Comprehensive Condition-Information-Based Risk Assessment Method for Power Transmission Line

The traditional risk assessment methods do not adequately consider the specific operating environment and characteristics of a power transmission line. Therefore, a new risk assessment method considering the information of comprehensive conditions is proposed in this paper. First, a risk assessment model is designed via statistical methods based on the information on preceding failures and defects of a power transmission line. Then, the quantization processes of the state parameters of components in the transmission line are formed to get the risk assessment results considering historical failure information. These results are of failure probability meaning. Finally, the influences of the operating age of equipment in the transmission line, the position of the transmission line, the service time, and the working condition of the grid on the risk value of the transmission line are analyzed. An empirical example shows that the suggested method can predict the risks of transmission lines and offers a way for maintenance scheduling and equipment replacement.


Introduction
With the rising scale and significance of power systems, higher requirements are needed for the reliable operation of power systems.Power experts need to tackle the ongoing problems and know the possibility of equipment problems and the severity of problems through risk assessment.In addition, the risk assessment data are often used for the maintenance and scheduling of the power grid and provide references for operation, monitoring, maintenance management, fault handling, and on-site inspection of power enterprises [1][2].As a key part of the power system, it is necessary to take risk assessment on the transmission line.
Few types of research at home and abroad focus on transmission line risk assessment.The major methods adopted here are the Monte Carlo method, fault tree analysis, risk matrix analysis, etc.For qualitative analysis, risk matrix analysis regards the possibility of a risk happening and the severity of that risk as x and y coordinates on the same plane [3][4].The Monte Carlo method analyses the likelihood distribution of risk occurrence probability, risk factor, and loss resulting from risk and other variables [5].Fault tree analysis is adopted to study the logical relationship among faults, trigger events, direct causes, and indirect causes [6].The operation status of the grid is required to be considered in the risk assessment [7].For this, there are few relevant types of research at home and abroad.For different faults, the severity of the consequences is similar [8].Therefore, the discussion on risk consequence severity is omitted in this paper, and the failure probability directly corresponds to the risk of transmission lines, ensuring the science of risk assessment and making it simple.
According to the characteristics and operating conditions of transmission lines, this paper suggests a risk assessment method based on all working conditions information of transmission lines.Firstly, the line will be classified into several components.According to different components, the state parameters reflecting their operating status are exposed.The preceding and current information of these state parameters is analyzed statistically and eventually to obtain the comprehensive risk of line.Then, the risks are revised based on the operating age of components, the operating time segment information of lines, and the status of the power grid.Lastly, a specific case study is performed to verify the practicability and effectiveness of the advocated method.

Overall structure of risk assessment
Based on transmission line composition and risk assessment requirements, transmission lines consist of eight parts: foundation, tower, conductor, ground wire, insulator, fittings, auxiliary facilities, grounding device, and channel environment.For different components, several state parameters can reflect their risk status, like the inclination of the tower, the sag of the conductor, the self-detonation of the insulator, etc. Figure 1 shows the risk assessment model presented in this paper.
Figure 1 The risk assessment model for components of the power transmission line The data in the component risk assessment model has many sources, i.e., fault and defect statistics of components over the years and the operating age of components.The state parameters in the model are given by the online monitoring parameters and the manual inspection parameters that reflect component status.The calculation of component risk value needs two parameters: the quantification of state parameters information, and the degree of correlation between state parameters and component risk, namely the state parameter's level of component risk membership.

Quantification of risk level for state parameters
About the state parameters that can be quantified through observation, a statistics-based risk degree quantification method for state parameters is proposed.
First, the data and corresponding times of each state parameter in past years were counted, and the data distribution of the state parameter in Figure 2 was obtained by fitting.The monitored value ω of a state parameter is marked by the horizontal axis, and the happening times of the monitored value ψ are marked by the vertical axis.If a% of the statistical data is in normal status, the boundary value of a% of the overall data is defined as the normal limit value, denoted by μ.Among the statistical data of various state parameters of transmission lines, at least 90%−95% of the data belongs to normal status.To avoid missing judgment, a%=90% is taken.When the state parameter exceeds the normal limit, and the component is close to 100% failure, the value of the state parameter at this time is the failure limit, denoted by ξ, and 100% turns out to be the corresponding risk degree of the state parameter.
The fault limit is established according to the power regulatory standard's warning or attention value.Specifically, the fault limit is three times the attention value or 1.3 times the warning value.This dynamic multiple could be adjusted frequently according to the practice's feedback.
We suppose the parameter observation is x.Then the corresponding risk level is calculated by: 0 ( ) ( ) In Formula (1), k>1 is the trend index.It noted that the relationship between the value of the state parameter and the risk level is not linear.As the value of the parameter is just over the normal limit, the risk level increases slowly, and its growth rate increases with the increase of the state parameter.

Membership level and overall risk value of state parameter
The previous section demonstrates the risk level of each parameter under different conditions.However, it is not the risk value by the component.Therefore, based on the known risk degree, estimating the influence of each parameter is necessary, namely membership level, represented by α.
Supposing that the related state parameter occurring in the records of the faults and drawbacks of a transmission line's component in recent (n) years is denoted by Q S1 , Q S2 ,…, Q Sn , and the occurrence times of state parameters are t 1 , t 2 , …, t n , respectively, then the membership level α i of the component state parameter Q Si is determined by Formula (2).
Therefore, Formula (3) shows the single state parameter risk value r i of the i th state parameter Q Si of a certain component.The risk level of the component is denoted as r i and calculated by: Suppose a single state parameter risk value of all state parameters be r 1 ､r 2 ､…､r n , then Formula (4) demonstrates the risk value R of the component.
The risk value of each parameter can add up to maintain the component risk because the single state parameter risk value is considered to contribute to the overall component risk in the previous calculations instead of the likelihood leading to component failure.
As can be seen from Formula (5), the component risk value gives the component risk the meaning of fault likelihood, whose range is [0, 1].When the state parameter is placed in the worst situation, the component risk value will be 1.
We assume that the risk of eight transmission line components is R 1 , R 2 , … R 8 , then we adopt the probability addition method.It is possible to get the overall risk as indicated in Equation ( 6).

Age coefficient of component
During operation, power equipment will have various corrosion and abrasion to varying degrees, like metal structural and characteristic changes, mechanical abrasion, insulation damage, and other deterioration phenomena.Table 1 shows the links among component aging index, remaining life, and failure likelihood.The equipment aging index is the same for all types of equipment and could be calculated as follows: Here, t AG is the aging index to be achieved; 0 AG is the primary aging index, taking 0.5 generally; B is aging constant; T is the year when the aging index needs to be maintained; 0 T is the year when the equipment was put into operation; mod f is correction coefficient.
Generally, according to the expected life n given by the equipment manufacturer, the aging index will change from the initial 0.4 to the final 5.6 at the end of the outage under a normal operating environment ( mod 1 f  ).Thus, aging constant equipment is designed.ln(5.6 / 0.4) ln14 2.64 The correction factor mod f is determined by defect, load, environment, and other factors.The equipment's aging index in the final year of expected life is 10, and the associated correction coefficient is 10/5.5≈1.82under the worst possible conditions.Therefore, the value range of mod f is [1, 1.82] depending on the operating environment.
After the operation, the component's age coefficient in the t year can be obtained from Formula (9), and the maximum age coefficient is 1.5, which can be adjusted according to the feedback information of the operation.0.05 1 (9) For the foundation, tower, and grounding device, their age coefficient is defined according to the main components.All the equipment in the line is taken as their age coefficient for conductors, ground wires, insulators, fittings, and ancillary facilities.It is not necessary to consider the channel environment's age coefficient.

Transmission line operating environment coefficient
Based on features of transmission lines and operation experience, six environmental elements that have the greatest impact on transmission line risk, namely lightning stroke, mountain fire, typhoon, ice cover, external damage, and bird damage, are selected.
In the specific calculation, the impact of the operating environment on risk can be divided into p (geographical location coefficient) and t (time coefficient).For t, taking the southern operating environment as an example, the special time zones are determined according to the statistics of faults and defects over the years, as listed in Table 2.
Table 2 Particular time zones

Lightning stroke March to September
Mountain fire September to the following January Icing September to the following February Typhoon June to October Bird damage September to November For example, based on the statistics of icing fault-prone time over these years, the occurrence probability of icing failure, that is, the time coefficient of icing failure, presents a normal distribution N( 1  ) with time, as revealed in Equation (10). ) where x is evaluation time (taking month as a unit); 1  is the variance of t, and 1 K is the offset value of the normal distribution.
For the other five special operating environments, the distribution functions with time or geographical location as variables can be obtained according to the analysis method of icing.At the vertex of the distribution function, we take i t or i p =1.2, within the range, its specific value is obtained by the distribution curve; outside the range, we take i t or i p =1.
As for the operating environment's impact on risk, the transmission line should be located in a special time and geographical location segment at the same time.Therefore, the operating environment coefficient i tp can be obtained from Equation (11), where i=1~6 refers to lightning stroke, typhoon, mountain fire, bird damage, icing, and external damage, respectively.
, 1 Equation ( 12) represents the revised transmission line risk value R  .

Case study
For example, the 2013 online monitoring data and inspection records of a 500 kV line are used to verify the risk assessment method's consequence based on all condition information of transmission lines.
According to the obtained data, Table 3 shows the defects of this line in inspection and the associated state parameter risk degree of defects.

Conductors, ground wire
The clamp of 195# right overhead ground wire is offset 10 cm to the small number side 0.5 230# optical fiber drainage had 4 strands of breakages 0.8

Insulator
The composite insulator jumper string umbrella skirt of 159# small number side left phase had 4 damages 0.5 The first glass insulator in the front series of 207# right phase had self-detonation 0.2 Fittings 264# Tension insulator string connection fittings had corrosion 0.2

Channel environment
Three new plastic greenhouses are under the wires between 212# and 213# 0.5 10 cypress trees are below 175# large number side wire, the vertical distance from the wire is 5 meters.0.5 Take the ground wire and conductor as an example, the membership level of the state parameter "slippage of the conductor in line clamp" is 0.034.The membership degree of the parameter "OPGW cable broken strand" is 0.036.Hence, the component risk value of the conductor and ground wire is 0.5×0.034+0.8×0.036≈0.046before correction.Other components' risk values before correction can be calculated similarly.
The corresponding component age coefficient and risk are calculated and reported in Table 4.The components' risk value is 0 if no defects have been recorded.According to Table 4 and Formula (6), this transmission line's risk value is R =0.157.Since there is no abnormality during this period, its status coefficient is S=1.In other words, the power grid's operating status does not influence the transmission line's risk value, and the line's obtained risk value is 0.195, revealing that the line is at low risk.Therefore, in this section, maintenance of the line should be enhanced, and the current defects should be corrected as soon as possible.

Conclusion
In this paper, combining the features of transmission lines, the risk assessment method is studied and can fully reveal the transmission lines' risk situation.Yet, as the risk assessment of the lines is under development, such methods used in the assessment process and some specific parameter settings should be adjusted frequently based on the system feedback.Therefore, the proven results can more precisely indicate the risk of transmission lines.
In the meantime, after obtaining the risks of the lines, it is necessary to explore how to adapt the assessment results to the subsequent maintenance schedule, equipment upgrade, and other works.Thus, the auxiliary decision for transmission lines based on risk assessment requires further research.

Figure 2
Figure 2 Statistical data fitting of a state parameter

Table 1
Relationship between the aging index and failure likelihood

Table 3
Transmission line's defect data and its associated risk level and membership level

Table 4
Components' age coefficient and their risk before and after revision