Energy Consumption Analysis and Optimization of Substation Building in Cold Regions Considering Various Influence Factors

Abstract

Building-energy consumption constitutes a pivotal component of global energy systems, with the heating and cooling loads during the operational phase being particularly significant. Substation building, as nodes in the transmission and transformation network, deserve attention for their building-operating loads. This study investigates heating and cooling loads during substation operation in severe cold climates. By integrating energy consumption simulations with one-factor-at-a-time and orthogonal multivariate analyses, optimization strategies under key influencing factors are systematically explored. The impact analysis identifies the following order of influence magnitude on substation total loads: indoor equipment heat generation, ventilation rate, roof U-value, exterior wall U-value, and window U-value. The heating- and cooling-load characteristics exhibit distinct patterns depending on indoor equipment heat generation. The total building load can be reduced by 61.23 per cent under multifactor optimal de-sign conditions, highlighting the critical role of systemic design coordination. This study provides a case study reference for energy efficient design of heating and cooling loads in substations, especially where significant changes in equipment heat occur, and highlights the importance of controlling indoor heat sources to achieve optimal energy efficiency.

1. Introduction

The construction industry is a significant global energy consumer, responsible for approximately 37% of global energy demand and carbon emissions in the energy sector. Additionally, it accounts for about 30% of building-operational-energy consumption (BOEC) [1]. In China, the energy consumption linked to buildings has grown substantially, reaching 2.27 billion tonnes of standard coal equivalent (tce) in 2020, which represented 45.5% of the nation’s total energy consumption. This includes energy used across three primary phases: the production phase of building materials (1.11 billion tce), the construction phase (0.09 billion tce), and the operational phase (1.06 billion tce). The production phase alone constitutes 22.3% of China’s total energy consumption, the construction phase accounts for 1.9%, and the operational phase represents 21.3%. These figures underscore the substantial energy demand within the building sector, particularly in the production of materials and the operational use of buildings, highlighting the need for energy efficiency improvements at all stages of the building lifecycle [2].

There exists a plethora of methodologies by which to achieve energy-efficient building design. These include the estimation of building consumption through the utilization of historical data, the employment of statistical methods, and the analysis of factors affecting building-energy consumption. This process facilitates the identification of key points for building-energy management and the implementation of improvements [3] and the prediction and analysis of building-energy consumption using machine learning [4] and deep learning models [5]. The use of simulation software to study the impact of input values for building-energy model parameters on simulation results is a crucial tool for exploring energy-efficient building design. For instance, dynamic simulations with software such as TRNSYS [6], EnergyPlus [7], and DeST [8] allow for detailed modeling and analysis of various parameters, including building dimensions, thermal properties, lighting, solar radiation, airflow, and the thermal characteristics of the building envelope. This approach helps to identify correlations between energy consumption and these factors, ultimately contributing to the optimization of energy efficiency, improvement of indoor environmental quality, and reduction in energy costs.

Various methods are available for analyzing building-energy consumption, including one factor at a time (OFAT) and variance-based sensitivity analysis, which can be integrated with EnergyPlus through MATLAB [9] to identify key variables influencing energy efficiency during the early stages of building design. The impact of different parameters is assessed through orthogonal design and numerical analysis [10]. In the decision-making process for building energy-saving renovation strategies, a multi-factor optimization approach that combines orthogonal experiments with entropy weighting is applied [11]. Genetic algorithms are also integrated with numerical simulations to evaluate parameter designs [12]. Research examines various factors affecting energy consumption, considering both the individual and combined effects of multiple variables.

In terms of building energy efficiency technology (BECT) research, passive building technology is a design concept that aims to minimize a building’s heating and cooling energy demand by fully utilizing natural resources and rationally planning the characteristics of the building itself. The key points of passive building technology include the following: adiabatic enclosure, window translucent curtain walls, minimizing thermal bridges in the envelope design, efficient ventilation and heat recovery systems, optimization of the building’s orientation and the internal layout to facilitate natural ventilation, shading, and lighting [13]. This approach focuses on reducing energy consumption while maintaining a comfortable indoor environment through design strategies that harness the natural surroundings and resources effectively. As demonstrated in extant research, the implementation of energy-efficient measures for curtain wall retrofit [14] in conjunction with optimal external window performance [15] has been shown to result in a reduction in cooling- and heating-energy consumption. Furthermore, strategies such as natural ventilation [16] and night ventilation measures [17] have been shown to mitigate overheating. The effect of climate on the building to achieve energy savings, the optimal design of the building envelope/HVAC system and natural cooling strategies [18], the optimization of the building mass group form, the geometry [19], and the combination of active and passive technologies are technical means to achieve energy efficiency in buildings.

Substations play a crucial role as integral components within the power transmission and transformation process in the power system. They enable voltage transformation, receive and distribute electrical energy, control the direction of power flow, and regulate voltage levels [20]. As industrial structures, substations have significantly different energy consumption patterns compared to typical public and residential buildings. Certain rooms within substations, such as distribution rooms and secondary equipment rooms, house a large amount of electrical equipment and numerous electrical panel cabinets. Due to the specialized industrial processes at substations, electrical equipment operates continuously, 24 h a day, leading to substantial heat generation [21]. This heat is transferred to the indoor air through convection and radiation, causing an increase in indoor temperature. Most electrical equipment have specific temperature requirements for optimal operation as well as inspector comfort requirements, which directly contribute to higher air conditioning energy consumption. In outdoor substations, main transformers and primary high-voltage electrical equipment are usually placed outdoors to promote heat dissipation via natural ventilation. The energy consumption of substation buildings is influenced by various factors, including heat dissipation from secondary equipment [22], building insulation and ventilation, human activities, local weather conditions, and the energy efficiency of auxiliary equipment.

Heat generation within substations varies across different types of rooms, depending on their specific functions. The Code for Energy Efficiency of Public Buildings’ typically specifies an internal heat density of 25 W/m² for public buildings. In indoor substations, key functional rooms such as comprehensive protection rooms, station transformer rooms, and control rooms exhibit internal heat densities ranging from 20 to 80 W/m². In contrast, heat densities in main transformer cooling rooms can exceed 1200 W/m² [23]. For outdoor substations, heat generation values for critical cooling rooms, including the main control room, secondary equipment room, monitoring room, transformer room, and battery room, range from 23.01 to 100 W/m² [24,25,26].

This paper analyzes the application effect of several typical BECT (building energy consumption technology) in substation buildings from the design point of view, based on the difference in heat generation of indoor equipment in substation buildings. The methods of model construction and analysis are first described. Subsequently, the construction of the studied model is described, and the analysis of the building HL and TL is discussed in terms of single and multi-factor. The novelty of this paper lies in the fact that the cooling- and heating-load characteristics of the substation building are obtained based on the characteristics of the climatic background with different indoor equipment heat generation. And according to the characteristics of discrete building parameters in practice, an optimization scheme is given by using the method of multi-factor analysis. Case references and suggestions are provided for the energy-saving design of substation buildings with high heat generation.

2. Model Development

2.1. Model Construction Method

DeST is a dynamic simulation and analysis software for building and HVAC systems developed by Tsinghua University. The operations in this energy simulation software are performed on a user interface developed based on AutoCAD, where the user sets up various data (materials, internal disturbances, etc.) of the building, and a large number of case studies [27] and theoretical validations [28] have been carried out so far. The DeST version number used in this article is 20220712, based on AutoCAD version 2019.

This study integrates substation design principles, employing DeST software for comprehensive annual analysis of building heating and cooling loads. The DeST software is utilized for modeling, facilitating the specification of envelope structures and indoor-outdoor design parameters according to design requirements and relevant standards. Numerous factors in practice influence the energy performance of buildings, including variations in thermal mass among envelope structures, insulation strategies for exterior walls (EW), and the window to wall ratio (WWR). These factors interact with each other and fluctuate with changes in climate and indoor conditions. Direct and precise analysis of these relationships, building environments, and annual energy consumption relies on hourly dynamic simulations, enabling effective planning and implementation of improvements to systems, structures, and control strategies.

2.2. Building-Energy Modeling

The climatic characteristics of the region in which the building is located are a determining factor in the direction of building-energy-saving design. Based on the thermal design zoning and outdoor meteorological conditions outlined in the Design standard for energy efficiency of public buildings [29], this paper focuses on the analysis of a substation located in the cold zone A (Areas in China where the average temperature of the coldest month is ≤−10 °C or the average daily temperature is ≤5 °C for ≥145 days). Similar regions of the world are concentrated in the mid-to-high latitude continental interiors of the Northern Hemisphere, such as the interior of southern Russia and Siberia, the area north of the Great Lakes in North America, northern Japan, and parts of northern Europe. This study utilizes meteorological data specific to cold regions, including hourly outdoor dry bulb temperatures and solar radiation intensity, to ensure that the thermal design of the substation building meets the climatic conditions of its location. The basis of the time-dependent meteorological parameters in DeST is the meteorological observation data provided by the Meteorological Information Center of the China Meteorological Administration (CMA). A typical meteorological year is formed by constructing “average months” (or “standard months”) [30]. Figure 1 shows the monthly average maximum, average, and minimum outdoor dry-bulb temperatures for a typical meteorological year in the study area, which expresses the characteristics of the temperature in the region.

The structure and geometry of the substation building are constrained by the Code for design of 35 kV~110 kV substation [31], and building dimensions are field-measured parameters. The substation structural model established in this paper is shown in Figure 2. The building is arranged in a linear configuration, located in the Greater Khingan Range area of Northeast China. The rooms are equipment rooms containing numerous secondary devices that operate continuously throughout the day. Regular inspection personnel visit the site daily to perform routine checks. The normal operation of the equipment and the personnel inspection activities require the room temperature to be adjusted accordingly. The building dimensions are 17.5 m × 11.808 m × 3.213 m, with a single floor. The window dimensions are 2.0 m × 1.2 m, with a total window area of 206.64 m². The window-to-wall ratios for the east, south, west, and north walls are 0.06, 0.26, 0.06, and 0.04, respectively.

The present study investigates the impact of different thermal conductivity coefficients of envelope materials on building-energy consumption. In addition, it quantitatively analyses the thermal and cooling loads of buildings. To this end, this study constructs several typical external wall configurations to modify the wall’s thermal transmittance. Various wall configurations with different heat transfer coefficients (U) are achieved by adjusting the thickness of insulation materials, a common method in engineering practice. It is important to note that the external wall materials presented in this study serve as reference examples, and various constructions can achieve the same thermal transmittance criteria in practical applications. The heat transfer coefficient of various wall structures is calculated through a thermal resistance analysis, taking into account their thermal inertia in subsequent calculations. Additionally, the rates of indoor and outdoor ventilation significantly affect the building’s heat and cooling loads. Therefore, it is necessary to investigate the impact of different air exchange numbers on the building load. According to the relevant literature [32], the number of air exchanges in the building with fully open windows is 10 times/h, while no natural ventilation measures result in the number of air exchanges being 0.32 times/h.

Due to the characteristics of substations, indoor secondary electrical equipment continuously generates heat and operates throughout the day. To ensure the comfort of personnel conducting inspections and the stability of the thermal environment for equipment operation, cooling supply is required during the building’s operational phase, while also ensuring the provision of heating sources in cold regions. According to the “Unified Standards for Energy Efficiency Design of Industrial Buildings” (GB 51245-2017) [33], the thermal performance parameters of the building envelope for industrial buildings with heating and air-conditioning systems must meet the following requirements: U_rf ≤ 0.35, U_ew ≤ 0.4 and, when the window-to-wall ratio is less than 0.2, U_ow ≤ 2.5. The maximum allowable heat transfer coefficient for the building envelope design should not exceed U_rf ≤ 0.5, U_ew ≤ 0.6, and U_ow ≤ 3.

This study focuses on the calculation of the cooling and heating loads during the building’s operational phase. After completing the 3D model of the building, it is necessary to define the characteristics of building components, such as walls, roofs, and windows, as well as internal heat disturbances from sources such as personnel, lighting, and equipment, before proceeding with load calculations. The key parameters of the building are presented in

Table 1. Initial key parameter settings for buildings.

Factors	Set	Value
EW setting	Cement Mortar + Lightweight Aggregate Concrete + EPS + Lightweight Aggregate Concrete + Lime Mortar	0.137–0.54 (W/m2·K)
RF setting	Cement Mortar + Expanded Perlite + Reinforced Concrete + Cement Mortar	0.199–0.556 (W/m2·K)
OW setting	General insulating glass(6 + 12A + 6)	2.9–2.0 (W/m2·K)
VN setting	Open at night from 18 to 45 weeks	Maximum 2–10 times/h, minimum 0.32 times/h
Air-conditioning settings	Open all day	Upper limit 26 °C, lower limit 18 °C
Equipment thermal disturbance setting	Open all day	25–80 (W/m2)

.

The operation energy consumption of buildings is affected by many factors. The variables associated with passive building-energy-efficient design, as discussed in related research, primarily encompass several aspects. These include the insulation of the building’s envelope (e.g., the U-value of exterior walls and roofs), indoor thermal disturbances (e.g., personnel, lighting, equipment operation, and heat density), operational strategy settings (e.g., air conditioning temperature, ventilation, and shading), building-geometry parameters (e.g., shape coefficient and window-to-wall ratio), and the optimization of air-conditioning equipment design. This paper focuses on substation buildings as the research subject. Typically, lighting, personnel activities, and equipment operation follow a work–rest schedule, while the air conditioning temperature is maintained within the standard range of 18–26 °C. The window-to-wall ratio is calculated to vary within the range of 0.2–0.4, and the maximum change in total load is 1.58 kWh/m², which has little impact and will be limited by safety regulations in practice. For the building orientation, because the substation studied in this paper is rectangular, the southern orientation of the longer wall is the optimal condition. Moreover, the substation is located in the suburbs, so it is not necessary to consider the shielding of surrounding buildings. The above factors are not considered as variables because they are unchanged or have little impact in practice. The U value of the enclosure structure, the ventilation volume generated by the electrical secondary equipment in the room, and the heat of the equipment are selected as the analysis factors to study the quantitative relationship between them and the cooling and heating loads.

3. Analysis Methods

To analyze the relationship between the factors on the cooling and heating loads of the building, a linear equation was fitted using the least squares method [34]. The goal of the least squares method is to minimize the sum of squared errors (residual sum of squares):

$$y=mx+b$$

(1)

$$S(m,b)=\sum _{i=1}^n{\left(y_i-(m{x}_i+b)\right)}^2$$

(2)

where m represents the slope, and b is the intercept. y_i is the actual observed value, while mx_i + b is the predicted value derived from the linear model, both in kWh/m², representing the building-cooling and -heating loads.

The impact of various energy-saving measures on building-operational energy consumption can be assessed using the sensitivity coefficient (SC) [9]. The calculation formula for SC is as follows:

$$SC={\displaystyle \frac{OP-O{P}_{\mathrm{rf}}}{O{P}_{\mathrm{rf}}}}/{\displaystyle \frac{IP-I{P}_{\mathrm{rf}}}{I{P}_{\mathrm{rf}}}}$$

(3)

where OP is the result of the simulation output parameter of the building, i.e., the cooling- and heating-energy consumption of the building; OP_rf is the result of the simulation output parameter of the reference building; IP is the value of the simulation input parameter of the building; and IP_rf is the value of the simulation input parameter of the reference building.

Orthogonal design is a method used to study multi-factor and multi-level experiments, with the aim of efficiently conducting multi-factor analysis. In this paper, orthogonal design is applied to arrange building-simulation scenarios for calculation and result analysis. Utilizing standardized orthogonal arrays (designated as ${\mathrm{L}}_n(q^m)$) to arrange experimental designs, perform computational analysis of results, and, ultimately, swiftly identify optimization solutions represents an efficient scientific method for addressing multi-factor optimization challenges. Within the orthogonal experimental analysis, a range analysis serves as the most intuitive method, characterized by its computational convenience and ease of generalization. The formula for calculating the polar deviation is shown below [10]. The objective of the range analysis is to quantify and illustrate the impact range of each factor by determining the difference between the maximum and minimum mean values of test results. By comparing the magnitudes of ranges, one can effectively determine the relative importance of factors.

$${\mathrm{R}}_{\mathrm{j}}=\max (\overline{K_{j1}},\overline{K_{j2}},\cdots ,\overline{K_{jm}})-\min (\overline{K_{j1}},\overline{K_{j2}},\cdots ,\overline{K_{jm}})$$

(4)

where K_jm is the sum of the experimental indicators corresponding to the j-th factor m level, ${\overline{K}}_{jm}$ is the average value o K_jm, and R_j is the extreme variance of the factor in column j.

An analysis of variance (ANOVA) determines whether one or more factors (independent variables) significantly impact a continuous dependent variable. It computes sums of squares and degrees of freedom to derive mean squares, then uses the ratio of the between-group mean square to within-group mean square to calculate an F-value for conducting an F-test.

4. Results and Discussion

4.1. OFAT Analysis

This section verifies the baseline impact of relevant factors on cooling and heating loads through OFAT analysis, which is used as a basis for further proposing ways to optimize substation building loads under different equipment heat generation. The calculation results are shown in the following graphs. In Figure 3, Figure 3a–c are the annual cumulative heat load, cold load, and total load with the change in heat transfer coefficient of the exterior wall, and Figure 3d–f are the annual cumulative heat load, cold load, and total load with the change in heat transfer coefficient of the roof, respectively. And the results of OW and VN are shown in Figure 4. Figure 4a–c represent the graphs of annual cumulative heat load, cooling load, and total load variation with heat transfer coefficient from external windows. Figure 4d shows the graph of annual cumulative cooling load variation with ventilation rate. The obvious difference between various types of substation building rooms in the calculation of the thermal parameter lies in the difference in the heat generation of indoor equipment, therefore analyzing the heat generation of different equipment in the substation room. Then, the indoor heat generation takes different values, respectively, for 25, 40, 50, 65, and 80 W/m².

Through the OFAT analysis of the U-value of the substation building envelope, this study reveals the variability of thermal performance optimization under different equipment heat generation conditions. As shown in Figure 3, when the U-value of the exterior wall decreases from 0.54 W/(m²·K) to 0.137 W/(m²·K), the building thermal loads show a gradient decreasing trend (decreasing from 19.59% to 71.76%), while the cooling loads increase in tandem (increasing from 12.98% to 14.9%). This heat-cooling load trend led to a condition-sensitive characteristic of the total load: the total load increased by 4.81% in the 80 W/m² equipment heat dissipation condition, whereas it decreased by 6.14–17.91% in the other operating conditions. The adjustment of the roof U-value (0.556–0.199 W/(m²·K)) showed a similar but more significant thermodynamic response, with heat load reductions ranging from 19.41% to 69.94%, but a total load increase of 2.13% for the 80 W/m² condition.

The results of the study show that (1) although the U-value reduction can effectively inhibit heat loss, it will lead to a simultaneous increase in the cold load demand, forming a double-edged sword effect of the optimization of the thermal performance; (2) the total load response has a critical sensitivity to the strength of the internal heat source, and the enhanced thermal insulation leads to a rise in the total energy consumption instead when the equipment heat dissipation density is more than 80 W/m². When the equipment heat dissipation density is larger, the total load shows a non-linear relationship; (3) the roof U-value has a higher influence coefficient on the total load than that of the external wall by 18.7%, indicating that the thermal performance of the roof has a priority regulation value in the energy saving of the substation building.

The analysis demonstrates a positive correlation between higher external window U-value and increased building heating loads. Across five internal heat emission scenarios (25–80 W/m²), reducing the window U-value from 2.9 to 2.0 W/(m²·K) yielded progressive HL reductions of 6.20% to 28.60% (12.41–5.78 kWh/m²), while cooling loads exhibited moderate increases of 3.78% to 4.45% (0.41–7.14 kWh/m²). Notably, total load decreased with a lower window U-value in most scenarios, except under the 80 W/m² condition, where TL increased by 0.69% (−1.36 kWh/m²). Compared to walls and roofs, the influence of the window U-value on thermal loads is less pronounced due to the inherently low window-to-wall ratio in cold-region substations, where daylighting requirements are minimal, and opaque exterior wall dominate.

The ventilation rate (applied in weeks 18–45, corresponding to the summer cooling period) mainly affects the CL, with a sensitivity proportional to the intensity of internal heat dissipation. At higher equipment heat output (80 W/m²), increasing the ventilation rate reduces the CL by up to 47.97% (85.09 kWh/m²), showing a non-linear cooling efficiency gain. This trend diminishes at lower heat emission levels, highlighting ventilation as a key control variable for high heat density environments.

The TL of the building varies to different extents in response to changes in four key factors. At higher internal equipment heat emission levels (80 W and 65 W), alterations in the ventilation rate lead to greater absolute variations in TL compared to changes in the U of the building envelope. At moderate heat emission levels (50 W), the effects of changes in ventilation rate on TL are similar in magnitude to the impact of variations in the U-value of the exterior wall and roofs. However, at lower equipment heat emission levels (25 W and 40 W), changes in the U-value of the building envelope result in a more significant absolute variation in TL than changes in the ventilation rate.

4.2. Regression Equation Fitting

In single-factor analysis, revealing the relationships between variables is the core of the analysis. Fitting a regression equation, as a commonly used method, establishes a mathematical model between the independent and dependent variables, uncovering their correlation and quantifying the effect of the independent variable on the dependent variable. The coefficients in the regression equation provide the foundation for hypothesis testing and predictive analysis. Through equation fitting, this analysis explores the relationship between building heating and cooling loads and the related factors.

Based on the results of OFAT analyses, regression equations were fitted for the relationship between the different factors and the HL and CL indicators of the building. Load data for 50 W equipment heat generation condition were selected. The independent variables for the regression equation were chosen as follows: the U of the exterior walls, roof, and windows; ventilation rate; and equipment heat generation. The dependent variables were the HL and CL. The regression equations were fitted using the least squares method, and the coefficient of determination (R²) was calculated.

Table 2. Relationship between building heat load and contributing factors.

Fitting Results	R2	Independent Variable
$y_1=81.02x_1+55.66$	1	U-value of the EW
$y_1=99.34x_2+55.23$	1	U-value of the RF
$y_1=11.51x_3+66.36$	1	U-value of the OW
$y_1=-3.17x_5+258.69$	0.97	Equipment heat generation

shows the fitting results of the building’s heating load with various factors, and

Table 3. Relationship between building cooling load and contributing factors.

Fitting Results	R2	Independent Variable
$y_2=-21.28x_1+69.72$	1	U-value of the EW
$y_2=-20.81x_2+67.78$	0.98	U-value of the RF
$y_2=-2.88x_3+66.73$	1	U-value of the OW
$y_2=-4.25x_4+57.58$	0.66	Ventilation rate
$y_2=1.41x_4^2-18.33x_4+87.14$	0.92	Ventilation rate
$y_2=136e^{\left(-{\displaystyle \frac{x_4}{1.35}}\right)}+28.15$	0.99	Ventilation rate
$y_2=3.14x_5-82.88$	0.95	Equipment heat generation

shows the fitting results of the building’s cooling load with various factors.

It can be observed that the HL and CL of the building exhibit a significant linear relationship with the thermal transmittance of external walls, roof, and windows, as well as the amount of heat generated by equipment. However, the linear relationship between CL and ventilation rate is not strong. It is more suitable to use quadratic regression methods for fitting the CL–ventilation relationship, and applying an exponential decay function provides a better fit.

4.3. One-Factor Sensitivity Analysis

Calculate the dimensionless sensitivity coefficient for energy consumption subsequent to the application of various measures. Various design parameters yield distinct sensitivity coefficients for the building’s total energy consumption. The absolute value of the sensitivity coefficient indicates the extent of influence that a design parameter exerts on the corresponding output. A positive sensitivity coefficient signifies that an increase in the design parameter value will lead to an increase in the corresponding output value, and vice versa. The total energy consumption is calculated as the sum of HL and CL. Typically, in practical scenarios, the substation’s cooling-energy consumption is supplied by air conditioning and its heating-energy consumption by electric heaters. The efficiency of air conditioning in cold regions is set at 2.9 and that of electric heaters at 0.99 L for the calculation of ETL. The result is shown in Figure 5.

The results indicate that the greater the equipment heat generation, the more significant the impact of the thermal insulation performance of the building envelope on the heat load. The influence of various factors on the cooling load reaches its maximum under moderate equipment heat generation. The impact of all factors on the total load is most pronounced when the equipment heat generation is moderate. Excessive or insufficient equipment heat generation is detrimental to adjusting building loads. The sensitivity coefficients for EW and RF are close to each other, indicating that their influence on the load is comparable. This suggests that the substation is more affected by roof characteristics than high-rise buildings due to its low-rise structure. In terms of positive and negative values, the SC of VN on building loads is negative, indicating that the increase in ventilation rate brings about a reduction in the total building load in the study area, and the SC of ETL on the exterior wall U, roof U, and exterior window U are positive, indicating that reducing the value of the envelope U reduces the loads.

4.4. Multivariate Analysis

In practical engineering applications, the determination of thermal parameters for buildings is often not continuous. Instead, due to constraints related to material selection and engineering conditions, these parameters are typically represented by discrete specific values. These discrete values are then combined to form a range of design options. The use of orthogonal experimental design and multifactorial analysis of energy consumption in buildings provides a methodology for selecting optimal design solutions under discrete conditions. In this paper, five factors were selected to characterize the energy consumption characteristics of the building, each containing five levels, using the orthogonal design table L₂₅ (5⁵), with specific numerical settings shown in

Table 4. Factor settings for orthogonal design schemes.

Experiment Number	EW (W/m2·K)	RF (W/m2·K)	OW (W/m2·K)	VN (times/h)	EHG (W/m2)
1	0.556	0.199	2.7	10	50
2	0.443	0.199	2.0	6	80
3	0.443	0.347	2.7	2	65
4	0.556	0.269	2.5	8	80
5	0.137	0.199	2.3	4	25
6	0.345	0.447	2.9	6	50
7	0.345	0.540	2.7	4	80
8	0.252	0.199	2.9	8	65
9	0.137	0.269	2.0	2	50
10	0.252	0.447	2.3	2	80
11	0.345	0.347	2.0	8	5
12	0.345	0.269	2.3	10	65
13	0.556	0.540	2.9	2	25
14	0.443	0.447	2.5	10	25
15	0.252	0.540	2.0	10	40
16	0.556	0.447	2.0	4	65
17	0.252	0.347	2.5	4	50
18	0.345	0.199	2.5	2	40
19	0.137	0.447	2.7	8	40
20	0.443	0.540	2.3	8	50
21	0.443	0.269	2.9	4	40
22	0.252	0.269	2.7	6	25
23	0.556	0.347	2.3	6	40
24	0.137	0.540	2.5	6	65
25	0.137	0.347	2.9	10	80

.

At five factors and five levels, an orthogonal experiment was designed with the L25 table, encompassing 25 scenarios. Simulation calculations were conducted based on the parameters set for each condition, deriving the building’s HL and CL. Following data processing, TL was determined. Significant analysis of building-energy consumption factors was then performed using range and variance analysis. The results of the range analysis are in

Table 5. Orthogonal design range analysis results.

Item	EW	RF	OW	VN	EHG
k1	149.411	152.067	147.564	167.156	212.874
k2	162.196	165.271	151.688	145.293	144.775
k3	141.400	152.864	143.244	139.404	120.277
k4	153.552	135.491	147.855	125.629	115.762
k5	136.858	137.723	153.066	165.935	149.729
R	25.338	29.780	9.821	41.527	97.112

, and the variance analysis in

Table 6. Orthogonal design variance analysis results.

Project	Sum of Squares	Degrees of Freedom	Mean Square	F	Significance
revised model	41,767.684	20	2088.384	6.803	0.038
intercept	552,661.402	1	552,661.402	1800.257	0.000
EW	1998.180	4	499.545	1.627	0.324
RF	2991.293	4	747.823	2.436	0.205
OW	298.749	4	74.687	0.243	0.900
VN	6340.088	4	1585.022	5.163	0.070
EHG	30,139.375	4	7534.844	24.544	0.004
inaccuracies	1227.961	4	306.990

. The value in Table 5 is the calculated building load, and the unit is kWh/m². Table 6 is the variance statistical calculation result without unit.

According to the results of the range analysis, the U value represents the average value of total load under each level of each factor, R value represents the range under different levels of each factor, the minimum value of U value under the same factor represents the optimal level, and the ranking of R value represents the ranking of influence. The greater the F value, the stronger the explanatory ability of the factor to the dependent variable, and the significance reflects the significance of the impact of the factor on the load. According to the ranking results of range and significance, the influence of various factors on the total energy consumption in the building operation phase is ranked from large to small as equipment calorific value > ventilation rate > roof U value > exterior wall U value > exterior window U value. In the design of building-energy efficiency, the largest factor should be given priority. The optimal level is selected according to the minimum U value of each factor. The minimum energy consumption level corresponding to each factor is 0.137 W/m²·K for the exterior wall U value, 0.268 W/m²·K for the roof U value, 2.5 W/m²·K for the exterior windows, 4 times/h of ventilation, and a heat generation of 40 W/m².

According to the selected value of each factor under the above optimal level, the corresponding levels of each factor for minimum energy consumption can be determined. By adjusting the model input conditions accordingly, the minimum energy consumption results are obtained. The total annual HL for the building is 40.41 kWh/m², the total annual CL is 41.40 kWh/m², and the annual TL is 81.81 kWh/m². This represents a decrease of 129.20 kWh/m² (61.23%) compared to the maximum energy consumption of 211.01 kWh/m².

Considering ETL and using the same method to calculate energy consumption, the minimum energy consumption level is 0.137 W/m²·K for exterior wall, 0.269 W/m²·K for roof, 2.5 W/m²·K for exterior window, 4 times/h for ventilation, and 65 W/m² for equipment. Under these conditions, the annual ETL during building operation is 53.75 kWh/m², which is lower than the ETL of 55.08 kWh/m² observed under minimum TL conditions. This represents a reduction of 152.14 kWh/m² (73.90%) compared to the maximum ETL of 205.88 kWh/m². Based on the 2021 average carbon dioxide emission factor for the Northeast region, which is 0.6012 kgCO₂/kWh, an annual reduction of 91.47 kgCO₂ in emissions can be achieved.

The results indicate that to achieve lower cooling- and heating-energy consumption in the building, it is essential to manage indoor equipment heat generation as an influencing factor in a reasonable manner. This can be performed by selecting multiple rooms and implementing appropriate equipment arrangement strategies, controlling heat density, and using air-cooled or oil-cooled methods to transfer heat between the equipment and the heat dissipation room for a more efficient distribution. In the ventilation rate on the impact of energy consumption, a reasonable layout of the room to strengthen air circulation, make full use of wind pressure and natural cooling sources for cooling the room.

5. Conclusions

(1)

In cold regions, the operational cooling and heating load energy consumption of substation buildings is significantly higher compared to general public buildings, due to elevated requirements for cooling load influenced by internal EHG. The effects of individual factors under different EHG vary in their impact on building loads. As the U of the building envelope decreases, under conditions of lower internal EHG, the TL of the building decreases, whereas under higher internal EHG conditions, the TL increases. Specifically, at higher EHG levels (80 W and 65 W), changes in VN cause a greater absolute TL variation compared to changes in the building envelope’s U. For moderate heat generation (50 W), both changes caused a similar degree of change in the absolute value of TL. Under lower EHG conditions (25 W and 40 W), changes in the building envelope’s U lead to greater absolute TL variations compared to TL variations caused by changes in VN.

(2)

In the regression analysis of the impact of individual factors on building-heating and -cooling loads, there is a clear linear relationship between the variations in the U-value of the exterior walls, roof, and windows, as well as the heat generated by equipment and the building’s HL and CL. However, the linear relationship between the ventilation rate and the building’s operational load is less pronounced, making it more suitable for regression fitting using an exponential decay function. The effect of equipment heat generation on building load is greater than that of the thermal insulation properties of the building envelope. The TL of the building is highly influenced by VN at high equipment heat generation and highly influenced by envelope insulation at moderate equipment heat generation.

(3)

Orthogonal design was used to obtain results, which were then subjected to range analysis and variance analysis. The results indicate that the factors affecting BOEC, in order of significance, are EHG > VN > RF > EW > OW. The factor levels corresponding to the minimum values of TL and ETL are different. The minimum TL is 81.81 kWh/m², while the minimum ETL is 53.75 kWh/m², representing a reduction of 152.14 kWh/m² (73.90%) compared to the maximum ETL of 205.88 kWh/m². Therefore, in building-energy conservation, it is crucial to consider both the internal EHG and the efficiency levels of HVAC systems to design optimal energy consumption levels for different buildings.

This paper analyzes and discusses the analysis of single and multiple influencing factors on the cooling- and heating-energy consumption of substation buildings. However, the paper is limited to the selection of parameters with discrete distributions, and there are deficiencies in the analysis of continuous changes in parameters. It is worthwhile to explore how to analyze the effects of continuous changes in parameters on the cooling and heating loads of the buildings in future research and how to realize and construct the specific parameters in practice.