Reliability Calibration of Tower Members in Transmission Line System Crossing High-Speed Railway

Due to the rapid development of high-speed railway in China, transmission line systems across the high-speed railway system have become more prevalent in recent years, which highlights the importance of reliable design for systems. e design is generally based on the structural reliability theory, in which the determination of target reliability level is the key. To determine the reliability level of a transmission line across high-speed railway, the statistical parameters of load eects and resistances of towermembers are derived and reliability calibration of tower elements satisfying the minimum design requirement in Chinese codes is performed by JCmethod. Furthermore, the reliability level of a transmission line across high-speed railway is divided into three classes, in which Class 1 is the strongest. According to the calibration results, the minimum target reliability indices are recommended. e results show that the reliability level is similar to the reliability of tower elements in the U.S. but higher than that in Canada. e target reliability indices with values of 3.7, 3.2, and 2.7 are recommended for Class 1, Class 2, and Class 3, respectively.


Introduction
With the rapid development of the national economy, the electrical power system has attained prominent achievements in China. e transmission line system, as the electric power carrier, is a lifeline project, which is becoming more and more complex. erefore, it is necessary to present new requirements for the design philosophy. Meanwhile, the phenomenon of transmission line system crossing the highspeed railway system is increasing substantially with the rapid development of high-speed railway. To ensure the safe operation of high-speed railway, the higher performance requirements for transmission line crossing high-speed railway must be satis ed. Since transmission line systems and rail facilities in China were designed by using di erent design approaches with various material strengths, partial factors, or safety factors, the safety of transmission system and railway facility cannot be clearly identi ed only based on the design speci cations or codes. For this reason, the identical criterion should be employed to determine the safety of high-speed railway and transmission line system.
During the past 40 years, rapid and signi cant development has arisen in the eld of structural safety. e primary theme in structural safety is reliability analysis, which can be de ned as the consistent evaluation of structure safety using probability theory [1]. e uncertainties associated with load e ects, material properties, physical dimension, and calculation model are fully taken into account in reliability analysis. Reliability analysis in the industries of buildings, bridges have been extensively carried out, and target reliability indices have been calibrated in the aforementioned industries. However, very little research e ort has been conducted on reliability of transmission line system [2][3][4][5].
As is known to all, reliability-based design method has been introduced to design standards or codes of electric system in many countries and applied to the design of overhead transmission line, such as "National Electric Safety Code (NESC C2-2002)" [6], "Guidelines for Electrical Transmission Line Structural Loading (ASCE 74-2009)" [7], "Canadian Electrical Code: Overhead Systems (CSA C22. 3 No.1-2001)" [8], "Overhead Electrical Lines Exceeding AC 45 kV (EN 50341-1)" [9], and "Design Criteria of Overhead Transmission Lines (IEC 60826-2003)" [10]. e probability-based design method has also been applied to Chinese design codes of overhead transmission line, such as "Technical Regulation of Design for Tower and Pole Structures for Overhead Transmission Line (DL 5154-2012)" [11], "Code for Design of 110 kV∼750 kV Overhead Transmission Line (GB 50545-2010)" [12], and "Code for Designing of ±800 kV DC Overhead Transmission Line (GB 50790-2013)" [13]. In the field of rail engineering, reliability-based limit state design method is specified in Chinese standard "Unified Standard for Reliability Design of Railway Structures (GB 50216-2019)" [14].
For ultra-high voltage (UHV) transmission line, in order to satisfy the safety requirement of transmission line across high-speed railway, target reliability indices of the components of transmission line satisfying the minimum requirements should be calibrated. e research is limited to calibration of reliability index and definition of reliability level of transmission tower for transmission line crossing high-speed railway.

Statistical Parameters of Load Effects
and Resistance e transmission line is composed of power transmission towers, insulators, fittings, conductors, and Earth wires. To serve its purpose, transmission line system must be safe against various loads, such as the weight of structure itself, wind load, and ice load. In the reliability analysis and design of overhead line structures, probabilistic distributions and statistical parameters of structural load effects and resistance capacity must be determined first.

Statistical Parameters of Load Effect.
Load effects are the moments, shears, and axial forces resulting from the loads on the structure. Statistical distributions of load effects are consistent with that of the loads. Statistical parameters of the loads are defined in terms of the bias factor (the ratio of mean value to nominal value) and coefficient of variation (the ratio of standard deviation to the mean value, abbreviated as COV). It should be pointed out that loads are classified by their characteristics varying with time, that is, permanent load and variable loads such as wind load or ice load.

Permanent Load.
e major part of permanent load is the weight of towers, insulators, and fittings. Based on the existing researches, permanent load follows a normal distribution [15]. Based on the results of statistical analysis of building structure, the bias factor of permanent load k G � 1.06 and COV δ G � 0.07 are suggested in the article [15].

Wind Load.
Wind load is one of the major variable loads acting on overhead transmission line. e action mechanism of wind load on overhead transmission line is very complicated. e same statistical parameters and probability distribution as that used in building structures are employed in this work. Wind load normally follows a Gumbel distribution. Two load cases will be considered as described below.
(1) e Extreme Wind Load Case. 30 m/s was taken as reference wind speed in the extreme wind load case and the speed can be treated as standardized wind speed which is used to define the standard value of wind pressure W k in Chinese code. e bias factor and COV of the extreme wind load are taken as k WT � 0.908 and δ WT � 0.193, referencing the statistical parameters of building structure designed using Chinese codes [16].
(2) e Combined Wind and Ice Load Case. 10 m/s is taken as the reference wind speed in the combined wind and ice load case. It is assumed that the wind speed of 10 m/s is the maximum wind speed for the condition of annual maximum ice thickness, so the average wind load effect can be estimated by the following equation: where μ wt is the mean of wind load for the combined wind and ice load case; μ WT and σ WT are the mean and standard deviation of wind load in the extreme wind load case, respectively; α WT is the scale parameter of Gumbel distribution, taken as 1.2826; t is 1 month; and T is the return period of wind load, taken as 50 years. Hence, the bias factor of wind load equals 0.034 for the combined wind and ice load case. e COV of wind load is still assumed to be taken as 0.193.

Ice Load.
When drops of water in the air and wet snow are in contact with components of overhead transmission line, such as conductors, insulators, and fittings, the coagulated ice on power transmission conductors may occur. ere are many natural factors affecting icing, such as weather, terrain, altitude, wind speed and direction, and cable conductor.
Given that the icing thickness is not even, substantial uncertainties cannot be neglected in this problem. Ideally, the statistics of ice thickness used for reliability analysis should be obtained from the local weather stations. However, due to the lack of statistical data available related to the coagulated ice, the statistical results of observation station in Enshi of Hubei Province are used in the article [17]. Based on the previous statistical data [17] and engineering experience, ice load is assumed to follow a lognormal distribution with the bias factor k I � 1.1 and the COV δ I � 0.3.

Statistical Parameters of Resistance.
Resistance is the ability of components and structures to resist load effects. ere is indeed a high degree of uncertainty associated with structural resistance, mainly from three aspects: the uncertainties in material properties, the uncertainties associated with geometric dimensioning, and calculation model.

Strength Calculation of Axially Loaded Members.
According to Chinese code "DL/T 5154-2012" [11], the resistance for strength calculation of axially loaded tower members can be expressed as where Ω P is the model error of Equation (2), m is the strength reduction factor; A is the cross-sectional area of tower members; and f is the material strength of tower members. e bias factor and COV can be, respectively, described by the following relationships: in which k P , k A , and k f as well as δ P , δ A , and δ f denote, respectively, the bias factor and COV of the calculation model, cross-sectional area, and material strength of tower members, as shown in Table 1 [18]; f k , A k , and R k are the characteristic value of steel strength, cross-sectional area of tower members, and resistance, respectively. Substitution of the parameters in Table 1 into (3) and (4) can estimate the statistical parameters of resistance for strength calculation of axially loaded members of tower, namely, k R � 1.134 and δ R � 0.117.

Stability Calculation of Axial Compression Members.
Similarly, based on DL/T 5154-2012, the resistance for stability calculation of axial compression members of tower can be written as where ϕ is the stability coefficient of axial compression tower member and m N is the stability reduction factor of compression chord member. erefore, the bias factor and COV of resistance can be, respectively, computed by where k ϕ and δ ϕ are the bias factor and COV of ϕ. Statistical parameters of Ω P in (6) and (7) are listed in Table 2 [18]. Statistical parameters of ϕ shown in Table 2 are derived by the following method. According to Chinese code DL/T5154-2012 [11], ϕ can be calculated as in which λ is defined as where λ is slenderness ratio of tower member; f y is steel yield strength; E is elastic modulus of steel; and α 1 , α 2 , and α 3 are factors specified in Chinese code DL/T 5154-2012 [11]. Since the stability coefficient ϕ shown in (8) is a piecewise function, the definition of probability distribution function and statistical parameters of ϕ are complicated. In order to determine the statistical parameters of the stability coefficient ϕ, steel yield strength f y is considered as random variable following a normal distribution with a bias factor of 1.08 and a COV of 0.08 [18]. Generally speaking, the variability of elastic modulus E is very small and a constant value of 2 × 10 5 MPa is adopted in the article. en, 100,000 random numbers for f y is achieved through Monte Carlo simulation (MCS), which are substituted into (8), thereby obtaining the random values of ϕ. According to the aforementioned statistical analysis, the statistical parameters of ϕ for λ of 10, 50, and 100 are given in Table 3, in which Section a and Section b are classifications of a cross-section of members in Chinese code DL/T 5154-2012 [11]. e readers are referred to [11] for detailed information about the classification of section and steel classes (i.e. Q235 and Q345). For illustrative purposes, slenderness ratio λ � 50 of Section b is taken as an example, and the frequency histogram of ϕ/ϕ k is depicted in Figure 1, where ϕ k is the characteristic value of ϕ. e bias factor of ϕ is equal to 1.025, COV of ϕ is equal to 0.068, as listed in Table 3.

Advances in Civil Engineering
Consequently, substituting statistical parameters of all random variables shown in (6) and (7) can obtain that the bias factor and COV of resistance for stability calculation of axial compression members are 1.185 and 0.150, respectively.

Summary of Statistical Parameters for Load Effects and
Resistance. A summary of the statistical parameters is given in Table 4 including bias factor, COV, and distribution type.

Primary Expressions in Chinese Design
Codes. e reliability indices for structures or components designed according to Chinese codes representing the minimum design requirements can be determined by calibration of reliability. Based on Chinese codes "DLT5154-2012" [11] and "GB 50545-2010" [12], design expression with partial safety factor for ultimate limit state is written as where c 0 is the coefficient for importance of structure, not less than 1.1 for the important transmission line, 0.9 for the temporary transmission line, 1.0 for the other transmission line;cG is the partial safety factor of permanent load, not more than 1.1 in favorable conditions, 1.2 in unfavorable conditions; c Qi is the partial safety factor of variable load Q i taken as 1.4; S Gk , S Qik are, respectively, characteristic value of permanent load effect and variable load effect; R d is design value of components or structures, and different formulas are adopted for different tower members in Chinese code GB 50545-2010 [12]; Ψ is combination coefficient of variable load, 1.0 for normal operation condition, 0.9 for the design condition of conductor breaking, installation, and uneven icing, and 0.75 for checking calculation. According to (10), the design value of resistance satisfying the minimum design requirements specified by Chinese codes can be expressed by

Expression of Resistance
(1) Strength Calculation of Axially Loaded Members. e strength of tower member in axial stress given by Chinese codes is written as [11] where A n is the net cross-sectional area of tower member and f d is the design value of steel strength. e axial force should meet the minimum design requirements described by (12) and can be expressed as follows: Substitution of (13) into (11) leads to in which N Gk and N Qik are, respectively, the characteristic value of permanent load and variable loads.

(2) Stability Calculation of Axial Compression Members.
Correspondingly, the stability calculation of axial compression member should satisfy the following expression given by Chinese code "DL/T 5154-2012" [11]: where ϕ k is the characteristic value of stability coefficient ϕ for stability calculation of tower member in axial compression. en, Equation (11) can be rewritten by the following formula:

Calculation of Reliability
Index. e variable loads of transmission tower include wind, ice, conductor, or ground wire broken loads, but for the sake of simplicity, only the extreme wind load case and the combined wind and ice load case are considered in this work. As shown in Table 4, in addition to the statistical parameters of resistance, the function and statistical parameters of load effects are taken as the same values for both strength and stability calculation of axially loaded tower members. erefore, for illustrative purpose, calculation of reliability index for strength calculation of axially loaded tower members is taken as example.
(1) Extreme Wind Load Case. e performance function of transmission tower for strength calculation of axially loaded member can be expressed as follows: where R is the resistance; N G is the axial force resulting from permanent load acting on tower member; and N W is the axial force resulting from the extreme wind load acting on tower member. Advances in Civil Engineering e mean and standard deviation of permanent load and wind load can be expressed as follows: where k N G , k N W , δ N G , and δ N W are the bias factors and COV of permanent load and wind load, respectively; ρ W is the wind load effect ratio, which is the ratio of the characteristic value of wind load to the characteristic value of permanent load and can be expressed as ρ W � N Wk /N Gk . e mean of resistance can be expressed as follows: e characteristic value f k and the design value f d of steel strength for various steel classes and thicknesses are summarized in Table 5.
Wind loads on power transmission tower include the wind loads acting on the tower members as well as conductors, overhead ground wires, insulators, and fittings. When determining wind loads acting on conductors and ground wires designed by Chinese code "GB50545-2010," wind load adjustment factor β c shown in Table 6 should be taken into account. Experience shows that wind loads acting on conductors, overhead ground wires, insulators, and fittings can be taken half of the total wind loads on tower. erefore, by substituting (14) into (20), the following equations can be obtained: As seen in (18), (19), and (21), the mean and standard deviation of structure resistance, permanent load effect, and wind load effect are proportional to the characteristic value of permanent load N Gk . It is well known that the value of reliability index depends on the wind load effect ratio ρ W rather than the specific value of load effects and resistance. Analyses of numerous actual power transmission towers have shown that the ratio ρ W tends to lie within a wider range of values ranging from 0.1 to 100. e most widely used approach for reliability analysis is JC method. en reliability indices β of tower members for various steel classes with different wind load adjustment factors β c are calculated according to JC method in this work, as shown in  Table 7.
It can be seen from Figures 2 and 3 that the reliability indices of power transmission tower members relate directly to the ratio ρ W . For ρ W ≤ 1.0, the reliability indices for strength calculation of axially loaded tower members increase with the increased ρ W ; meanwhile, for ρ W > 1.0, the reliability indices decrease with the increased ρ W . As shown in Table 7, the average values of reliability indices increase with the wind load adjustment factor β c , while the average values of reliability indices for stability calculation of axial compression members are always lower than those for strength calculation.
(2) e Combined Wind and Ice Load Case. Axially loaded tower members in strength calculation are still taken as examples herein. In the combined wind and ice load case, the performance function of transmission tower member for strength calculation of axially loaded member can be expressed as follows: So, the mean values and standard deviations of wind load and ice load in the combined wind and ice load case are written as follows: where k N WI ,k N I , δ N WI , and δ N I are, respectively, the bias factors and COV of wind load and ice load in the combined wind and ice load case; ρ W I is the wind load effect ratio, ρ W I � N Wk /N Gk ; ρ I is the ice load effect ratio,ρ I � N Ik /N Gk .
According to Table 6, β c for the combined wind and ice load case is equal to 1, since the wind speed of 10 m/s is assumed to be the maximum wind speed for the condition of annual maximum ice thickness. erefore, the mean of resistance for stability calculation of axial compression member is defined as: where c Q 2 is the partial factor of ice load. For the combined wind and ice load case, ice loads would dominate the reliability of tower. In accordance with the analysis results of actual power transmission towers, the ratio ρ W I tends to lie within the range 0.02∼10, while ρ I ranges from 0.02 to 5.
To compute the reliability indices of tower members in the combined wind and ice load case, JC method is employed in the present article. e reliability indices for different ρ W I and ρ I are depicted in Figures 4 and 5; meanwhile, the average values of reliability indices for various steel classes are shown in Table 8.

Advances in Civil Engineering
As can be seen in Figures 4 and 5, when the ice load effect ratio ρ I is taken a fixed value, the reliability indices increase with the ratio ρ W I , while the reliability indices decrease with ρ I in the case of a fixed wind load effect ratio ρ W I . Since the reliability in the combined wind and ice load case is governed mainly by ρ I , a higher reliability level is produced for ρ W I > 0.1 and ρ I < 0.1, as compared with the extreme wind load case. However, the average reliability indices of axial compression member under stability are lower than those of axially loaded member for strength calculation, as shown in Table 8.

Reference Value of Reliability Index β 0 .
e reliability analysis is carried out for transmission tower designed by Chinese codes "GB 50545-2010" and "DL/T 5154-2012," and average values of the above-mentioned reliability indices shown in Tables 7 and 8 are listed in Table 9.
It can be seen from Table 9 that the average values of reliability indices for strength calculation of axially loaded member are 3.1028 in the extreme wind load case and 3.3479 in the combined wind and ice load case, respectively. e average reliability indices for stability of axial compression member are, respectively, 2.9949 in the extreme wind load case and 3.1983 in the combined wind and ice load case. Based on the analysis and adjustment of the average reliability indices, the reference value of reliability index of transmission tower is presented for transmission line employing 50-year design reference period and crossing high-speed railway in this work, as shown in Table 10. e reference value shown in Table 10 represents the reliability level of transmission tower meeting the minimum design requirements specified by "GB 50545-2010" and "DL/T 5154-2012."

Safety Classes.
In general, it is difficult to quantificationally calculate the loss due to structural failure.  erefore, safety classes of engineering structure are usually defined by a qualitative analysis method combined with engineering experience. Safety of engineering structure can be classified into three classes in design codes of many countries, such as Eurocode "Basis of structural design (EN 1900:2002)" [19] and Chinese standard "Unified standard for reliability design of engineering structures (GB 50153-2008)" [15]. erefore, Classes 1, 2, and 3 are specified for transmission line crossing high-speed railway, with Class 1 being the strongest. Tension section across high-speed railway is defined as Class 1, tension section not crossing high-

Recommendation of Target Reliability
Index. e numerical values of the reliability are often described on the basis of the reliability index β defined by β = −Φ −1 (P f ), in which P f is the failure probability. e relationship between β and P f is given in Table 11. It is well recognized that the safety degree of engineering structure is medium and high for 10 -3 < P f < 10 -4 and 10 -4 < P f < 10 −5 [20].
Calibration method is a very simple and practicable method applied to define the structural safety level. erefore, target levels for reliability are often based on calibration [20]. Target reliability indices β T (i.e., recommended minimum values for reliability index) stipulated in Eurocode "EN 1900:2002 and Chinese standard "GB 50153-2008" [15] are listed in Table 12.
As can be seen in Table 12, there is a difference of about 0.5 for reliability indices between the adjacent safety classes, while the probability of failure can approximately differ by an order of magnitude.
Based on the above section, the acceptable safety level of transmission tower Class 2 should be in agreement with the average reference values described in Table 10. us, a target reliability index of 3.2 is suggested for Class 2. en a reliability index of 3.7 is recommended for Class 1 in this study for the reason that adjacent safety classes have a difference of about 0.5 for reliability indices. Since transmission line of Class 3 is temporary, functional failure does not lead to serious consequences, reliability can be reduced accordingly. Consequently, a reliability index of transmission tower for Class 3 can be 2.7.

Discussion
tIn order to determine the reliability of transmission line across high-speed railway, reliability calibration for steel angle tower of transmission line satisfying the minimum design requirements specified by Chinese codes was performed.
e reliability indices of the tower for the extreme wind load case and the combined wind and ice load case are shown in Figures 2-5, respectively. Based on the aforementioned results, some discussions are provided as follows: (1) In the extreme wind load case, reliability indices for strength calculation of axially loaded member approximately range from 2.60 to 3.61, whereas the indices for stability calculation of axial compression      It can be seen that the reliability level of transmission tower designed by Chinese code is about similar with that specified in NESC and higher than that specified by CAS for extreme wind load case and the combined wind and ice load case. (3) ere are some deficiencies in statistical parameters used for the reliability analysis in this work. Ideally, the statistical data of various load effects, such as wind load and ice load used for the reliability analysis should be obtained from the field. However, due to the lack of such information required, an alternative approach was used in this work. For instance, parameters of building structure stipulated in Chinese code were used as statistical parameters of permanent load and wind load in the extreme wind load case. Statistical parameters of wind load in the combined wind and ice load were derived by 50 year return annual maximum wind speed. Parameters of ice load were determined based on a statistical analysis of limited field data. Statistical parameters of resistance of tower for stability calculation were derived by considering the uncertainty of stability coefficient. (4) e calibration results show that the reliability level of tower satisfying the minimum requirements in Chinese code needs to be improved. In this study, the recommended target reliability indices were given based on the average reliability indices obtained from reliability calibration and the recommended values of engineering structures suggested in the codes of many countries.

Conclusions
Based on the analysis results obtained, the following conclusions can be drawn: (1) e reliability indices achieved lie within a large range, which may be caused by a wide range of ρ w and ρ I . e reliability level of transmission tower is similar with that of American standards but higher than that of Canadian specifications. (2) e average reference value of reliability index was taken as approximately 3.2. e acceptable safety level of transmission tower Class 2 should be in agreement with this value.
(3) Classes 1, 2, and 3 were specified for transmission line crossing high-speed railway, with Class 1 being the strongest. For tower of transmission line across high-speed railway, the target reliability indices of Class 1, Class 2, and Class 3 were recommended as 3.7, 3.2, and 2.7.

Data Availability
e data used to support the findings of this study are included within the article.

Disclosure
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of Shandong Provincial Natural Science Foundation, China and Natural Science Foundation of Liaocheng University.

Conflicts of Interest
e authors declare that they have no conflicts of interest.