基于多源信息融合的配电开关柜状态评价方法研究

发布时间：2018-05-08 06:39

本文选题：配电开关柜 + 状态评价　；参考：《华北电力大学(北京)》2016年硕士论文

【摘要】：开关柜具有构造封闭、安装数量多、分布面积广等特点,是配电网中故障率最高的设备之一,其故障大概占到所有设备故障的20%左右,其检修费用已经超过了整个配电系统设备维修费用的一半。作为配电系统的主要设备之一,开关柜起着控制和保护其他电力设备的作用,其状态的好坏直接影响着电网的运行可靠性和用户的用电质量。因此,对配电开关柜进行合理准确的状态评价可以为开关柜的性能评估、合理使用和安全工作提供依据,进而为维护整个配电系统的安全稳定运行做出贡献。本文针对现有开关柜状态评价方法中的不足,结合配电开关柜的结构和安装特点,从多方面选取表征开关柜状态的特征量,根据重要性的不同,对特征量进行分类,建立分层空间模型；接着以分层空间模型作为信息融合的结构基础,采用物元可拓学方法对各层各特征量信息进行融合,实现对开关柜当前状态的评价,并对传统物元可拓学方法中权重分配和关联度计算两方面进行改进,优化了实际评价效果；最后,基于开关柜的历史及当前评价结果,采用动态贝叶斯方法进行概率推理,实现开关柜未来状态的预测,为状态检修工作提供了良好的基础。文中通过实际算例对评价过程进行了说明,将当前评价结果与开关柜的实际状态以及采用传统物元可拓法得到的评价结果作对比,证明了所提方法的有效性；对同一开关柜的历史及当前状态评价结果进行贝叶斯概率学习,实现了对其未来状态的预测,并将预测结果与开关柜的实际情况作比较,证明了预测的准确性。
[Abstract]:The switchgear has the characteristics of closed structure, large number of installation, wide distribution area and so on. It is one of the equipment with the highest failure rate in the distribution network, and its faults account for about 20% of all the equipment failures. The maintenance cost has exceeded half of the total maintenance cost of the distribution system equipment. As one of the main equipment of distribution system, switchgear plays a role in controlling and protecting other power equipment, and its state directly affects the reliability of power grid and the quality of power consumption. Therefore, the reasonable and accurate state evaluation of the switchgear can provide the basis for the performance evaluation, rational use and safety work of the switchgear, and then contribute to the maintenance of the safe and stable operation of the entire distribution system. In this paper, aiming at the deficiency of the existing state evaluation methods of switchgear, combined with the structure and installation characteristics of the switchgear, the characteristic quantity representing the state of the switchgear is selected from many aspects, and the characteristic quantity is classified according to the difference of importance. The layered spatial model is established, and then the layered spatial model is used as the structural basis of information fusion, and the information of each layer characteristic quantity is fused by using matter-element extension method to realize the evaluation of the current state of the switchgear. The traditional matter-element extension method is improved in weight distribution and correlation degree calculation to optimize the actual evaluation effect. Finally, based on the history and current evaluation results of switchgear, dynamic Bayesian method is used to carry out probability reasoning. It can predict the future state of switchgear and provide a good basis for condition maintenance. In this paper, the evaluation process is explained by practical examples. The effectiveness of the proposed method is proved by comparing the current evaluation results with the actual state of the switchgear and the evaluation results obtained by the traditional matter-element extension method. The history and current state evaluation results of the same switchgear are studied by Bayesian probability learning. The prediction of the future state of the same switchgear is realized, and the accuracy of the prediction is proved by comparing the prediction results with the actual situation of the switchgear.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TM591

【相似文献】