大型发电站简化故障诊断方法。该发电站包括：发电机组，升压变压器，厂用变压器，发电站母线和自耦变压器。所提出的故障诊断方法利用大型发电站内的保护继电器和断路器的信息来建立Petri网模型的每一个故障原件或区域。故障原件或区域能够被发生状Petri网的最终状态诊断和估计出来。此外，将所提出的Petri网的有效性与性能与所呈现出的模糊关系和贝叶斯网络进行了比较。
该方法在水电厂的发电公司的埃及子公司15.75/500 kV高坝发电站内进行测试，测试结果表明，所提出的方法推理较为简单，故障诊断模型具有较强实用性，同时，它能够帮助和支持发电站控制室运行人员做出正确的决定。
2.Petri网建模方法
2.1 Petri网的定义
Petri网是对离散分布式系统的一些数学和图形表示[17, 18]。作为一种模型化语言，它将分布式系统的结构图形化描述为定向二分图形决策支持系统。
2.2 Petri网建模能力
通过典型特征激活动态事件驱动系统所具有的典型特征，比如，并发，决策，同步和优先级，就能够有效利用Petri网进行建模[20]。
顺序执行:图2（a）中的变迁元素t2只能够在t1发生之后发生。这就强行限制程序必须为“t2在t1之后”。这种程序限制是动态系统中的一部分典型的执行方式。同时，这个Petri网结构模型的临时关系也在激活的范围当中。
冲突:变迁节点t1和t2在图2（b）中处于冲突中。无论是允许还是任何变迁节点都会导致其他变迁节点失去能力。于是会出现一种情况，比如，一个机器可以在部件类型之间选择或者一个部件可以在许多机器之间选择。由此产生的冲突可能会利用不确定性方法或者概率方法解决，通过适当分配冲突变迁节点概率进行。
并发性:在图2（c）中，变迁节点t1和t2是并发的。并发性是系统交互性的一个重要属性。对于一个变迁节点并发来说，分叉变迁节点的存在，也就是将一个令牌存放在两个或更多输出库所里是一个必要条件。
同时性:在本文中，一个动态的系统需要断路器和继电保护提供多重的相关信息。由此产生的相关信息的同步性会被变迁节点所捕捉，见图2d。当且仅当库所p1和p2都接受一个令牌是变迁节点t1才能执行。在Petri网络的其他地方，当令牌进入到每个库所中，可能对运行产生一个复杂的结果。实质上，变迁节点t1已经建模到了已经运行的系统了。
互斥性：由于在共享信息使用上的程序限制，如果它们不能同时运行，这两个程序将会相互排斥。图2（e）显示了这种结果。例如，一个命令同时受两件装载和卸载设备的影响。这样的两种结构是平行互斥和连续互斥的。 优先顺序：这种建模能力通过引入抑制弧来表现。抑制弧连接于输入库所和变迁节点，并且用在弧的终点画一个圈来表示。抑制弧连接于输入库所和变迁节点用以改变过渡条件，在抑制弧的存在下，如果每一个输入库索用正常的弧连接变迁节点，则我们认为这个变迁节点是被激活的（弧线的终点画上箭头来表示）。至少令牌的数量要等于弧的量，如果没有令牌，那每一个输入库所到变迁节点用抑制弧来连接。对于通常连接库所，变迁节点实施的规则也是相同的。然而，实施时不能改变抑制弧连接库所的标记。用抑制弧表示的Petri网见图图2（f）。如果p1包含令牌，则t1被激活，当p2包含令牌p1没有令牌，则t2被激活。则t1优先于t2。
(a)顺序(b)冲突
(c)相竞 (d)同步
(e)相互排斥(f)优先
图2 Petri网系统特征基本代表
