摘 要

本文总结了台风灾害下电力系统风险评估的研究现状和前人的研究成果，首次使用了逻辑斯蒂模型对台风灾害下的输电线路进行风险评估。以线路的停运事件作为二分变量，以停运率为风险指标，以台风信息、线路信息、沿线树木信息、土质类型信息作为自变量，建立了输电线路的逻辑斯蒂模型。使用了模拟数据作为建模的基础，并对模型进行了检验，表明该模型具有较好的预测效果，但是受限于数据的数量，模型目前并不完美。在此模型的基础上，提出了基于电力公司损失计算的严重度函数，并且分为了粗略计算和精确计算两种算法。对两种算法进行了计算比较，结果表明当负荷的分时差异较明显时，粗略计算误差较明显。

本文基于经验拟定了140条数据，其中70条用来表示事件发生，另外70条用来表示事件不发生。用Y=1表示发生停运事件，用Y=0表示不发生停运事件，并用模拟数据建立了台风灾害下考虑多种因素的输电线路停运可能性模型。以概率P=0.5区分停运事件发生和不发生，并对停运事件进行预测。本文提出的严重度函数是基于电力公司的经济损失，综合了有功负荷的利润损失和输变电设备损毁的损失，其中前者是随时间变化的，而后者是固定损失。本文利用逻辑斯蒂模型填补了该领域的空白，并且今后会在此基础上改进此项研究，可能会利用图像处理技术使预测结果更加直观。

关键词：台风；电力系统；停运率；逻辑斯蒂回归；严重度函数

Abstract

This paper summarizes the research status of power system risk assessment under typhoon disaster and the previous research results，and it was the first use of the Logistic model for the risk assessment of transmission line under typhoon disaster. Taking the outage event of the line as the dichotomous variable, and taking the outage rate of the line as the risk index, the typhoon information, the line information, the tree information along the line and the soil type information as the independent variables, the logical model of the transmission line was established. In this paper, artificial data was used as the basis of modeling, and the model was tested. It showed that the model had good prediction effect, but limited to the number of data, the model was not perfect at present. On the basis of this model, the severity function based on the loss calculation of power companies was proposed, which was divided into two algorithms: rough calculation and accurate calculation. The two algorithms were calculated and compared. The results showed that when the difference of load in different duration is large, the error of the rough calculation is obvious.

This article creates 140 data based on experience, 70 of which are used to simulate the occurrence of an event, and the other 70 to simulate the situation where the event does not occur. Y=1 indicates that an outage event has occurred, and Y=0 indicates that no outage event has occurred. And the analog data was used to establish a model of the possibility of transmission line outage considering a variety of factors under typhoon disaster. Using probability P=0.5 to distinguish between the occurrence of outage events and outage events that does not occur, and the outage events was predicted. The severity function presented in this paper is based on the economic loss of the utility company, which combines the loss of profits of the active load and the loss of power transmission equipment. The former is time-varying, while the latter is a fixed loss. This paper uses the Logistic model to fill the gaps in the field and will improve the study on the basis of this, and may use the image processing technology to make the prediction more intuitive

Key words: typhoon; power system; outage rate; logistic regression; severity function

目录

第一章 绪论 1

1.1本研究的目的及意义 1

1.2国内外研究现状 2

1.2.1台风模型的建立 2

1.2.2关于气候变化和用台风模型进行风险评估 3

1.2.3台风灾害下电力系统风险评估方法及比较 5

1.3台风造成电力系统设备损毁的原因 6

1.4本文研究内容及章节安排 7

第二章 逻辑斯蒂模型介绍 8

2.1逻辑斯蒂模型的原理 8

2.2逻辑斯蒂回归分析的一般步骤 10

2.2.1自变量间的多重共线性检验 10

2.2.2模型系数评估 10

2.2.3系数的显著性检验 11

2.2.4模型评估 11

2.3 SPSS软件介绍 12

2.4 本章小结 12

第三章 线路停运可能性建模 13

3.1变量选择和事件模拟 13

3.1.1选择自变量和因变量 13

3.1.2经验下确定性事件的逻辑模拟 15

3.2建模流程 16

3.2.1进行自变量共线性检验 16

3.2.2建立逻辑斯蒂模型 16

3.2.3系数的显著性水平检验 17

3.2.4模型的评估 17

3.3断电预测及建议 17

3.4严重度函数 18

3.4.1用平均值计算损失 18

3.4.2用日负荷曲线计算损失 18

3.5本章小结 18

第四章 算例分析 20

4.1负荷为市政用电负荷 22

4.2负荷为钢铁工业用电负荷 22

4.3本章小结 23

第五章 结论与展望 24

5.1研究结论 24

5.2研究展望 24

参考文献 26

致谢 28

附录A 毕业设计期间发表的主要论文 29

附录B 自变量共线性检验结果 30

附录C 建模所用原始数据及预测结果 33

第一章 绪论

1.1本研究的目的及意义

目前，全球进入自然灾害多发期，日益严重的厄尔尼诺现象导致恶劣气候、台风灾害等连锁反应不断发生，给社会、经济和人民生活造成了严重的影响和损失。极端气象条件引发的电力系统停运事故，给人民生产、生活及电力系统正常运行带来了严重危害。台风对电力系统的危害很大，不仅会造成断线、倒杆、倒塔、跳闸等故障，给人民生活、工业生产和国民经济造成巨大的损失，还直接间接地造成人员伤亡。

台风对电力系统造成的损害可以从几个案例说明^[1-5]，见表1.1。因此，为了研究台风对电力系统输变电设备的影响，本文提出了利用逻辑斯蒂模型对台风各项信息和电力系统各项信息进行回归分析，找出台风灾害下造成断电的因子，并拟合出因子与断电灾害发生概率之间的关系表达式，在此基础上提出严重度函数。主要的目的是验证逻辑斯蒂模型在研究台风对区域输变电设备损毁可能性时的有效性，并针对预测结果给出合理的建议。