数据挖掘技术在电力系统故障诊断中的应用

发布时间：2018-04-10 00:14

本文选题：数据挖据　切入点：决策树　出处：《华北电力大学》2017年硕士论文

【摘要】：随着信息技术的日新月异,计算机技术的应用遍布各行各业,海量的信息融入人们的方方面面,现今人们正处在一个信息爆炸的年代。如何有效的处理海量的数据信息,从中找到有价值的数据,就成了各方研究的焦点。在电力行业,电力设备越来越多,电网结构越来越复杂的情况下,如何把采集到的电网数据通过提取和分析,得到有价值的运行诊断信息,数据挖据技术的出现给这种用于电力系统故障诊断中海量信息的处理分析提供了方法。本文中,我们把数据挖据技术应用到了电力系统的故障诊断中,构建了电力系统故障诊断的模型。此模型主要利用决策树把知识的获取和表示结合在一起,使两者的处理能够同步进行。在电力系统故障诊断的模型中,我们改进了数据挖据技术中的决策树算法,并用它完成知识的获取与之结合。本文系统由推理机、解释器、决策树算法模块和图形界面四大模块组成。推理机的主要功能是以决策树算法生成的规则进行推理,解释器主要对推理机所返回的规则和决策树生成或者专家录入的规则进行翻译,决策树算法主要作用是构架一个高效易用的决策树并且生成分类准确的规则供推理机使用,图形界面负责直观的显示输出结果。本实验系统利用.NET平台构架,系统还可以扩展适用的数据样本或者使用数据挖据中其它的算法来处理其它方面的问题。
[Abstract]:With the rapid development of information technology, the application of computer technology is spread all over the industry, massive information is integrated into all aspects of people, now people are in an era of information explosion.How to effectively deal with massive data information and find valuable data has become the focus of research.In the electric power industry, with more and more power equipment and more and more complex power network structure, how to get valuable operation diagnosis information by extracting and analyzing the collected power grid data,The emergence of data mining technology provides a method for processing and analyzing mass information in power system fault diagnosis.In this paper, we apply the data mining technology to the fault diagnosis of power system, and construct the model of power system fault diagnosis.This model mainly uses decision tree to combine knowledge acquisition and representation so that the two processes can be processed synchronously.In the power system fault diagnosis model, we improve the decision tree algorithm in data mining technology, and use it to complete the knowledge acquisition and combination.The system consists of four modules: inference machine, interpreter, decision tree algorithm module and graphical interface.The main function of the inference machine is to infer the rules generated by the decision tree algorithm. The interpreter mainly translates the rules returned by the inference machine and the rules generated by the decision tree or input by experts.The main function of the decision tree algorithm is to construct an efficient and easy to use decision tree and to generate accurate classification rules for use by the inference machine. The graphical interface is responsible for displaying the output results intuitively.This experiment system uses .NET platform framework, the system can also extend the applicable data samples or use other algorithms in data mining to deal with other problems.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13;TM711

【参考文献】