基于模糊推理脉冲神经膜系统的电网故障诊断方法

发布时间：2018-11-04 10:08

【摘要】：随着电网规模不断扩大和互联程度不断提高,电网发生故障时,若故障不能得到及时有效处理,会演变成大面积停电事故,进而对社会造成重大经济损失。电网故障时,要求调度人员快速准确地完成故障诊断,从而将故障区域隔离并为后续供电恢复工作提供基础。而大量故障多源信息在短时间内涌入调度中心,在提供诊断依据的同时,也给调度人员进行快速、准确的故障诊断增加了难度,会导致调度人员出现误判和漏判的情况,从而使故障情况加剧,造成更严重的停电事故。因此,电网故障诊断方法的研究对及时有效诊断故障,保证电网安全、稳定运行具有重要实际意义。模糊推理脉冲神经膜系统是将模糊理论引入膜计算理论而产生的一种可以处理不确定性因素的计算模型,具有直观的图形表示、并行性、动态性以及不确定性等特点。而电网故障是一个离散动态演变过程,它由故障、保护动作、断路器跳闸等一系列事件组成,且受到不确定性因素影响。模糊推理脉冲神经膜系统的特点使其适合求解电网故障诊断问题,近年来被应用于电网故障诊断领域,得到一定程度的发展。所以,本文在现有基于模糊推理脉冲神经膜系统的电网故障诊断方法的基础上继续深入,研究一种以调度中心获得的故障多源信息为诊断依据的基于模糊推理脉冲神经膜系统的电网故障诊断方法,使其能够利用警报时序信息在警报信息存在不确定性因素的情况下识别出故障元件并且能利用故障电气量信息判断出故障类型,从而给调度人员提供辅助决策。论文首先利用警报时序信息,给出了基于时序模糊推理脉冲神经膜系统的电网故障元件识别方法。该方法首先建立了基于时序模糊推理脉冲神经膜系统的故障元件识别模型,其次利用时序一致性约束检查警报信息有效性,从而修正了保护和断路器动作信息的初始置信度,进而提高了识别结果准确性。最后应用IEEE 39节点电网模型的算例测试了方法的有效性,并与其它方法的结果进行了对比。利用输电线路故障电气量信息,给出了一种将小波变换和模糊推理脉冲神经膜系统相结合的输电线路故障分类方法。该方法首先利用小波变换和奇异值分解提取故障信号特征,然后利用基于模糊推理脉冲神经膜系统的故障分类模型对故障特征进行分类,从而确定故障类型。以500kV输电线路故障模型为例,做了大量仿真分析。仿真结果表明,该方法不受故障初始角、故障位置和过渡电阻等因素的影响且对线路参数以及噪声有较好的适应性。
[Abstract]:With the continuous expansion of power network scale and the continuous improvement of interconnection, if failure occurs, if the fault can not be dealt with in time and effectively, it will evolve into a large-scale blackout accident, which will cause great economic losses to the society. In order to isolate the fault area and provide the basis for the subsequent work of power supply recovery, the dispatcher is required to complete the fault diagnosis quickly and accurately. However, a large number of fault multi-source information poured into the dispatching center in a short time, which not only provided the diagnostic basis, but also increased the difficulty of rapid and accurate fault diagnosis for the dispatcher, which would lead to the misjudgment and omission of the dispatcher. As a result, the fault situation is aggravated, resulting in a more serious blackout. Therefore, the research of fault diagnosis method has important practical significance for timely and effective fault diagnosis, ensuring power grid safety and stable operation. Fuzzy reasoning impulsive neural membrane system is a kind of computing model which can deal with uncertain factors by introducing fuzzy theory into membrane computing theory. It has the characteristics of intuitive graphical representation, parallelism, dynamic and uncertainty. Power network fault is a discrete dynamic evolution process, which is composed of a series of events, such as fault, protection action, circuit breaker tripping and so on, and is affected by uncertain factors. The characteristics of fuzzy inference pulse neural membrane system make it suitable to solve the power network fault diagnosis problem. In recent years, it has been applied in the field of power network fault diagnosis, and has been developed to a certain extent. Therefore, based on the existing fault diagnosis methods of power network based on fuzzy reasoning impulse neural membrane system, this paper continues to go deep. A fault diagnosis method based on fuzzy inference pulse neural membrane system based on multi-source fault information obtained from dispatching center is studied. It can use the alarm sequence information to identify the fault elements and to use the fault electrical information to judge the fault type, which can provide auxiliary decision for the dispatcher. In this paper, a method of fault element identification based on time-series fuzzy inference pulse neural membrane system is presented based on alarm timing information. In this method, a fault element identification model based on time-series fuzzy reasoning pulse neural membrane system is first established, and then the validity of alarm information is checked by time sequence consistency constraint, thus the initial confidence degree of protection and circuit breaker action information is corrected. Thus, the accuracy of the recognition results is improved. Finally, the effectiveness of the proposed method is tested by an example of IEEE 39 bus power network model, and the results are compared with those of other methods. A fault classification method based on wavelet transform and fuzzy inference pulse neural membrane system is presented based on the electrical information of transmission line fault. Firstly, wavelet transform and singular value decomposition are used to extract fault signal features, and then fault classification model based on fuzzy inference pulse neural membrane system is used to classify fault features to determine fault types. Taking 500kV transmission line fault model as an example, a lot of simulation analysis is done. The simulation results show that the method is independent of the fault initial angle, fault location and transition resistance, and has good adaptability to line parameters and noise.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM711

【相似文献】