Selection of 110 kV Prefabricated Steel Substations Considering Seismic Vulnerability in China

Abstract

Prefabricated modular substations are expected to become the mainstream construction type for substations in China. However, there is a lack of scientific basis for structural selection and seismic performance evaluation. Taking a 110 kV substation as an example, this study compares the construction cost of cast-in situ reinforced concrete (RC) substations and prefabricated steel (PS) ones. Two types of PS structures are considered: one with H-section steel columns and the other with box-section steel columns. A seismic vulnerability analysis is performed to compare the probability distribution of various damage states of substation building structures under different seismic damage levels. Results indicate that the construction cost of PS structures is approximately 27.9% higher than that of cast-in situ concrete. When using H-section steel columns, there is a significant difference in the flexural stiffness in two horizontal directions, resulting in reduced seismic performance in the weak-axis direction. The construction cost of using box-section steel columns is slightly higher than that of the H-section steel case, but its seismic performance is significantly improved. Although the probability of slight and moderate damage states for the box-section steel column scheme is generally higher than that of the cast-in situ RC scheme, the probability of collapse is reduced. Thus, box-section steel columns are recommended for prefabricated modular substation building structures.

1. Introduction

Prefabricated modular structures offer several advantages, including standardized design, factory processing, and modular construction. These features effectively shorten construction periods, reduce energy consumption, and promote green construction. Substation buildings, which are characterized by complex layouts, large spans, uneven load distribution, and partial hollow areas, benefit significantly from modular approaches. Steel structures, which are easy to fabricate and install, are commonly employed to achieve prefabrication and modularity in substation buildings. Engineering practice has demonstrated that prefabricated steel structure (PSS) substations can effectively address various construction challenges, such as material transportation difficulties, harsh construction conditions, and long construction periods, thereby significantly enhancing the quality and efficiency of substation construction [1,2].

Seismic activity is one of the controlling factors in substation structural design [3,4] and the seismic resistance of substation building structures has been extensively investigated. Wen et al. [5,6,7,8] conducted a series of studies on the nonlinear dynamic behavior of substations, discussing the interaction between electrical equipment and buildings, and conducting seismic vulnerability research. Kang et al. [9] examined the seismic design challenges faced by indoor substations in high-intensity areas, researching seismic fortification objectives and corresponding design methods. Zhang Xiuli et al. [10] established a theoretical system for assessing the seismic resilience of substations, quantitatively studying the functional level and seismic resistance of substations. However, most of these studies focus on cast-in situ reinforced concrete (RC) substation structures, with relatively little consideration given to prefabricated steel (PS) substation structures. Compared to RC structures, PS ones have lighter self-weight, lower stiffness, and different structural arrangements, leading to significant differences in seismic performance. Zhang Cui [11] conducted a static pushover analysis and incremental time history analysis to analyze the seismic performance of a 220 kV steel structure indoor substation; however, this study mainly focused on the elastic stage. Therefore, there is an urgent need to explore the seismic performance of prefabricated steel structure substations, discuss their differences from traditional reinforced concrete structures, and conduct related research.

This paper reviews the design, mechanical characteristics, and construction cost of three types of substation building structural schemes: reinforced concrete columns, prefabricated H-section steel columns, and prefabricated box-section steel columns. Utilizing seismic vulnerability analysis methods, the damage states of substation building structures are determined based on probability distributions under various seismic damage levels. Then, seismic performance among different schemes is thoroughly compared and their differences are discussed, which provides an essential reference for the rational selection and seismic design of substation building structures.

2. Schemes of RC and PS Structures

The 110 kV substation is located in the seventh-degree anti-seismic zone in South China. The total height of the building is 17.80 m and the ground area is 859 ${\mathrm{m}}^2$ with a total floor area of 2576.1 ${\mathrm{m}}^2$. The design follows the Southern Power Grid “Standard Design for 110 kV to 500 kV Substations” code. The building is primarily divided into four levels: one underground level with a height of 4.6 m, two above-ground levels with heights of 5 m and 4.5 m, respectively, and one rooftop level with a height of 4.5 m, as depicted in Figure 1. For simplicity, the planar long side represents the x-direction, the short side represents the y-direction, and the vertical direction represents the z-direction hereafter (see Figure 1). Following the same architectural layout, three structural schemes are considered as shown in Figure 2:

Scheme One: RC structure. RC structures are frequently employed in indoor substations due to their inherent stability, well-established construction techniques, and cost-effectiveness. Grade C30 concrete, as per Chinese standards, is utilized for both beams and columns. Based on the basic architectural data and floor load conditions, the column cross-section dimensions for the underground level are 1000 mm × 1000 mm square, while the beam dimensions are 250 mm × 800 mm. For the first and second above-ground levels, the column dimensions are reduced to 700 mm × 700 mm and 600 mm × 400 mm, respectively, accompanied by beam dimensions of 250 mm × 600 mm and 250 mm × 500 mm, respectively. The reinforcement ratio of the columns is 1.37% in the basement whereas it ranges from 1.80~2.0% on the ground floors.

Scheme Two: PS Structure with H-section Columns. PS structures replace original reinforced concrete structures to improve construction quality and shorten construction time. H-section steel columns, known for their uniform cross-sectional extension and low internal stress during rolling, are widely utilized in steel structures. For underground steel columns, concrete wrapping forms composite steel-concrete columns for two primary reasons: (1) mitigating steel corrosion in humid underground environments and (2) enhancing load-bearing capacity under heavy floor loads, thus optimizing steel usage. For the above-ground floors, only steel columns are used for support. The strong axis of the H-section steel column aligns with the x-direction of the structure, while the weak axis aligns with the y-direction. Steel beams, also H-section steel, are connected to columns via high-strength bolts. Connections are categorized into strong-axis and weak-axis connections based on the orientation of the H-section columns’ axes, as depicted in Figure 3. Research conducted by Lu [12] investigates the static mechanical properties of beam–column joints, advocating for this connection method to ensure rigid joints at the nodes.

Scheme Three: PS structure with box-section columns. The RC columns are replaced with equivalent box-section columns. The beams remain the same as in Scheme Two. Rigid connections between beams and columns employ the strong-axis method depicted in Figure 3. Compared to the H-section steel column, the box-section profiles have equal bending stiffness in two horizontal directions, thereby effectively enhancing the section’s torsional stiffness.

In normal operating conditions, columns primarily withstand compression, while beams primarily resist bending. In both Scheme Two and Scheme Three, the design of columns adheres to the principle of equal axial load-bearing capacity, while beams are designed based on equal bending resistance within the reinforced concrete (RC) structure. Final adjustments are made according to static load conditions. The results of the column sections for Scheme Two and Scheme Three are illustrated in Figure 2.

Table 1. Comparison of bending rigidity of columns in different schemes of 110 KV substation.

Scheme	Position	Flexural Stiffness (EI, 104 kN·m2)
		In the x-Direction (Longitudinal Direction)	In the y-Direction (Transverse Direction)
Reinforced concrete structure	Underground floor	262.50	262.50
	Above-ground floor	63.03	63.03
Prefabricated steel structure (H-shaped columns)	Underground floor	141.81	132.40
	Above-ground floor	13.46	4.62
Prefabricated steel structure (box-shaped columns)	Underground floor	143.95	143.95
	Above-ground floor	12.46	12.46

compares the bending stiffness (EI) of the columns across three structural schemes. Both the RC structure and the PS structure with box-section columns exhibit equal stiffness in the x- and y-directions. In contrast, the PS structure with H-section columns displays varying bending stiffness between the two directions. Notably, at underground levels where external concrete wrapping predominantly influences flexural stiffness, directional stiffness disparity is minimal. However, on above-ground floors, the x-direction stiffness is approximately three times greater than that in the y-direction due to the asymmetric nature of the H-section column. Moreover, RC columns generally exhibit greater stiffness compared to PS columns on each floor. Specifically, H-section steel columns demonstrate similar x-axis stiffness to box-section steel columns, both of which are lower than those of reinforced concrete structures. Notably, at the above-ground level, the stiffness of the PS column section is merely 1/5 that of the reinforced concrete structure.

The construction cost of the substation structure is a primary concern for stakeholders.

Table 2. Comparison of construction cost of different schemes of 110 KV substation.

Scheme	Material Usage			Construction Cost (CNY)
	Steel Plate Q355 (t)	Reinforcement HRB400 (t)	Concrete C30 (m3)
RC structure	0	58.5	1072.1	7,300,000
PS structure with H-shaped columns	298.2	23.0	421.2	9,600,000
PS structure with box-shaped columns	310.4	23.4	428.4	9,730,000

Notes: The main structural cost consists of the material cost and other relevant costs, including labor costs, machine usage costs, and so forth. The calculation is according to the internal regulations of the Southern Power Grid; however, the information is not open to the public due to financial and confidentiality reasons.

compares the material usage and resulting construction costs of the three schemes. The amount of steel used in the PS structure with box-section columns is slightly higher than in the PS structure with H-section columns, while the other material usage is similar across the two schemes. Additionally, adopting PS structure schemes increases the construction cost of the substation compared to conventional reinforced concrete structures. Specifically, the PS structure with H-section columns and PS structure with box-section columns schemes induce additional costs of CNY 2.4 million and CNY 2.56 million, respectively, representing an increase of approximately 26.4% and 27.9% over the RS Scheme. Note that only the construction costs are considered here. When factoring in the entire investment of the substation, including equipment procurement fees, installation engineering fees, and other expenses, the increase in construction costs for PS structures decreases significantly, by no more than 10%.

3. Finite Element Model and Elastic Seismic Analysis

The three-dimensional nonlinear finite element model of the substation building structure was established using ETABS finite element software (V18). Figure 4 illustrates the schematic diagram of the finite element model for the substation building structure. Beam–column elements were utilized to simulate columns and beams, incorporating bending plastic hinges to capture the nonlinear behavior of the columns and beams. Plastic hinges were placed on both ends of the beams with defined moment–rotation relationships, without considering axial force effects. Columns were modeled using P-M-M hinges to account for the coupling effect of axial force and bidirectional bending moment, following the specifications of ASCE41-17 [13]. Steel materials were represented using a bilinear model, whereas the concrete material employs Mander’s Model for the simulation, as depicted in Figure 5. Additionally, Figure 4 shows the uniform distribution of gravity loads from electrical equipment in each room, based on load values referenced from the Technical Code for the Design of Substation Buildings and Structures [14]. It is noted that these loadings are substantially higher than typical building loads (normally 4 kN/m distributed), indicating the electrical equipment would have a great impact on the seismic response of the substation building structure.

The seismic response of structural schemes needs to be assessed within the elastic range according to the “Code for Seismic Design of Buildings” [15]. Therefore, an elastic analysis of the structure is conducted, and the structural response under minor seismic excitations is calculated using the modal decomposition response spectrum method. The response spectrum analysis is simple to conduct while giving a general view of the seismic performance of three substation structure schemes.

In the elastic analysis, the damping ratio is set to 0.05 for reinforced concrete structures and 0.04 for prefabricated steel structures. The results of the elastic analysis are summarized in

Table 3. Elastic Design Results.

Type		RC Structure	PS Structure with H-Shaped Columns	PS Structure with Box-Shaped Columns	Limit Requirements in Chinese Code
Period/s	T1 (Y-direction lateral displacement)	0.704	1.242	1.129	-
	T2 (X-direction lateral displacement)	0.598	1.207	0.936	-
	T3 (Torsion)	0.544	0.871	0.890	-
Cycle ratio	T3/T1	0.77	0.70	0.79	<0.9
Total mass/t		2680.32	1294.67	1306.54	-
Maximum inter-story drift	X-axis Y-axis	1/1283 1/1286	1/1331 1/1165	1/1397 1/1208	<1/550 <1/550

. It is observed that all three schemes meet the seismic performance specified in the code with respect to inter-story drift. However, they exhibit different response characteristics. Specifically, the RC structure demonstrates a relatively shorter period, and there is no significant difference in inter-story drift angles between seismic actions in the x- and y-directions. Conversely, the PS structure shows a noticeable increase in natural periods and exhibits varied inter-story drift between the x- and y-directions. While both schemes display similar inter-story drift in the x-direction, the PS structure with H-section columns experiences the greatest inter-story drift in the y-direction. This disparity can be attributed to its reduced stiffness in that direction, resulting in the development of larger inter-story drift.

4. Seismic Vulnerability Analysis

Under the effect of major earthquakes, substation building structures may develop an inelastic response, leading to structural damage or potential collapse. A seismic vulnerability analysis plays a crucial role in predicting the likelihood of different damage levels under varying seismic intensities. Conducting seismic vulnerability analyses on substation building structures facilitates a reasonable assessment of their seismic performance across different structural types. This further helps designers improve the seismic resistance of structures in a targeted manner based on their vulnerability. In this study, seismic vulnerability analysis methods are employed to compare the seismic performance of the three types of substation building structures mentioned above.

4.1. Theories of Seismic Vulnerability Analysis

The seismic vulnerability of a structure macroscopically describes the relationship between structural damage and seismic intensity, and its mathematical expression is

$$P_f=P(DM>C|IM=x)$$

(1)

where $P_f$ represents the vulnerability function of the structure, also known as the failure probability; $IM$ is the seismic intensity index, in this paper, it is the peak ground acceleration (PGA); $DM$ represents the seismic demand parameter of the structure; and $C$ denotes the structural seismic capacity parameter.

Extensive research [16,17,18,19] indicates that the regression relationship between the seismic intensity parameter $IM$ and the structural seismic demand parameter $DM$ can be described as

$$DM=\alpha {\left(IM\right)}^{\beta }$$

(2)

Taking the logarithm of both sides of the above equation, it yields

$$\mathrm{l}\mathrm{n}DM=\mathrm{l}\mathrm{n}\alpha +\beta \mathrm{l}\mathrm{n}\left(IM\right)$$

(3)

where $\mathrm{l}\mathrm{n}\alpha$ and $\beta$ are obtained through statistical regression of data.

On the other hand, the structural seismic demand parameter $DM$ and the structural seismic capacity parameter $C$ are variable in practical engineering. It can be assumed that both of them follow a lognormal distribution:

$$\begin{gathered} \mathrm{l}\mathrm{n}DM~N\left(u_{\mathrm{d}},{\sigma }_{\mathrm{d}}^2\right) \\ \mathrm{l}\mathrm{n}C~N\left(u_{\mathrm{c}},{\sigma }_{\mathrm{c}}^2\right) \end{gathered}$$

(4)

where $u_{\mathrm{d}}$ and ${\sigma }_{\mathrm{d}}$ are the logarithmic mean and standard deviation of the structural seismic demand parameter, respectively, and $u_{\mathrm{c}}$ and ${\sigma }_{\mathrm{c}}$ are the logarithmic mean and standard deviation of the structural seismic capacity parameter, respectively. Substituting Equation (4) into Equation (1), the failure probability $P_f$ can be calculated as follows:

$${\mathrm{P}}_{\mathrm{f}}=\mathrm{P}\left(\frac{{\mathrm{u}}_{\mathrm{d}}-{\mathrm{u}}_{\mathrm{c}}}{\sqrt{{\mathrm{\sigma }}_{\mathrm{d}}^2+{\mathrm{\sigma }}_{\mathrm{c}}^2}}>0\right)=\mathrm{\Phi }\left(\frac{\ln \mathrm{\alpha }+\mathrm{\beta }\ln \left(\mathrm{PGA}\right)-{\mathrm{u}}_{\mathrm{c}}}{\sqrt{{\mathrm{\sigma }}_{\mathrm{d}}^2+{\mathrm{\sigma }}_{\mathrm{c}}^2}}\right)$$

(5)

where $\mathrm{\Phi }(·)$ is the standard normal distribution function. In Equation (5), the parameter ${\sigma }_{\mathrm{d}}$ can be determined based on the statistical regression of the IDA analysis results. However, obtaining the parameter ${\sigma }_{\mathrm{c}}$ from IDA or alternative methods remains challenging. For simplicity, $\sqrt{{\mathrm{\sigma }}_{\mathrm{d}}^2+{\mathrm{\sigma }}_{\mathrm{c}}^2}$ is taken to be 0.5 according to the US HAZUS99 specification [20,21,22,23], which has been widely accepted and applied as a practical solution.

4.2. Seismic Demand Analysis

The site of the case substation is categorized as Zone 7 according to the “Code for Seismic Design of Buildings” [15] characterized by medium–hard soil with a shear wave velocity ranging from 250 to 500 m/s. Based on the site condition, 10 ground motion records—with a moment magnitude in the range of 6.0~8.0 and a scale factor from 0.25 to 4—from the PEER earthquake database were initially selected as seismic inputs [24]. Figure 6 illustrates a comparison between the response spectrum and the design seismic response spectrum. The mean values of the spectra of the 10 ground motion records generally fit in well with the design response spectrum, indicating that these ground motions can be used for seismic response analysis.

The Incremental Dynamic Analysis (IDA) method is utilized for conducting nonlinear time history analyses on three models. The peak ground acceleration (PGA) of the seismic motion ranges from 0.1 g to 1.5 g. By integrating the effects of various structural components and floor heights and also having a high correlation with the degree of structural damage, the maximum inter-story drift at each floor level is used to define the structural seismic demand parameter $DM$ under the corresponding PGA. The seismic input is applied in either the longitudinal or transverse direction of the substation building structure, while the seismic demand parameter DM represents the maximum response under both loading cases. Using the logarithm of the PGA values as the horizontal axis and the logarithm of the inter-story drift angles as the vertical axis, the parameters $\mathrm{l}\mathrm{n}\alpha$ and $\beta$ of the exponential distribution function can be obtained through linear regression of the IDA data results using the least squares method according to Equation (3).

4.3. The Establishment of Vulnerability Curves

HAZUS [20] defines four levels of seismic damage states: slight damage, moderate damage, severe damage, and complete damage, which are widely recognized and applied.

Table 4. Structure damage state and corresponding inter-story displacement angle limit.

Structural Type	Slight Damage	Moderate Damage	Severe Damage	Collapse
Reinforced concrete substation	1/150	1/75	1/45	1/40
Steel structure substation [25,26,27]	1/250	1/100	1/50	1/25

provides the corresponding inter-story drift angle limits for different damage states for both reinforced concrete and steel structures.

By substituting the parameters $\mathrm{l}\mathrm{n}\alpha$ and $\beta$ obtained from the seismic demand analysis into Equation (5) and taking $u_{\mathrm{c}}$ as the limit value from Table 4, vulnerability curves for different damage states of the structures can be calculated, as shown in Figure 7. Compared to reinforced concrete structures, the vulnerability curves of PS structures with H-shaped columns exhibit abrupt changes, indicating a rapid increase in exceedance probability with higher seismic intensity. The exceedance probability of the PS structure with box-shaped columns is slightly lower than that with H-shaped columns but generally higher than RC structures.

According to regulations, for a Zone 7 seismic area, the corresponding PGAs for minor, moderate, and major earthquakes are 0.035 g, 0.1 g, and 0.22 g, respectively. Substituting these values into Equation (5), the failure probabilities $P_f$ are calculated as shown in

Table 5. Seismic damage probability matrix of 110 kV structure schemes.

Structural Scheme	Design Seismic Intensity	Damage State
		Slight Damage	Moderate Damage	Severe Damage	Collapse
Reinforced concrete structure	Minor earthquake	2.10%	0.40%	0.10%	0.08%
	Moderate earthquake	14.30%	4.40%	1.50%	1.10%
	Major earthquake	33.30%	14.30%	6.20%	5.00%
Prefabricated steel structure (H-shaped columns)	Minor earthquake	6.87%	0.87%	0.11%	0.01%
	Moderate earthquake	21.16%	8.34%	1.98%	0.32%
	Major earthquake	65.54%	31.16%	12.19%	3.29%
Prefabricated steel structure (box-shaped columns)	Minor earthquake	0.14%	0.01%	0.00%	0.00%
	Moderate earthquake	2.97%	0.43%	0.07%	0.05%
	Major earthquake	18.37%	4.99%	1.41%	1.01%

.

From Table 5, it can be observed that when subjected to minor earthquakes, all three substation structures are minimally affected, with the probability of experiencing slight damage not exceeding 7%. The probabilities of experiencing moderate damage, severe damage, and collapse are all below 1%. Regarding moderate earthquakes, both RC and PS structures with box-section columns have probabilities of experiencing the four damage states below 15%. However, PS structures with H-section columns have a probability of slight damage exceeding 20% while probabilities of moderate damage, severe damage, and collapse are below 15%. Major earthquakes significantly impact PS structures with H-section columns, where probabilities of slight damage exceed 50%, severe damage exceed 10%, and collapse exceed 3%. Meanwhile, RC structures and PS structures with box-section columns have probabilities of slight damage, moderate damage, and severe damage exceeding 15%, 5%, and 1%, respectively.

The discussion above indicates that the seismic performance of PS structures with H-shaped columns is generally lower compared to the other two schemes. This is attributed to the H-shaped steel column’s lower flexural stiffness and deformation resistance in the weak-axis direction, resulting in larger seismic responses. With symmetric stiffness in two directions, the seismic performance of PS structures with box-shaped columns is deemed suitable for substation structures’ seismic resistance. The probabilities of experiencing four damage states are lower than those of RC structures.

5. Conclusions

This paper discusses the design, mechanical characteristics, and construction costs of cast-in-place RC structures and PS structures for a 110 kV substation. Seismic vulnerability analysis methods are employed to analyze the seismic demand probability distribution of different structures, establishing seismic vulnerability curves for various damage states. This analysis provides probabilities of structural damage under three different seismic design levels. A comparative assessment of seismic vulnerability between cast-in-place RC and PS substation structures is conducted to inform the selection of prefabricated substation structures. The main research conclusions of this paper are as follows:

(1)

PS structure substations can utilize either H-section or box-section steel columns, with minimal difference in construction costs between the two options, approximately increasing by 27.9% compared to the cast-in-place RC structures.

(2)

H-section steel columns exhibit significant differences in flexural stiffness between horizontal directions, resulting in greater seismic response in the weak axis under seismic action. Consequently, their seismic performance is inferior to RC structures, with higher probabilities of experiencing slight damage, moderate damage, severe damage, and collapse.

(3)

In contrast, box-section steel columns demonstrate a more rational structural stiffness arrangement, leading to reduced seismic responses compared to the H-section steel column. Meanwhile, the probabilities of slight damage, moderate damage, severe damage, and collapse are marginally lower than those for RC structures. Therefore, for cost-effective substation construction that meets required standards, PS structures with box-section steel columns are recommended for prefabricated modular substation structures.