Construction and Case Analysis of a Comprehensive Evaluation System for Rural Building Energy Consumption from an Energy–Building–Behavior Composite Perspective

Abstract

A comprehensive evaluation system for rural building energy consumption from an innovative composite perspective was established, suitable for southwest of China. The index system was established by brainstorming and the Delphi method, the weights of the comprehensive evaluation model were calculated by the analytic network process (ANP) method, and the scoring criteria of all evaluation indexes were levelled based on fuzzy evaluation theory. The system model was verified by case analysis, in the countryside around Chengdu Second Circle. Taking into account the highest weight, lowest comprehensive score, and widest range of comprehensive scores, three key factors were identified, namely percentage of clean energy use, thermal performance of exterior walls, and implementation rate of energy-saving measures. The distribution of comprehensive indicators and evaluation factors had certain spatial distribution characteristics, and the overall spatial distribution was characteristically high in the southeast and low in the northwest. Finally, based on key factors and regional distribution characteristics, energy-saving measures are proposed from three aspects: increasing sunrooms, adding wall insulation layers, and standardizing air conditioning temperature settings.

1. Introduction

Environmental problems such as global warming, pollution, and extreme weather caused by energy consumption have attracted widespread attention from scholars. The whole life cycle of a building consumes a large amount of energy, making buildings among the largest components of energy consumption in today’s society. In the United States and Europe [1], the construction sector accounts for 39% and 40% of energy consumption, and 38% and 36% of carbon dioxide emissions, respectively. At the same time, it accounts for a quarter of China’s total emissions from energy consumption [2].

The most recent edition of the China Energy Statistics Yearbook (2022) reveals that per capita domestic energy consumption in rural areas has increased by over threefold, from 132 kg of standard coal (kgce) in 2000 to 529 kg of standard coal (kgce) in 2021. As rural energy consumption continues to rise, China has placed a significant focus on addressing the issue of rural energy consumption and carbon emissions. This is evidenced by the proposal to promote the energy-saving renovation of rural housing, the construction of green rural housing, and the use of clean energy sources [3].

In order to identify the causes of high energy consumption in buildings, scholars have conducted research into a number of factors that affect energy consumption in buildings [4,5]. These studies aim to identify factors that significantly impact energy consumption. They propose alternative building designs, such as window designs or choice of roofing materials [6,7], and characterize the impact of single factors such as wall thickness, external windows, solar chimneys, etc., on the building’s energy consumption in the context of building characterization [8,9,10,11]. Some scholars have studied the impact of environmental changes on building energy consumption [12]. This encompasses the impact of temperature fluctuations [13,14], and solar radiation [15]. Furthermore, scholars have examined the impact of facade geometry on visual comfort and energy consumption across four distinct climatic conditions in Iran [16]. They have also explored the impact of energy consumption behaviors and the building energy sector [17]. For instance, a statistical analysis of factors such as occupant behavior and awareness of energy efficiency identified three distinct behavioral types: proactive, intermediate, and careless. Subsequently, these behaviors have been subjected to analysis in order to ascertain their influence on the consumption of energy by buildings [18,19,20].

The previous descriptions have only focused on single factors that affect energy consumption, such as buildings, the environment, or energy usage behavior, and they have frequently failed to acknowledge that the phenomenon of building energy consumption is a complex dynamic system. A systematic analysis of energy consumption from a composite perspective was necessary [21], with the aim of elucidating the interactions among various influencing factors. The operational energy consumption of buildings has been recognized as a significant contributor to overall energy consumption [22]. To attain the goals of low energy consumption and reduced carbon emissions in rural housing, an evaluation of building energy consumption from the integrated energy–building–behavior” perspective was deemed necessary.

Meanwhile, in the past, the comprehensive evaluation of rural living environments and green buildings in China was predominantly applied to the eastern coastal areas and northern plain regions, leaving a significant gap in the comprehensive evaluation system specifically tailored for mountainous rural areas. To bridge this gap in the comprehensive evaluation of rural building energy consumption in southwest China, a scientific and reasonable comprehensive evaluation model of the region’s rural building energy consumption was constructed.

With this constructed model, on-site research and data analysis were carried out on the building energy consumption in 20 villages surrounding Chengdu, with the building energy consumption level of each village being accurately quantified. Moreover, the key factors influencing building energy consumption were thoroughly explored and identified. Based on the identified key factors, practical and feasible energy-saving renovation plans for rural buildings were devised. Through the comparison of building energy consumption data before and after the renovation, the effectiveness and practicality of the constructed comprehensive evaluation model and the energy-saving renovation plan were verified.

2. Research Process

2.1. Research Framework

Figure 1 illustrates the research framework. The influencing factors were sorted through the application of both brainstorming and a literature review. An evaluation index system for rural building energy consumption was established, and the weights of each index were determined using the Delphi method and the analytic network process (ANP). The scoring criteria for each index were determined through numerical simulation and fuzzy theory. Case selection was conducted for data collection. ArcGIS 10.7 software was employed to analyze the evaluation results and survey data, using spatial interpolation analysis, thereby deriving the spatial distribution pattern of the low carbon intensity (LCI) of building energy consumption. Based on the evaluation outcomes, targeted building energy efficiency renovation schemes are proposed.

2.2. Construction of Evaluation Models

The construction of the evaluation model was divided into six stages, as follows:

2.2.1. Indicator Factor Sorting

Given the diverse factors influencing rural building energy consumption, which necessitated a comprehensive analysis of multiple indicators, an evaluation index system framework was established through analysis of the literature. This framework comprised four levels: the target level, the criterion level, the sub-criterion level, and the factor level [23]. The target layer clarified the core objectives and the ultimate results expected to be achieved. The criterion layer was the benchmark or basis for evaluating or judging the degree of achievement of the target level. The sub-criterion layer was a further refinement of the criterion layer. The factor layer included the analysis of various factors that affected the achievement of goals. An innovative comprehensive evaluation metric, low carbon intensity (LCI), is proposed to be utilized for quantifying the level of energy consumption in buildings. Consequently, the target level was defined as LCI, and the criterion level was summarized into three aspects: cleanliness of energy (C), energy efficiency of buildings (E), and self-discipline of residents (S), based on the energy’s inherent attributes, the spatial carrier of energy usage, and the energy implementer. This led to the CES model being defined. The framework of the index system is illustrated in Figure 2.

The preliminary indicators were obtained through two rounds of brainstorming, where members brainstormed together at the same time to select the indicators. Subsequently, the indicator factors underwent optimization through two rounds of Delphi methodology. In the first round, focused primarily on open-ended consultations, the experts were presented with the preliminarily drafted indicators and their corresponding explanations. A total of 12 experts from relevant fields were invited to provide feedback on the evaluation system. The recognition rate for the criterion level and sub-criterion level of the evaluation index system achieved 100% among the experts.

The second round of expert consultation questionnaire adopted the Likert scale method, with five judgments for each indicator: “important”, “relatively important”, “average”, “less important”, and “not important”, assigned 5 points, 4 points, 3 points, 2 points, and 1 point, respectively. Initially, in order to improve data quality, the Kendall Wa coefficient was used to evaluate the degree of coordination between experts’ judgments on all indicators, which was particularly suitable for scenarios that required multiple evaluators to participate in the evaluation. This coefficient was used as a correlation measure to calculate the degree of correlation between multiple level variables, which could be used to select good works or good raters objectively. Subsequently, the weighted average score, weighted standard deviation, and weighted coefficient of variation for each indicator were calculated, serving as the basis for indicator screening. The calculation results consistently indicated a high degree of coordination and consensus among experts’ judgments, with the consistency test being successfully passed. Through these two rounds of expert consultation methods, a final evaluation index system comprising eight sub-criterion levels and 26 factor levels was established, with a distinction made between objective and subjective indicators for each. The evaluation index system is presented in the corresponding columns of

Table 1. Comprehensive evaluation index system of energy consumption of rural residential buildings.

Criterion Layer	Weight	Sub-Canonical Layer	Weight	Factor Layer	Weight	Normalized Values
						Q ∈ [80,100]		Q ∈ [60,80)		Q ∈ [40,60)		Q ∈ [20,40)		q < 20
Energy Cleanliness (C)	0.559	Energy Supply and Demand C1	0.071	Clean Energy Demand Satisfaction C11 (subjective)	0.041	Satisfaction with clean energy demand is high		Satisfaction with clean energy demand is relatively high		Satisfaction with clean energy demand is average		Satisfaction with clean energy demand is relatively low		Satisfaction with clean energy demand is low
				Energy Price Stability C12 (subjective)	0.016	Energy prices are stable		Energy prices are relatively stable		Energy prices vary in general		Energy prices are relatively highly volatile		Energy prices are highly volatile
				Energy Subsidies and Satisfaction C13 (subjective)	0.013	Residents are highly satisfied with energy subsidies		Residents are satisfied with energy subsidies		Residents’ satisfaction with energy subsidies is average		Residents’ satisfaction with energy subsidies is relatively low		Residents’ satisfaction with energy subsidies is low
		Energy Use C2	0.285	Electricity Consumption per capita C21	0.064	Electricity consumption Q ∈ [646,727]		Electricity consumption Q ∈ (727,808]		Electricity consumption Q ∈ (808,889]		Electricity consumption Q ∈ (889,970]		Electricity consumption Q ∈ (970,1051]
				Gas Consumption per capita C22	0.041	Gas consumption G ∈ [72,81]		Gas consumption G ∈ (81,90]		Gas consumption G ∈ (90,99]		Gas consumption G ∈ (99,108]		Gas consumption G ∈ (108,117]
				Proportion of Energy Use from Commodities C23	0.074	Commodity energy use/total energy × 100 per cent
				Percentage of Clean Energy Use C24	0.105	Total clean energy usage/total energy usage × 100 per cent
		Energy Sustainability C3	0.204	Biomass Energy Utilization C31	0.102	Meets the requirements of biogas digester on-site use and has a high frequency of use	Meets the requirements of biogas digester on-site use and the frequency of use is average		Meets the requirements of biogas digester on-site use and is used less frequently		Does not meet the requirements for use or does not use modern biomass energy		Conventional biomass energy is used
				Solar Energy Systems C32	0.102	30 points for solar thermal equipment, 30 points for solar photovoltaic equipment, 20 points for setting up a sunshine room, and the cumulative score is calculated.
Building Energy Efficiency (E)	0.297	Architectural Design E1	0.098	Building Site Selection E11	0.043	According to the definition of the rationality of building site selection with relevant specifications, five main conditions are established to determine the evaluation criteria for building site selection based on the number of buildings.
						5 conditions are met	4 conditions are met		3 conditions are met		2 conditions are met		0–1 conditions are met
				Building Orientation E12	0.016	The growth rate of energy consumption is 0 per cent–3 per cent, corresponding to the direction.	The growth rate of energy consumption is 3 per cent–6 per cent, corresponding to the direction.		The growth rate of energy consumption is 6 per cent–9 per cent, corresponding to the direction.		The energy consumption growth rate of 9 per cent–12 per cent, corresponding to the direction.		The energy consumption growth rate is greater than 12 per cent, corresponding to the direction.
				Architectural Space Layout E13	0.025	Floor height 2.7 ≤ h ≤ 3.0	Floor height 3.0 < h ≤ 3.3		loor height 3.0 < h ≤ 3.3		Floor height 3.6 < h ≤ 3.9		Floor height 3.9 < h ≤ 4.2
				Building form Factor E14	0.013	0.35 ≤ Tx ≤ 0.45	0.45 < Tx ≤ 0.55		0.55 < Tx ≤ 0.75		0.75 < Tx ≤ 0.95		0.95 < Tx ≤ 1.2
		Envelope Structure E2	0.131	Thermal Performance of Exterior Walls E21	0.041	0.6 ≤ Km ≤ 1.0	1.0 < Km ≤ 1.4		1.4 < Km ≤ 1.8		1.8 < Km ≤ 2.2		2.2 < Km ≤ 2.6
				Thermal Performance of Exterior Windows E22	0.041	1.4 ≤ Kw ≤ 2.4	2.4 < Kw ≤ 3.4		3.4 < Kw ≤ 4.4		4.4 < Kw ≤ 5.4		5.4 < Kw ≤ 6.4
				Thermal Performance of Roofing E23	0.022	0.8 ≤ Kr ≤ 1.4	1.4 < Kr ≤ 2.0		2.0 < Kr ≤ 2.6		2.6 < Kr ≤ 3.2		3.2 < Kr ≤ 4.0
				External Shading Measures E24	0.027	2.0 ≤ L ≤ 2.7	1.5 < L ≤ 2.0		1.0 < L ≤ 1.5		0.5 < L ≤ 1.0		0 < L ≤ 0.5
		Building Material E3	0.068	Building Materials Localization Ratio E31	0.026	City-wide use of building materials/total use of building materials × 100 per cent
				Utilization Rate of Environmentally Friendly Construction Materials E32	0.042	Green building materials used/total building materials used × 100 per cent
Residents’ Self-discipline (S)	0.144	Awareness Management S1	0.042	Widespread Awareness of Low Carbon S11 (subjective)	0.016	Residents have a high level of low-carbon knowledge	Residents have a relatively high level of low-carbon knowledge		Residents’ low-carbon knowledge is average		Residents’ understanding of low-carbon knowledge is relatively low		Residents’ low-carbon knowledge is low
				Acceptance of Low-Carbon Living S12 (subjective)	0.011	Low-carbon lifestyles mainly involve green consumption, food conservation, residential energy-saving renovation, energy-saving household appliances, garbage classification, and clean travel
						Meets 5–6 items	Meets 4 items		Meets 3 items		Meets 2 items		Meets 0–1 items
				Responsiveness to Low-Carbon Construction S13 (subjective)	0.015	The village residents are supportive of infrastructure construction	Residents are in favor of the development of rural infrastructure and hardware		Residents generally support the construction of rural infrastructure and hardware		The construction of rural infrastructure is less supported by residents		Residents do not support the construction of rural infrastructure
		Behavior Management S2	0.101	Proportion of Equipment Designed to Save Energy S21	0.036	Number of energy-saving devices in the dwelling/Total number of devices in the dwelling × 100 per cent
				Implementation Rate of Energy-saving Measures S22	0.037	The number of energy-saving behaviors achieved by residents/10 × 100 per cent
				Waste Recycling S23 (subjective)	0.015	Utilize household waste to its full potential	A significant proportion of household waste is utilised.		Household waste is partly utilised		A small quantity of domestic waste is utilised		Household waste is not utilised
				Indoor Air Quality Discipline S24	0.013	The cumulative score is calculated by assigning 30 points for indoor planting of green plants, 30 points for indoor air purifiers, and 20 points for window ventilation.

.

2.2.2. Construction of a Model for the Mutual Influence Relationship between Indicators

The analytic network process (ANP) method is a multi-criteria decision analysis approach. The ANP method can consider the influence of the interrelationships between indicators on the weight size. Therefore, in order to eliminate the impact of the interrelationships between indicators on weight allocation and make the calculation more reasonable and scientific, the ANP method was adopted. The control layer encompassed indicators of the target level and the criterion level, whereas the network layer comprised indicators of the sub-criterion level and the factor level. The interdependent relationships among these indicators were determined through expert consultations and questionnaire surveys. A total of 12 experts were consulted via questionnaires, and when the number of experts who perceived a correlation between two indicators was equal to or greater than one, it was determined that there existed an influencing relationship between those two indicators; otherwise, no influencing relationship was assumed. Based on the dependencies and feedback relationships among the indicators, a network structure model diagram (Figure 3) was constructed using Super Decisions (yaanp 2.7) software [24]. In the ANP structure, the comprehensive evaluation of building energy consumption (energy, building, and residents) comprised the control layer, whereas the assessment factors under each criterion comprised the network layer.

2.2.3. Construction of Judgment Matrix

Based on the indicator network hierarchy diagram, judgment matrices were formulated to evaluate the superiority and inferiority of indicators at both the control layer and the network layer. With considerations given to feasibility and representativeness, six experts were selected from the previously mentioned twelve, who were highly relevant in the field of human settlement environments, to ensure their corresponding professional competence and credibility. These experts were then tasked with conducting pairwise comparisons of factor importance using Saaty’s 1–9 scale, where 1 indicates that two elements are equally important, 9 indicates that one element is much more important than the other, and the middle numbers represent the degree of relative importance. Among them, four experts self-assessed their familiarity with the indicators as 1 (fully familiar), while the remaining two assessed their familiarity as 0.75 (moderately familiar). Subsequently, the judgments on superiority/inferiority were weighted and averaged according to Equation (1), where Q_i represents the weighted average superiority/inferiority value, Cs_i denotes the expert’s self-assessed familiarity with the study, and n refers to the total number of experts consulted.

$$Q_i={\displaystyle \frac{{Cs}_i}{{\sum }_{i=1}^{i=n}{Cs}_i}}$$

(1)

The construction of a judgment matrix was carried out by using indicators at higher levels or those affecting other factors as criteria for the judgment matrix (located in the upper left corner), with the indicators belonging to its lower levels or affected by others being arranged in the first row and first column. The criteria layer indicators were independent of each other, and a cluster judgment matrix was established based on the degree of direct superiority or inferiority. There existed a mutually influential relationship between the sub-criteria layer indicators, necessitating not only the construction of a node judgment matrix between the sub-criteria layer indicators but also the comparison of each sub-criteria layer indicator with its associated sub-criteria layer indicators. The factor layer indicator initially required the construction of a node judgment matrix between the factor layer elements under each sub-criterion layer indicator. Secondly, based on the correlation between elements, it was necessary to construct a comparison of the dominance between elements that had mutual influence on a certain element when it was used as a criterion.

Table 2. Energy supply and demand cluster judgement matrix under the criterion layer.

Energy Supply and Demand	Architectural Design	Envelope Structure	Energy Use	Energy Sustainable	Awareness Management	Behavior Management
Architectural design	1	2	1/2	1/2	3	2
Envelope structure	1/2	1	1/2	1/2	2	2
Energy use	2	2	1	1	2	2
Energy sustainability	2	2	1	1	3	2
Awareness management	1/3	1/2	1/2	1/3	1	1
Behavior management	1/2	1/2	1/2	1/2	1	1
consistency test: λmax: 6.174113; CR = 0.0276 < 0.1

displays only the cluster judgment matrix under the indicators of the energy supply and demand criteria layer. Owing to the numerous indicators and complex correlations, in addition to Table 2, there were seven cluster judgment matrices and 87 node judgment matrices. The final judgment matrix comprised 95 matrices. Due to limited space, only a portion of it has been presented.

2.2.4. Consistency Check of Judgment Matrix

Through consistency testing, the logical consistency of the decision maker’s input in the judgment matrix was verified. In the ANP method, the maximum eigenvalue (λ_max), CI (consistency index), and CR (consistency ratio) are important parameters for evaluating the consistency of a judgment matrix.

The maximum eigenvalue represents the maximum scaling ratio of a matrix during the transformation process, usually calculated using computer programs or mathematical software. In ANP, the maximum eigenvalue is used to calculate the CI value, which is a quantitative indicator used to measure the consistency of the judgment matrix. The calculation formula for CI is shown in Equation (2), where n is the order of the judgment matrix (representing the number of factors). The smaller the CI value, the better the consistency of the judgment matrix, indicating that the decision makers’ judgments tended to be consistent.

To eliminate the influence of matrix order on CI values, the CR metric was introduced. The CR calculation formula is shown in (2.3), where RI is related to n and can be obtained by consulting a table. The smaller the CR value, the better is the consistency of the judgment matrix. When CR ≤ 0.1, it was determined that the judgment matrix had passed the consistency check. According to the calculations that were performed, the consistency ratio for the judgment matrix presented in Table 1 was found to be 0.0276, which was less than 0.1 (similarly, all judgment matrixes had their consistency ratios determined to be less than 0.1).

$$CI={\displaystyle \frac{{\lambda }_{max}-n}{n-1}}$$

(2)

$$CR={\displaystyle \frac{CI}{RI}}$$

(3)

2.2.5. Calculation of Indicator Weights

After the expert scoring results for the judgment matrices were obtained and their consistency was verified, the data for the matrices was entered into yannp 2.7 software [25]. The input was facilitated through a questionnaire format, which captured the individual experts’ scoring results for the judgment matrices. Subsequently, the judgment outcomes of superiority and inferiority for eight cluster judgment matrices and 87 node judgment matrices were derived. By selecting the appropriate options within the yannp 2.7 software, the unweighted supermatrix, weighted supermatrix, and limit supermatrix were calculated. The final determined weight results are presented in the corresponding weight columns of Table 1.

2.2.6. Index Classification Criteria and Determination of LCI

We used linear interpolation combined with numerical simulation, national or local standards, current regulations, and statistical yearbooks to score objective indicators. The data sources included field measurements and observations, while questionnaires were used to collect information and record the subjective opinions of respondents. Data sources encompassed field measurements and observations, while questionnaires were utilized to gather information and document the subjective perceptions of respondents. Subsequently, the subjective indicators were quantified through the application of fuzzy mathematics theory. Using “low”, “relatively low”, “average”, “relatively high”, and “high” to indicate the degree of evaluation of the indicators by the research subjects, the evaluation opinions were scored 20 points, 40 points, 60 points, 80 points, or 100 points respectively. The standardized values corresponding to the scoring criteria for specific indicators are presented in Table 1.

LCI was used to measure energy consumption levels and has a certain functional relationship with the indicators of the three criteria layers from a composite perspective:

$$\mathrm{L}\mathrm{C}\mathrm{I}=F\left(C,E,S\right)$$

(4)

The correlation between the three criterion levels of C, E, and S and the sub-criterion levels is described by Equation (4). Equation (4) can be written as $\mathrm{L}\mathrm{C}\mathrm{I}=F\left(f_C\right(C_i),f_E(E_i),f_S(S_i\left)\right)$, where C_i, E_i, and S_i represent the indicators of the sub-criteria layer, and i represents the number of indicators of the sub-criteria layer. The process of pushing secondary functions is as follows:

$$C=f_C\left(C_i\right)=\sum _{i=1}^{i=3}{w_{Ci}C}_i=\sum _{i=1,j=1}^{i=3,j=m}{w}_{ij}{C}_{ij}$$

(5)

$$E=f_E\left(E_i\right)=\sum _{i=1}^{i=3}{w_{Ei}E}_i=\sum _{i=1,j=1}^{i=3,j=n}{w}_{ij}{E}_{ij}$$

(6)

$$S=f_S\left(S_i\right)=\sum _{i=1}^{i=3}{w_{Si}S}_i=\sum _{i=1,j=1}^{i=3,j=l}{w}_{ij}{S}_{ij}$$

(7)

The weight coefficients of each indicator are represented by “w”. “j” represents the number of factor levels, while “m”, “n”, and “l” represent the number of indicators in the sub-criteria level. Equation (8) shows the functional relationship between LCI and the criterion layer, where 0.559, 0.297, and 0.144 are the indicator weight coefficients of the criterion layer, obtained from the weight calculation method described earlier.

$$\mathrm{L}\mathrm{C}\mathrm{I}=0.559C+0.297E+0.144S$$

(8)

2.3. Application of Evaluation Model

As the center of the southwest region, the investment in energy infrastructure in Chengdu significantly exceeds that in villages in other areas. Although this demonstrates the achievements in rural development made in recent years, it also highlights the relatively weak energy infrastructure in rural areas. The second-tier districts within a 30 km radius to the east, south, west, and north of Chengdu’s main urban area, including Pidu District, Xindu District, Longquanyi District, Shuangliu District, and Wenjiang District, were selected through a multi-stage stratified sampling method, ensuring the objectivity of the research subjects. A total of 20 villages, 6 designated as demonstration villages and 14 as ordinary villages, were selected as samples. The specific locations of the households that were sampled are depicted in Figure 4. A total of 550 households were surveyed, yielding 521 valid samples, with an effective questionnaire rate of 94.73%.

The data were collected through on-site observation, measurement, and a questionnaire survey. On-site observation and collection of data included indicators such as envelope structure (E2) and building materials (E3) (materials and structures of walls, doors and windows, shading, roofs, etc.). On-site measurements were conducted to collect data on the indicators of architectural design (E1), such as building orientation, building depth, and floor height. The questionnaire survey collected data on indicators such as energy supply and demand (C1), energy use (C2), energy sustainability (C3), awareness management (S1), and behavior management (S2) (household energy consumption structure, energy supply and demand satisfaction, various types of energy consumption, etc.). Figure 5 shows the specific research process and the tools used for on−site measurement.

3. Results

By integrating the research data with indicator grading standards, factor scores were obtained, and the comprehensive evaluation values for the indicators of both the criterion layer and the target layer were calculated using Equations (5)–(8). The results were subsequently derived as follows.

3.1. Energy Cleanliness (C) Sub-Evaluation Results

Energy cleanliness (C) encompassed nine factors, including satisfaction with clean energy demand (C11) and several others. The bar chart of the weights of each indicator arranged in descending order is shown in Figure 6; among them, the weight of C24 indicator is the most prominent. Although C31 and C32 indicators also play an important role in weight ranking, their actual scores were significantly lower than C24 (Figure 7). According to field investigations, the low utilization rate of C31 and C32 indicators in rural households is mainly due to high implementation costs and insufficient technological maturity. Therefore, when considering both weight and actual score, the C24 indicator was more critical.

Statistical analysis was conducted on the energy consumption of various types of buildings during the operation within the surveyed sample households. We calculated the percentage of various energy sources in relation to the total energy consumption, as shown in Figure 8. Electricity, firewood, liquefied petroleum gas, natural gas, solar energy, and biogas accounted for 39.29%, 30.30%, 18.77%, 8.13%, 2.53%, and 0.98%, respectively. The research data revealed that the natural gas penetration rate in the Shuangliu area was high, achieving a clean energy proportion of up to 75.30%. The C24 index was assigned an LCI score of 80.99. Conversely, in the Pidu area, the natural gas infrastructure was relatively underdeveloped, leading to a lower proportion of clean energy usage at 62.30%, which in turn resulted in the C24 index being awarded a lower LCI score of 64.96.

The rural residents exhibited a lower utilization rate of solar energy and biogas. On-site investigations and statistical analyses were conducted on the reasons for not utilizing solar energy and biogas. The main reason why residents did not use solar energy was that the intensity of solar radiation may not meet the demand for hot water supply. Some held the belief that there was a shortage of raw materials for biogas production, whereas others argued that the availability of liquefied petroleum gas, natural gas, and other alternatives rendered the use of biogas unnecessary.

To facilitate a deeper understanding of the geographical changes in energy consumption patterns in rural areas surrounding Chengdu, the spatial distribution characteristics of LCI scores for the C index are visually depicted in Figure 9. The map shows a significant decreasing trend in LCI scores from southeast to northwest, indicating that the impact of energy utilization on the environment varied. The map exhibits a pronounced downward trend in LCI scores from the southeast to the northwest regions, indicating varying levels of environmental impact stemming from energy utilization.

Huaguo Village, located in Longquanyi District, stood out with an LCI score of 77.14, which could be attributed to several factors. Among them, the strategic location of the village within the Longquan Mountain Range, where abundant solar radiation is received, played a part. Additionally, the household energy structure in Longquanyi District was reported to be highly diversified and environmentally friendly, with natural gas accounting for a significant proportion (up to 80%) of energy consumption. This reduction in dependence on fossil fuels, in turn, contributed to a cleaner environment.

On the other hand, Jinbai Village in Pidu District received an LCI score of merely 56.37, the lowest recorded in the analysis. This relatively low score stemmed from the challenges faced by the area in developing natural gas infrastructure. Households in Pidu District rely heavily on firewood to meet their energy needs, a traditional fuel source that, despite its abundance, has significant impacts on air pollution and deforestation. Consequently, the C index for the region reflected a lower energy–environmental performance, with an LCI score of 66.56.

3.2. Building Energy Efficiency (E) Sub-Svaluation Results

The building energy efficiency (E) index included multiple factors. Differences were present in the overall LCI scores, with each factor having been assigned a unique score. Figure 9 shows the LCI scores for each factor, with the scores exhibiting marked diversity, spanning from 56.24 to 75.97, and averaging 65.78. As discussed previously, the score for the C24 indicator was notably high. Notably, the LCI score for the thermal performance of exterior walls (E21) index was particularly singled out, ranking last among all factors with a score of 56.24, which was 14.5% lower than the average score of 65.78. This low scoring not only underscored the shortcomings of external wall thermal performance under the prevailing evaluation framework but also pointed to the obstacles encountered in terms of energy conservation, emission reduction, and enhancing building energy efficiency.

The exterior walls of rural buildings in the surroundings of Chengdu encompassed clay solid brick walls, sintered hollow brick walls, sintered porous brick walls, and concrete hollow blocks, each having distinct thermal conductivity coefficients of 1.89 W/(m·K), 0.63 W/(m·K), 1.26 W/(m·K), and 0.315 W/(m·K), respectively. Based on the benchmark model, the physical parameters, boundary conditions, and initial conditions of the wall were set in DeST-h 2.0 software, and the temperature distribution of the wall and the periodic fluctuations in indoor temperature were calculated. After the simulation was completed, the indoor temperature fluctuation data were extracted and plotted into a chart (Figure 10). Notably, concrete hollow blocks, due to their minimal thermal conductivity, restricted indoor air heat dissipation, resulting in a more pronounced decrease in room temperature. In contrast, solid clay bricks, possessing the highest thermal conductivity, facilitated greater heat dissipation from indoor air.

Figure 11 shows the proportion of primary material for exterior walls in rural buildings around Chengdu. The findings had revealed that only 6.3% of those rural buildings had employed hollow concrete blocks for their exterior walls, whereas the proportion of clay solid brick walls stood as high as 40.9%. Consequently, the overall LCI score of the E21 indicator in rural Chengdu was merely 56.24. Pidu District stood out as a notable example, where the majority of buildings in Jinbai Village had been constructed over an extended period by villagers themselves. As a result of the high thermal conductivity of these exterior walls and their tendency to absorb more internal heat, the LCI score of the E21 index in this area was low, reaching only 42.19.

As depicted in Figure 12, the spatial distribution of LCI scores for the E indicator showed a notable decline from the southwestern regions towards the northeast. Specifically, the pinnacle of 81.35 for the E indicator’s LCI score was achieved in Liyuan Demonstration Village, located within Shuangliu District. In stark contrast, the lowest score of 51.79 was recorded in Yituan Village, situated in Xindu District. This disparity primarily stemmed from the fact that some villages in Xindu had endured protracted construction periods, coupled with a severe degradation of their overall architectural integrity. Notably, in these villages, there was a prevalence of solid brick walls, known for their inferior insulation properties, and outer windows constructed from either single-layer plastic steel or wooden materials, both of which contributed significantly to heat loss. Consequently, the average LCI score for the E index in Xindu District hovered at a mere 59.00, underscoring the urgency to enhance energy efficiency.

In stark juxtaposition, Liyuan Village in Shuangliu District stood out as a beacon of sustainability. The local government had embraced a proactive approach, having integrated greening and comfort considerations into every facet of building design, construction, and operation. This holistic methodology yielded a commendable LCI score of 68.89 for the E index within the region, attesting to the realized potential for sustainable development and improved energy performance in rural areas.

3.3. Residents’ Self-Discipline (S) Sub-Evaluation Results

The residents’ self-discipline (S) index, which pertains to residents’ implementation behavior, encompasses factors related to their awareness of and attitudes towards energy conservation. The maximum and minimum LCI scores for each factor, as shown in Figure 13, exhibited a marked difference. This disparity manifested not only in the span of scores but also in the varying effectiveness of factors, with factors exhibiting differing capabilities in carrying out energy-saving and emission reduction measures and advancing environmental sustainability. Overall, the broad range between the maximum and minimum values underscored that while some households had attained remarkable achievements in energy conservation and emission reduction, others presented considerable opportunities for improvement.

Subsequently, specific attention was turned to the implementation rate of energy-saving measures (S22) indicator, whose extreme value difference was highlighted as the most notable and pronounced among all factors. The S22 index attained its maximum score of 83.43 points in Liyuan Village, Shuangliu District, which not only significantly surpassed the average by 16.10 percentage points but also underscored the outstanding performance of energy-saving practices in the village. Conversely, the minimum score of 55.02 points for the S22 index, recorded in Renyi Village, Pidu District, fell considerably below the average by 23.43 percentage points, thereby revealing a clear deficit in the village’s energy-saving awareness at that time.

To delve deeper into the execution of energy-saving practices, a meticulous record was kept of residents’ air conditioning temperature preferences during the scorching summer months. As illustrated in Figure 14, a significant 77% of residents opted to maintain their air conditioning settings within the range of 21 °C to 26 °C, while an even higher percentage of 84% of households kept the temperature below 26 °C. However, with 7% of households having set their air conditioners to a frigid temperature below 20 °C, which indicated a potential disregard for energy efficiency, this aspect was noteworthy. The fact that a modest increase of 1 °C in the set temperature of a household air conditioner can lead to energy savings of 8% to 12% underscores the importance of mindful temperature settings.

Evidently, rural residents surrounding Chengdu tended to set their air conditioning temperatures too low, reflecting a lack of energy-saving awareness. The LCI score for the implementation rate of energy-saving measures (S22) indicator in Jinbai Village, Pidu District, stood at 58.34. This figure served as a stark reminder that significant energy-saving potential could have been unlocked through optimizing air conditioning temperature settings. It underscored the need for targeted interventions and promotional activities aimed at fostering a stronger energy-saving mindset within the local community.

Figure 15 reveals regional differences and spatial distribution trends of LCI scores under the indicator of residents self-discipline (S) index. Specifically, a decreasing trend in the LCI score of this indicator was observed from southwest to northeast. A high LCI score of 84.67 was attained in Gaoshan Village, Wenjiang District, which demonstrated the region’s outstanding performance in energy conservation, emission reduction, and the promotion of low-carbon living. Conversely, in Jinbai Village, Pidu District, a sharp drop in this value to 59.59 was seen, reflecting the apparent shortcomings in the adoption of low-carbon living practices within the region in the past.

Further analysis was conducted, revealing the aging of equipment and the prolonged use of traditional wood stoves in areas like Jinbai Village and Pidu to be phenomena that directly contributed to a low proportion of energy-saving equipment being utilized. Additionally, as a result of many residents in the area having relocated from other places in the past, the concept of low-carbon living may not have been fully embraced and internalized during their adaptation to the new environment. Consequently, in their daily routines, particularly for cooking and hot water supply, there had remained a strong reliance on traditional energy sources such as firewood, which undoubtedly led to increased carbon emissions and underscored the weak carbon awareness of the residents at that time.

3.4. Results of the Comprehensive LCI Evaluation of Energy Consumption in Rural Buildings

Figure 16 presents the comprehensive evaluation results of LCI scores for energy consumption in rural buildings surrounding Chengdu. The average variation among the LCI scores of different districts was found to be minor, yet the extreme disparities in scores among individual buildings within each district were significantly pronounced. Longquanyi District was noted as having the highest average LCI score, at 75.19, with scores ranging from a low of 56.62 for a building in Lianhe Village to a high of 86.92 for a building in Baosheng Village. Conversely, Pidu District was observed to have the lowest average LCI score, at 66.54, while scores varied from 44.63 for a building in Jinbai Village, the lowest within the district, to 87.66 for a building in Qinjiamiao Village, the highest.

From a spatial distribution perspective, the overall pattern was characterized by high scores being concentrated in the southeast, lower scores in the northwest, and moderate scores averaging out in the central region. This distribution was found to have been significantly influenced by key indicators such as C24, E21, and S22, which had played pivotal roles in shaping the LCI scores and their spatial distribution across rural buildings in the vicinity of Chengdu in the past.

Based on the actual situation, the 20 sample villages were classified into different levels according to their LCI scores (

Table 3. The carbon level of each sample village was comprehensively assessed.

Carbon Level	LCI Score	Name of the Village
Low carbon	[80,100]	Baosheng Village, Huaguo Village, Qinjiamiao Village, Gaoshan Village, Liyuan New Village.
Medium–low carbon	[70,80)	Satellite Village, Gonghe Village, Shuangyi Village, Helin Village, Mitsui Village, Sanxin Village.
Medium carbon	[60,70)	Tiangong Village, Yongning Village, Renyi Village, Jingshan Village.
Medium–high carbon	[50,60)	Lianhe Village, Wuyi Village, Jinbai Village, Liyi Village, Huoshiyan Village.
High carbon	[0,50)	None.

). From Figure 16a, it can be observed that the LCI scores of the surveyed 510 sampled households were mainly concentrated between 50 and 80, with not many sample families scoring outside this interval. At the same time, experts in relevant fields were consulted and scores divided within the range of 50–80 into 10 points per interval. The energy consumption level of the sample villages was between low carbon and medium–high carbon, and no villages were classified as high carbon. This trend highlights the significant progress made by rural communities around Chengdu in promoting low-carbon construction.

Among the low-carbon villages, four out of five villages were designated as demonstration villages, highlighting their exemplary status. These demonstration villages exhibited superior building performance and a diversified energy consumption portfolio, characterized by a heavy reliance on clean energy sources for daily consumption. Consequently, their LCI levels surpassed those of the non-demonstration villages. However, an exception to this pattern was Huaguo Village, whose elevated LCI level stemmed from factors distinct from those of the demonstration villages. Specifically, Huaguo Village has benefited from its tourism-driven development, government-led infrastructural renovations, and a unique environmental context characterized by high altitudes, intense solar radiation, and widespread adoption of renewable energy sources.

When considering the medium-to-high carbon villages, a pattern emerged in the form of three recurring challenges: firstly, the suboptimal utilization of clean energy resources; secondly, the inadequacy of thermal insulation and performance of building envelope structures; and thirdly, the general lack of awareness and adoption of low-carbon behaviors among residents. Addressing these issues holds the key to further advancing low-carbon development in these villages and fostering a more sustainable future for rural communities in the Chengdu region.

4. Recommendations

Based on the actual situation in the southwest region, solutions for the energy-saving renovation of buildings in rural areas surrounding Chengdu were explored. Following an evaluation of building energy consumption, the energy consumption issues present in certain buildings were identified. Adhering to the principles of open sourcing and cost savings, renovation plans for building energy use were investigated, integrating considerations of economic applicability, environmental friendliness, and social sustainability. The specific content of these plans is outlined as follows:

4.1. Transformation of Energy Efficiency

In terms of the factor of the C index, the rural areas surrounding Chengdu had a diverse energy structure, yet the utilization rate of clean energy in these areas was not high. To facilitate the comprehensive exploitation of renewable resources like solar energy within the Chengdu region, the adoption of additional installation of solar houses was employed to realize the application of passive solar energy technology, with the aim of solar power being harnessed.

Taking the example of a building in Jinbai Village, Pidu District, its low clean energy utilization rate had led to a suboptimal LCI score for the C24 indicator in the past. Upon the addition of solar rooms with varying depths, an energy consumption simulation was performed on the building, as depicted in Figure 17. The results indicated that, as the depth of the solar room increased, the cumulative heat and cooling loads of the building were also found to increase. When the depth of the solar room was optimized at 1 m, the building was found to achieve the highest overall energy efficiency, with a cumulative total load of 151.43 kW·h/m² throughout the year and an energy efficiency rate of 14%.

Based on the prevailing conditions in the rural areas surrounding Chengdu at that time, the recommendation was made to set the depth of solar rooms, also known as sun houses, between 1 m and 1.5 m. When the depth of the solar room in the building was set to 1.2 m, the standardized score of the C24 index was observed to increase, resulting in an elevation of the LCI score from 55.58 to 70.21, marking a significant improvement in energy efficiency.

4.2. Transformation of Energy Carriers

The factor of E index was most notably plagued by the poor thermal performance of the exterior wall. As previously mentioned, the exterior wall of a building in Jinbai Village, Pidu District, which was constructed of solid clay bricks, underwent a change in its construction method. Using the benchmark building as the model, keeping other parameters unchanged but changing the exterior wall structure, DeST-h 2.0 software was used to simulate the energy consumption per unit area of the building. Simulation results were obtrained for various types of exterior walls of benchmark rural houses before and after energy-saving renovation (Figure 18).

For a 240 mm clay solid brick wall, the addition of a 20 mm thick extruded polystyrene board resulted in an energy-saving rate of 15.14% being achieved. Similarly, the addition of a 15 mm extruded polystyrene board to sintered porous bricks led to an energy-saving rate of 10.1%. However, for two different types of exterior walls, increasing the thickness of the insulation layer to 30 mm only marginally improved energy efficiency by 0.22% and 0.69%, respectively. Consequently, the thermal performance of the building’s exterior wall was improved by the passive addition of 20 mm thick extruded polystyrene board insulation material, which in turn, elevated the standardized value of the E21 index. This enhancement subsequently raised the LCI score of its E21 indicator from 42.19 to 61.24.

4.3. Implement Behavioral Guidance

Regarding the factor of the S index, residents were found to possess weak low-carbon awareness. To enhance the LCI score of the S22 indicator and steer residents towards setting appropriate air conditioning temperatures, DeST-h 2.0 software was utilized to simulate and analyze the impact of varying air conditioning usage behaviors on building energy consumption. The outcomes of this simulation are presented in Figure 19.

The energy consumption was found to be positively correlated with the set temperature of the air conditioning, with an increase of about 10% in the energy-saving rate for every 1 °C decrease in the air conditioning temperature. Survey data revealed that 70% of residents had their air conditioning temperatures set below 26 °C, indicating a lack of standardization in air conditioning temperature settings. The optimal temperature setting for air conditioning in Chengdu was identified as 26 °C. Subsequently, the energy consumption habits of a household in Jinbai Village, Pidu District, were standardized, resulting in an increase in the standardized score of the S22 index. Consequently, the LCI score was elevated from 58.34 to 72.36, significantly reducing the total energy consumption of rural buildings in the vicinity of Chengdu, a milestone achievement with profound implications for energy conservation.

5. Conclusions

A rural building energy consumption LCI evaluation model was constructed from the perspective of energy building behavior. The evaluation model was applied to 20 sample villages in the surrounding areas of Chengdu to verify its feasibility. The key factors affecting rural building energy consumption in Chengdu were identified, providing improvement solutions for energy conservation and emissions reductions in the region. Consequently, the aforementioned considerations led to the following key conclusions being drawn:

(1)

The LCI of rural building energy consumption was found to be influenced by a multitude of key factors. The percentage of clean energy use (C24), the thermal performance of exterior walls (E21), and the implementation rate of energy-saving measures (S22) were identified as the primary factors affecting the energy consumption of rural buildings in the Chengdu area, where significant potential for improvement was uncovered.

(2)

It was evident that both the LCI and the impact factor exhibited distinct regional distribution characteristics. The spatial distribution of the LCI of building energy consumption in the case area was found to adhere to a pattern characterized as high in the southeast, low in the northwest, and average in the center. Deficiencies in the utilization of clean energy, the thermal performance of external walls, and the awareness of energy-saving behaviors among residents were observed in certain villages within the case area. A more detailed account of these patterns would contribute to a more comprehensive understanding of the salient features of rural building energy consumption.

(3)

The established evaluation model has been proven theoretically feasible, from the composite perspective of energy–buildings–behavior. Validated through the use of illustrative examples, the model was proven effective. Apart from its applicability in evaluating rural buildings in Southwest China, the evaluation model has been demonstrated to be adaptable to areas with difficult transportation access by adjusting the factors and evaluation criteria. Facilitating the provision of more comprehensive and accurate support, along with relevant data, for the construction and renovation of rural green buildings, this approach has proven beneficial.