Impact of Urban Building-Integrated Photovoltaics on Local Air Quality

Abstract

Amidst the global energy structure transition and intensification of climate warming, the temperature control targets of the Paris Agreement and China’s “dual carbon” goals have driven the rapid development of building-integrated photovoltaics (BIPVs). However, solar cells in BIPV systems may produce exhaust gases that affect local urban air quality if exposed to extreme environmental conditions such as high temperatures during operation. In this study, eight air quality monitoring points were established around the BIPV system at Shenyang Jianzhu University as the experimental group, along with one additional air quality monitoring point serving as a control group. The concentrations of four air pollutant indicators (PM_2.5, PM₁₀, SO₂, and NO₂) were monitored continuously for 14 days. The weight of each indicator was calculated using the principle of information entropy, and the air quality evaluation grades were determined by combining the homomorphic inverse correlation function. The Entropy-Weighted Set Pair Analysis model was applied to evaluate the air quality of the BIPV system at Shenyang Jianzhu University. The results indicated that due to the high concentrations of SO₂ and NO₂, the Air Quality Index (AQI) grade at Shenyang Jianzhu University was classified as “light pollution.” Corresponding recommendations were proposed to promote the sustainable development of urban BIPV. Simultaneously, the evaluation results of the Entropy-Weighted Set Pair Analysis model were similar to those obtained using other methods, demonstrating the feasibility of this evaluation model for assessing the impact on air quality.

1. Introduction

In the context of global energy structure transformation and intensified response to climate warming [1], the temperature control targets established by the Paris Agreement, synergistically advanced with China’s ‘carbon peak and carbon neutrality’ goals, have imposed rigid constraints on carbon emission control throughout the building life cycle [2]. As a critical sector for energy consumption and carbon emissions in countries worldwide, the construction industry’s shift toward low-carbon and green development has become an inevitable trend [3]. Driven by the low-carbon and green development of the construction industry, Building-Integrated Photovoltaics (BIPV) has entered a period of rapid growth [2]. As an innovative approach that deeply integrates solar power generation technology with architectural design, BIPV has become a key force promoting sustainable development in the construction industry [4]. Germany was one of the first countries to incorporate BIPV into its national energy strategy [5,6]. The United States has developed BIPV technology for application in rooftop photovoltaic systems and solar communities [7,8]. India has also adopted BIPV technology to reduce energy consumption resulting from rapid urban construction [9,10]. Gercek et al., studying 24 BIPV projects at the University of Twente in the Netherlands, found that BIPV can contribute to reducing building energy consumption [11]. In May 2024, China’s State Council issued the “Action Plan for Energy Conservation and Carbon Reduction 2024–2025,” which emphasized: ‘Stringently enforce mandatory building energy conservation and carbon reduction standards, strengthen green design and construction management, vigorously develop prefabricated buildings, actively promote smart construction, and accelerate the development of Building-Integrated Photovoltaics.’ The introduction of this policy has facilitated the transition of BIPV from ‘technology demonstration’ to ‘large-scale application’ [12]. With the widespread adoption of BIPV globally, attention has primarily focused on its energy supply and economic benefits, while research evaluating its impact on urban air quality remains relatively limited. However, Copper Indium Gallium Selenide (CIGS) thin-film solar cells used in urban BIPV may, when exposed to extreme environmental conditions such as high temperature, high humidity, and strong radiation during operation, undergo material decomposition or deterioration, and structural changes in the thin films, potentially leading to the emission of exhaust gases that impact urban air quality [13,14].

Existing scholars researching the environmental impact assessment of photovoltaic power stations have focused on their effects on the air environment. Based on a comprehensive photovoltaic application science popularization demonstration park project in Wuhan, Yang Ying found that the sulfur hexafluoride and lead-acid batteries required for photovoltaic use release harmful substances that affect air quality [15]. Wang Yiwen, taking the Talatan Photovoltaic Power Generation Park as the research subject, concluded that the construction of photovoltaic power stations would significantly alter the CO₂ concentration in the air [16]. Yuan Lingya also argued that the rapid development of the photovoltaic industry will increase CO₂ emissions [17]. Abdul Ghani Olabi discovered that optimizing photovoltaic power generation design can effectively reduce the emissions of greenhouse gases and harmful gases [18]. Chang Tsangyao found that the use of renewable energy in the United States increases CO₂ emissions, which has a significant negative impact on the environment [19]. Using houses with photovoltaic panels on their roofs as the research subject, Pandiaraj conducted CO₂ concentration tests and concluded that the CO₂ concentration around houses with photovoltaic panels is significantly higher than that around houses without them [20]. Similarly, using a photovoltaic project in a carport in Finland as the research subject, Ranta also found that the operation of the photovoltaic project would increase CO₂ emissions [21]. Taking urban photovoltaic rooftops as the research subject, Chen found that the construction of photovoltaic rooftops affects the extent of PM_2.5 diffusion and increases the PM_2.5 concentration in the air [22]. In summary, existing scholars have concluded that the construction of photovoltaic power stations has an impact on the air environment; however, there is a lack of research on the evaluation of urban air quality using Building-Integrated Photovoltaics (BIPV). Therefore, this study focuses on analyzing and evaluating the impact of BIPV on urban air quality after its implementation, to better leverage the advantages of photovoltaic power stations and take effective measures to address potential problems, aiming to provide strong support and guarantee for the sustainable development of BIPV.

This study considers the BIPV project of Shenyang Jianzhu University as an example and measures the concentrations of PM_2.5, PM₁₀, SO₂, and NO₂ at Shenyang Jianzhu University after the completion of the project. The weight of each indicator was calculated using the principle of information entropy, and the air quality evaluation grades were determined by combining the homomorphic inverse correlation function. The Entropy-Weighted Set Pair Analysis model was applied to evaluate the Air Quality Index (AQI) of Shenyang Jianzhu University, and other air quality evaluation methods were used to compare the evaluation results to verify the model accuracy, thereby determining the accuracy of the evaluation system for the impact of urban BIPVs on local air quality and providing guidance for the future development of the BIPV industry.

2. Materials and Methods

2.1. Overview of BIPVs at Shenyang Jianzhu University

To explore the impact of urban BIPVs on local air quality, Shenyang Jianzhu University built BIPV projects on campus, as shown in Figure 1a–d. The project consisted of four 3 m × 6 m box units, including three flat roof units and one pitched roof unit, each with a building area of 54 m². Thirty-two pieces of 100 Wp CIGS photovoltaic modules were installed on the south and west façades, and sixteen flexible CIGS modules were installed on the roof, totaling 440 Wp. The photovoltaic system generates 15–20 kW/h of electricity per day, which fully meets the lighting, cooling, and heating requirements of the building.

The solar cells selected for the photovoltaic base were CIGS thin-film cells. Unlike ordinary solar cells, the raw material of CIGS thin-film cells in the photovoltaic base is not silicon but is made into a thin-film form by layering compounds such as copper, indium, gallium, and selenium [23]. Compared to traditional silicon-based solar cells, CIGS thin-film cells consume less energy and are less expensive during production [24]. Consequently, this technology has been widely used in BIPVs [25]. CIGS cells have a higher energy conversion rate and better adaptability and can produce a more stable power output [26] under different environmental conditions without harmful substances [27]. However, during photoelectric conversion, the elements contained in CIGS cells may produce exhaust gases such as sulfur-containing compounds when the cells are damaged or burned at high temperatures [28], thereby affecting air quality.

2.2. Air Quality Monitoring Method and Location Determination

2.2.1. Determine the Air Quality Monitoring Method

In determining the air quality monitoring method, a miniature air quality monitoring station, model FT-CQX12 (Shandong Fengtu Internet of Things Technology Co., Ltd., Weifang, China) was selected. This meteorological station is a highly integrated, low-power, rapidly deployable automatic weather observation device capable of real-time monitoring of meteorological elements such as temperature, wind speed, PM_2.5, PM₁₀, carbon monoxide (CO), nitrogen dioxide (NO₂), sulfur dioxide (SO₂), and ozone (O₃). The station was deployed at the monitoring points for 24 h real-time monitoring and data collection to ensure data accuracy and reliability. Finally, SPSS analysis software was used for data processing.

2.2.2. Set up Air Quality Monitoring Points

According to the principle that ambient air quality monitoring points must be representative, comparable, holistic, forward-looking, and stable [29], and considering that increasing the number and coverage of monitoring points is an important measure to improve the accuracy of air quality monitoring [30], this study fully considered the layout of BIPVs at Shenyang Jianzhu University and the complexity of the surrounding environment. A total of eight air quality monitoring points were established as the experimental group: four located 100 m east, south, west, and north of the BIPVs, and others near important surrounding buildings such as the school hospital and water sources. One additional air quality monitoring point was set up far from the BIPVs as the control group. The sampling inlets were placed 3 m above the ground to monitor air pollutant concentrations in real-time [31], thus forming a multi-point and wide-coverage air quality monitoring network, as shown in Figure 2. All nine air monitoring points were set up under identical conditions to ensure the comparability of the data obtained from each point.

2.3. Construction of the Entropy-Weighted Set Pair Analysis Model for Air Quality

2.3.1. Selection of the Entropy-Weighted Set Pair Analysis Model for Air Quality

Set Pair Analysis (SPA) is a novel method for the quantitative analysis of certain and uncertain problems, which has been widely used in the past for water quality evaluation [32,33]. Given that the atmosphere is a complex and uncertain system with multi-factor and multi-level coupling effects, Guo Shaoying [34] applied SPA by treating the connection degree components of atmospheric environmental quality and evaluation classification criteria as weights. They calculated the air quality of samples to be evaluated through weight differences, achieving favorable results. However, the aforementioned approach lacked in-depth research on haze during the study process, did not consider the impact of key pollutants PM_2.5 and PM₁₀ on the environment, and directly classified evaluation results by rounding off the calculated grade values, resulting in relatively low accuracy. The Entropy-Weighted Set Pair Analysis model is a comprehensive evaluation model that combines the objective weighting of information entropy with the ability of SPA to handle uncertainties. It is used to address the interrelationships between certainty and uncertainty in multi-indicator systems and make quantitative judgments [35]. Building on this, the present study expanded the categories of pollutant indicators in the evaluation system of the BIPV project at Shenyang Jianzhu University in accordance with ambient air quality standards [36]. Furthermore, the principle of information entropy was adopted to determine the weights of each evaluation indicator [37,38].Compared to traditional Life Cycle Assessment methods for environmental impact evaluation of BIPV projects, the Entropy-Weighted Set Pair Analysis model for air quality constructed in this study more accurately reflects the real-time and dynamic impact of BIPV projects on local air quality during the operational phase. It avoids the potential spatiotemporal averaging limitations inherent in full life cycle LCA. Simultaneously, by employing objective weighting through information entropy, the model fully accounts for the synergistic effects and uncertainties of multiple pollutants, offering a new approach for air quality assessment of urban BIPV systems.

2.3.2. Determination of Air Quality Evaluation Indicators and Classification Standard

Due to the wide variety of air pollutants, it is difficult to estimate the generation of secondary pollutants via complex reactions. Therefore, the effects of fluoride, lead, and acid rain on air quality were not considered in this model [39]. To make the model more acceptable to the general public and facilitate the verification of model results, the classification criteria should be consistent with China’s current air quality assessment system (AQI), and the selection of indicators should also be comprehensive. Meanwhile, the diffusion pathways and measured concentrations of air pollutants are also influenced by environmental meteorological conditions [40]. For example, as ambient wind speed and temperature increase, the concentrations of pollutants such as SO₂ and NO₂ in the air decrease accordingly [41]. Therefore, in this study, while monitoring PM_2.5, PM₁₀, SO₂, NO₂, CO, and O₃, wind speed and temperature at nine air quality monitoring points were simultaneously monitored to determine whether changes in pollutant concentrations were caused by differences in environmental meteorological conditions. The monitoring data were analyzed using the Bonferroni-corrected p-value method in SPSS 27.0 software to assess the significance of differences between the control and experimental groups at the air quality monitoring points. Indicators showing significant differences were selected as air quality evaluation indicators to further analyze the impact characteristics of BIPV on air quality. The significance of the eight indicators between the control and experimental groups at the air quality monitoring points is presented in

Table 1. Significance comparison of each indicator between the control group and the experimental group based on the Bonferroni-corrected p-value method.

Paired Samples	Wind Speed	Temperature	O3	CO	PM2.5	PM10	SO2	NO2
Control point and Point 1	0.943	0.999	0.863	0.703	0.012	0.001	0.029	0.001
Control point and Point 2	0.922	0.999	0.597	0.787	0.003	0.001	0.033	0.001
Control point and Point 3	0.928	0.996	0.870	0.772	0.01	0.001	0.026	0.001
Control point and Point 4	0.974	0.995	0.780	0.869	0.005	0.001	0.035	0.001
Control point and Point 5	0.906	0.999	0.851	0.824	0.006	0.001	0.029	0.001
Control point and Point 6	0.910	0.996	0.859	0.894	0.01	0.001	0.023	0.001
Control point and Point 7	0.961	0.999	0.65	0.799	0.004	0.001	0.029	0.001
Control point and Point 8	0.968	0.999	0.955	0.638	0.002	0.001	0.029	0.001

.

As can be seen from the table above, for the environmental meteorological indicators of wind speed and air temperature, there were no significant differences (p > 0.05) between the eight monitoring points in the control group and the experimental group. Therefore, the potential influence of differences in wind speed and air temperature on the air quality monitoring results of the control and experimental groups could be appropriately excluded. Meanwhile, for the air pollutant indicators of O₃ and CO, there were also no significant differences (p > 0.05) between the eight monitoring points in the control group and the experimental group. However, for the four air pollutant indicators—PM_2.5, PM₁₀, SO₂, and NO₂—significant differences (p < 0.05) were observed between the eight monitoring points in the control group and the experimental group. This indicates that the construction of the BIPV at Shenyang Jianzhu University may have a certain impact on the concentrations of PM_2.5, PM₁₀, SO₂, and NO₂ in the air. Consequently, PM_2.5, PM₁₀, SO₂, and NO₂ were ultimately determined as the air quality evaluation indicators for this study.

In accordance with the above principles and the national standards for pollutant concentration limits specified in the “Technical Regulation on Ambient Air Quality Index (on trial)” [42], the evaluation grades are divided into six levels from I to VI, corresponding to excellent, good, light pollution, moderate pollution, severe pollution, and serious pollution in the AQI grades. The specific indicators and classification criteria are listed in

Table 2. Air quality classification criteria.

Evaluation Grades
Indicators	I	II	III	IV	V	VI
PM2.5	35	75	115	150	250	350
PM10	50	150	250	350	420	500
SO2	50	150	475	800	1600	2100
NO2	40	80	180	280	565	750

Note: The data for PM_2.5, PM₁₀, SO₂, and NO₂ are all 24 h average concentrations, measured in μg/m³.

.

2.3.3. Determination of Entropy Weights in the Entropy-Weighted Method

The Entropy-Weighted method is an objective weight method that uses information entropy to calculate the weights of each indicator based on the degree of variation in the values of each indicator. This can, to some extent, avoid the bias caused by subjective consciousness. For a comprehensive problem with m days of air quality to be evaluated and n evaluation indicators, construct an initial data matrix x_ij of m × n. Then, the proportion of the pollutant concentration value on the i-th day in the j-th indicator is defined as:

$$y_{ij}={\displaystyle \frac{x_{ij}}{{\displaystyle \sum _{i=1}^D{x}_{ij}}}}(0\le y_{ij}\le 1)$$

(1)

The entropy is a measure of the degree of disorder in a system. The system may exist in different states and the probability of each state occurring is denoted by p_i. The entropy of the system is defined as:

$$e=-{\displaystyle \sum p_i\ln }{p}_i$$

(2)

If the air quality of each day under the i-th indicator is regarded as a system and the pollutant concentrations of each day as different states the system may be in, then the information entropy of the j-th indicator is defined as Equation (3). In Equation (3), K is the Boltzmann constant and K = 1/lnm.

$$e_j=-K{\displaystyle \sum _{i=1}^m{y}_{ij}\ln y_{ij},j=1,2,\cdots ,n}$$

(3)

The Entropy-Weighted method was used to estimate the weight of each indicator. Its essence is calculation based on the value coefficient of the indicator information. The greater the weight, the greater the contribution to the evaluation results. The smaller the information entropy e_j of an indicator, the greater the variation in the indicator’s value, the more information it provides, and the greater its weight in the overall evaluation. We obtain the final weight of the n-th indicator as follows:

$${\omega }_j={\displaystyle \frac{1-e_j}{{\displaystyle \sum _{i=1}^m(1-e_j)}}},j=1,2,\cdots ,n$$

(4)

2.3.4. Determination of Relevance in Set Pair Analysis

The key to using Set Pair Analysis for air quality evaluation is to determine the degree of connection between each evaluation index and the evaluation criteria. The membership function of the degree of association of each evaluation index with respect to the evaluation classification criteria was constructed to describe the quantitative relationships of identity, difference, and opposition with the criteria. Combining the six constructed pollutant indicators and six evaluation grades, the specific correlation calculation formulae are expressed as follows:

The connection degree of the j indicator to level I is:

$${\mu }_{j\mathrm{{\rm I}}}=\left\{\begin{array}{cc}1& x\in \left[0,\left.S_{j(1)}\right]\right.\\ 1+{\displaystyle \frac{2(x-S_{j(1)})}{S_{j(1)}-S_{j(2)}}}x& x\in \left[S_{j(1)},\left.S_{j(2)}\right]\right.\\ 1& x\in \left[S_{j(2)},\left.\infty \right]\right.\end{array}\right.$$

(5)

The connection degree of the j indicator to level k is:

$${\mu }_{jk}=\left\{\begin{array}{cc}1+{\displaystyle \frac{2(x-S_{j(k-1)})}{S_{j(k-1)}-S_{j(k-2)}}}& x\in \left[S_{j(k-2)},\left.S_{j(k-1)}\right]\right.\\ 1& x\in \left[S_{j(k-1)},\left.S_{j(k)}\right]\right.\\ 1+{\displaystyle \frac{2(x-S_{j(k)})}{S_{j(k)}-S_{j(k+1)}}}& x\in \left[S_{j(k)},\left.S_{j(k+1)}\right]\right.\\ -1& x\in \left[S_{j(k)},\left.S_{j(k+1)}\right]\right.\cup \left[0,\left.S_{j(k-2)}\right]\right.\end{array}\right.$$

(6)

The connection degree of the j indicator to level VI is:

$${\mu }_{jV\mathrm{{\rm I}}}=\left\{\begin{array}{cc}-1& x\in \left[0,\left.S_{j(4)}\right]\right.\\ 1+{\displaystyle \frac{2(x-S_{j(1)})}{S_{j(1)}-S_{j(2)}}}x& x\in \left[S_{j(4)},\left.S_{j(5)}\right]\right.\\ 1& x\in \left[S_{j(5)},\left.S_{j(6)}\right]\right.\end{array}\right.$$

(7)

In Equations (5)–(7), k is the evaluation level, an integer between 1 and 6; x represents the actual measured value of the evaluation indicator; S_jk represents the k level classification standard of the i-th indicator; and μ_jk represents the level k connection degree of the j-th indicator. Such a membership function of the degree of connection reflects a definite and uncertain relationship between the degrees of identity, difference, and opposition.

2.3.5. Determination of Air Quality Evaluation Grades in the Entropy-Weighted Set Pair Analysis Model

After obtaining the weights of each pollutant concentration indicator, the weighted average method was used, in which the degree of association of each indicator was multiplied by its corresponding weight, and the weighted average was used as the degree of association for each grade. The formula for the weighted average degree of association was as follows:

$${\mu }_{ik}={\displaystyle \sum _{j=1}^6{\omega }_j\times }{\mu }_{ik}$$

(8)

where μ_ik is the weighted average connection degree of the air quality assessment grade on i-th day relative to k-th grade.

The air quality evaluation grades in the Entropy-Weighted Set Pair Analysis model are as shown in Equation (9), where for a certain day to be evaluated, the grade corresponding to the maximum average connection degree is taken as the final evaluation grade.

$${\beta }_i=\max \left\{{\mu }_{i1},\right.{\mu }_{i2,}\cdots ,\left.{\mu }_{ik}\right\}$$

(9)

3. Results and Discussion

3.1. Determination of Air Quality Monitoring Results

In the air quality assessment of the BIPV system at Shenyang Jianzhu University, this study established a total of nine air quality monitoring points to conduct real-time monitoring of the concentrations of suspended particulate matter (PM_2.5, PM₁₀) and oxides (SO₂, NO₂) in air pollutants over a period of 14 days. The monitoring results are shown in Figure 3a–d. From the concentration trends of each pollutant indicator in Figure 3 and the significance results in Table 1, it can be observed that the concentrations of the four air pollutant indicators in the control group were significantly lower than those at the eight experimental air quality monitoring points over the 14-day period (p < 0.05), indicating that the BIPV system at Shenyang Jianzhu University may lead to increased concentrations of PM_2.5, PM₁₀, SO₂ and NO₂ in the air. Although the Bonferroni correction method was previously applied to exclude the influence of differences in wind speed and temperature among the nine air quality monitoring points on the air quality monitoring results, due to the short monitoring period in this study and the fact that air pollutant concentrations are also affected by urban traffic and industrial pollution sources, the potential increase in air pollutant concentrations attributable to the BIPV system requires further analysis and verification in subsequent studies.

To verify whether there was a correlation among the concentrations of the indicators at different air monitoring points, the PM_2.5, PM₁₀, SO₂, and NO₂ concentrations at the eight monitoring points on the test day were analyzed using SPSS analysis software. The Pearson correlation coefficients for the PM_2.5, PM₁₀, SO₂, and NO₂ concentrations at different monitoring points are presented in

Table 3. Pearson correlation coefficient of PM_2.5 concentrations at different monitoring points.

	1	2	3	4	5	6	7	8
1	1	0.999	0.998	0.998	0.993	0.993	0.995	0.996
2	0.999	1	0.998	0.998	0.995	0.990	0.996	0.994
3	0.998	0.998	1	0.997	0.995	0.992	0.994	0.992
4	0.998	0.998	0.997	1	0.991	0.991	0.993	0.995
5	0.993	0.995	0.995	0.991	1	0.977	0.994	0.985
6	0.993	0.990	0.992	0.991	0.977	1	0.980	0.991
7	0.995	0.996	0.994	0.993	0.994	0.980	1	0.990
8	0.996	0.994	0.992	0.995	0.985	0.991	0.990	1

,

Table 4. Pearson correlation coefficient of PM₁₀ concentrations at different monitoring points.

	1	2	3	4	5	6	7	8
1	1	1.000	1.000	1.000	0.999	0.998	0.998	0.996
2	1.000	1	1.000	1.000	0.999	0.998	0.998	0.996
3	1.000	1.000	1	0.999	0.999	0.998	0.997	0.995
4	1.000	1.000	0.999	1	0.999	0.998	0.998	0.995
5	0.999	0.999	0.999	0.999	1	0.995	0.998	0.994
6	0.998	0.998	0.998	0.998	0.995	1	0.993	0.994
7	0.998	0.998	0.997	0.998	0.998	0.993	1	0.992
8	0.996	0.996	0.995	0.995	0.994	0.994	0.992	1

,

Table 5. Pearson correlation coefficient of SO₂ concentrations at different monitoring points.

	1	2	3	4	5	6	7	8
1	1	0.997	0.997	0.994	1.000	0.993	1.000	1.000
2	0.997	1	0.988	0.999	0.997	0.981	0.997	0.997
3	0.997	0.988	1	0.982	0.997	0.999	0.997	0.996
4	0.994	0.999	0.982	1	0.993	0.974	0.994	0.994
5	1.000	0.997	0.997	0.993	1	0.994	1.000	1.000
6	0.993	0.981	0.999	0.974	0.994	1	0.993	0.993
7	1.000	0.997	0.997	0.994	1.000	0.993	1	1.000
8	1.000	0.997	0.996	0.994	1.000	0.993	1.000	1

and

Table 6. Pearson correlation coefficient of NO₂ concentrations at different monitoring points.

	1	2	3	4	5	6	7	8
1	1	0.894	0.991	0.999	0.990	0.988	0.996	0.988
2	0.894	1	0.882	0.874	0.824	0.855	0.851	0.815
3	0.991	0.882	1	0.991	0.982	0.966	0.987	0.979
4	0.999	0.874	0.991	1	0.994	0.990	0.998	0.992
5	0.990	0.824	0.982	0.994	1	0.984	0.999	1.000
6	0.988	0.855	0.966	0.990	0.984	1	0.988	0.983
7	0.996	0.851	0.987	0.998	0.999	0.988	1	0.997
8	0.988	0.815	0.979	0.992	1.000	0.983	0.997	1

.

Table 3, Table 4, Table 5 and Table 6 show that the Pearson correlation coefficients of the concentrations of PM_2.5, PM₁₀, SO₂, and NO₂ at different monitoring points are all >0.8, indicating significant correlations. Therefore, the arithmetic means of the PM_2.5, PM₁₀, SO₂, and NO₂ at concentrations the eight air monitoring points were combined to determine the air quality monitoring results at the BIPVs of Shenyang Jianzhu University, as shown in

Table 7. Final monitoring results of air quality.

Date	PM2.5	PM10	SO2	NO2
1.20	37.25	57.25	599	192
1.21	36.875	56.5	481.25	182
1.22	37.125	56.75	470	179
1.23	46	73	475.5	180.5
1.24	44	70	476	180.5
1.25	36.875	56	562.5	186.25
1.26	38.25	58	468.75	150
1.27	35.8	59	481	179.25
1.28	36.75	58.75	487.5	156.25
1.29	38.125	57.125	468.75	177.5
1.30	39.25	55.625	381.25	150
1.31	37.25	58.25	525	175
2.01	35.15	50	319	120
2.02	39.25	57.875	443.75	150

Note: The data for PM_2.5, PM₁₀, SO₂, and NO₂ are all 24 h average concentrations, measured in μg/m³.

.

3.2. Evaluation of the Impact of Urban BIPVs on Air Quality

First, the initial pollutant concentration matrix x_ij was constructed based on Equation (1), and then the weights of the four pollutant indicators were calculated using the Entropy-Weighted method using Equation (4). As shown in

Table 8. Weights of air quality evaluation indicators.

Indicators	PM2.5	PM10	SO2	NO2
Weights	0.06	0.16	0.39	0.39

, the main pollutants affecting air quality at Shenyang Jianzhu University are SO₂ and NO₂, followed by PM₁₀ and PM_2.5.

Then, we calculated the correlation degree of each pollutant evaluation index with respect to each evaluation grade from Equations (5)–(7) and presented the data and calculation results of the first four test days of this experiment, as shown in

Table 9. Correlation between each pollutant index and evaluation grade on the test day.

Date	Indicators	The Degree of Association of Each Evaluation Grade
		I	II	III	IV	V	VI
1.20	PM2.5	0.89	1	−0.89	−1	−1	−1
	PM10	0.86	1	−0.86	−1	−1	−1
	SO2	−1	−1	0.24	1	−0.24	−1
	NO2	−1	−1	0.76	1	−0.76	−1
1.21	PM2.5	0.91	1	−0.91	−1	−1	−1
	PM10	0.87	1	−0.87	−1	−1	−1
	SO2	−1	−1	0.96	1	−0.96	−1
	NO2	−1	−1	0.96	1	−0.96	−1
1.22	PM2.5	0.89	1	−0.89	−1	−1	−1
	PM10	0.87	1	−0.87	−1	−1	−1
	SO2	−1	−0.97	1	0.97	−1	−1
	NO2	−1	−0.98	1	0.98	−1	−1
1.23	PM2.5	0.45	1	−0.45	−1	−1	−1
	PM10	0.54	1	−0.54	−1	−1	−1
	SO2	−1	−1	0.99	1	−0.99	−1
	NO2	−1	−1	0.990	1.010	−1	−1

.

According to the data obtained in Table 9, the air quality monitoring data of SO₂ and NO₂ after the construction and operation of BIPV at Shenyang Jianzhu University are significantly higher than the secondary standards of “Ambient air quality standards (GB3095-2012) [39].” According to Equation (8), the weighted average correlations between air quality on January 20 and each evaluation grade were −0.59, −0.56, 0.20, 0.56, −0.61, and −1.00. The air quality of that day had the highest weighted average correlation with grade IV; therefore, the air quality of that day was classified as grade IV, which is “moderate pollution.” The model was applied to the final air quality monitoring results of the BIPVs at Shenyang Jianzhu University from 20 January to 2 February to obtain the corresponding air quality grades, and the evaluation results of the air quality grades are listed in

Table 10. Sample air quality evaluation grades.

Date	The Degree of Association of Each Rating Level						Grades
	I	II	III	IV	V	VI
1.20	−0.59	−0.56	0.20	0.56	−0.61	−1.00	IV
1.21	−0.59	−0.56	0.56	0.56	−0.97	−1.00	IV
1.22	−0.59	−0.54	0.59	0.54	−1.00	−1.00	III
1.23	−0.67	−0.56	0.66	0.56	−1.00	−1.00	III
1.24	−0.67	−0.56	0.66	0.56	−1.00	−1.00	III
1.25	−0.59	−0.56	0.33	0.56	−0.74	−1.00	IV
1.26	−0.60	−0.31	0.60	0.31	−1.00	−1.00	III
1.27	−0.38	−0.39	0.28	0.26	−0.39	−0.39	III
1.28	−0.39	−0.21	0.29	0.08	−0.39	−0.39	III
1.29	−0.39	−0.37	0.29	0.25	−0.39	−0.39	III
1.30	−0.40	−0.12	0.29	0.07	−0.39	−0.39	III
1.31	−0.39	−0.35	0.30	0.23	−0.41	−0.39	III
2.1	−0.39	0.14	0.27	−0.14	−0.39	−0.39	III
2.2	−0.40	−0.15	0.29	0.05	−0.39	−0.39	III

.

According to Table 10, the AQI grades of the air quality evaluation of BIPVs at Shenyang Jianzhu University from 20 January to 2 February are mostly “light pollution.” During 20, 21, and 25 January, the SO₂ and NO₂ concentrations were relatively high, which reduced the air quality grade of the days to some extent, with the air quality evaluation AQI grade being “moderate pollution.” Therefore, the BIPVs of Shenyang Jianzhu University should adopt measures such as vegetation restoration and ecological compensation to improve the air quality [43].

3.3. Verification of the Air Quality Evaluation Results of the Entropy-Weighted Set Pair Analysis Model

3.3.1. Traditional Set Pair Analysis Method

To verify the accuracy of the Entropy-Weighted Set Pair Analysis model on the air quality evaluation results, this study evaluated the air quality of Shenyang Jianzhu University over 14 days using traditional Set Pair Analysis and AQI method [44]. The traditional Set Pair Analysis method is a new system analysis method for the quantitative analysis of similarities, differences, and the reverse of definite and uncertain systems proposed by Zhao Keqin [45,46]. The basic idea and method of the traditional Set Pair Analysis method is to identify two sets, U and V, that have a certain connection for a specific problem to form a set pair, and to analyze the identity, difference, and opposition of their characteristics to establish the following expression: μ = a + bi + cj (a, b and c are the identity, difference, and opposition degrees, respectively) [47]. The traditional Set Pair Analysis method can effectively analyze regional atmospheric environmental quality by calculating the correlation between the concentrations of various pollutants in the air [48]. The traditional Set Pair Analysis method was used with Equations (5)–(7) to calculate the correlations of the evaluation grades of Shenyang Jianzhu University over 14 days, as listed in

Table 11. Air quality evaluation grades using the traditional Set Pair Analysis method.

Date	The Degree of Association of Each Rating Level						Grades
	I	II	III	IV	V	VI
1.20	−0.06	0	−0.19	0	−0.75	−1	II
1.21	−0.06	0	0.04	0	−0.98	−1	III
1.22	−0.06	0.01	0.06	−0.01	−1	−1	III
1.23	−0.25	0	0.25	0.01	−0.99	−1	III
1.24	−0.25	0	0.25	0.01	−0.99	−1	III
1.25	−0.05	0	−0.11	0	−0.83	−1	II
1.26	−0.08	0.16	0.08	−0.16	−1	−1	III
1.27	−0.06	0.003	0.05	−0.003	−0.99	−1	III
1.28	−0.07	0.12	0.05	−0.12	−0.98	−1	III
1.29	−0.07	0.02	0.07	−0.02	−1	−1	III
1.30	−0.08	0.29	0.08	−0.29	−1	−1	II
1.31	−0.07	0.02	−0.01	−0.03	−0.92	−1	II
2.1	0	0.54	0.01	−0.54	−1	−1	II
2.2	−0.09	0.20	0.09	−0.20	−1	−1	II

.

3.3.2. AQI Method

AQI is a dimensionless index used to describe the state of air quality. The AQI visually reflects the degree of air pollution by converting the concentrations of various pollutants into a uniform numerical value, and can determine the air quality grade for a certain day based on the value [49,50]. The larger the AQI value, the more severe the air pollution and worse the air quality. AQIs between 0 and 50, 51 and 100, 101 and 150, 151 and 200, 201 and 300, and >300 represent grades 1, 2, 3, 4, 5 and 6, respectively.

The AQI was calculated by first calculating the sub-index (IAQI) for each pollutant and then taking the maximum of these sub-indices as the final AQI value. The sub-index (IAQI) is calculated as follows.

$$IAQ{I}_p={\displaystyle \frac{IAQ{I}_{Hi}-IAQ{I}_{Lo}}{B{P}_{Hi}-B{P}_{Lo}}}(C_p-B{P}_{Lo})+IAQ{I}_{Lo}$$

(10)

where IAQI_p is the air quality sub-index of pollutant P, C_p is the mass concentration of pollutant P, BP_Hi is the higher value of the pollutant concentration limit similar to that of C_p, BP_Lo is the lower value of the pollutant concentration limit close to C_p, IAQI_Hi is the air quality sub-index corresponding to BP_Hi, and IAQI_Lo is the sub-index of air quality corresponding to BP_Lo. The formula for the AQI is shown in Equation (11):

$$AQI=\max \left\{IAQ{I}_1,\right.IAQ{I}_2,IAQ{I}_3,\cdots ,\left.IAQ{I}_n\right\}$$

(11)

The AQI of Shenyang Jianzhu University for 14 days was calculated and graded using the AQI method combined with Equations (10) and (11), and the results are shown in

Table 12. Air quality evaluation grades using the AQI method.

Date	Sub-Index of Air Quality IAQI				AQI	Grades
	PM2.5	PM10	SO2	NO2
1.20	52.81	53.62	167.61	156.00	167.61	IV
1.21	52.34	53.25	150.96	151.00	151.00	IV
1.22	52.66	53.38	149.23	149.50	149.50	III
1.23	63.10	61.50	150.08	150.25	150.25	III
1.24	61.25	60.00	150.15	150.25	150.25	III
1.25	52.34	53.00	163.46	153.12	163.46	IV
1.26	54.06	54.00	149.04	135.00	149.04	III
1.27	51.00	54.50	150.92	149.62	150.92	IV
1.28	41.56	54.38	151.92	138.0	151.92	IV
1.29	53.90	53.56	149.04	148.75	149.04	III
1.30	55.31	52.81	135.57	135.00	135.57	III
1.31	52.00	54.125	157.69	147.5	157.69	IV
2.1	50.19	0	126	120	126	III
2.2	67.81	53.93	145.19	135.	145.19	III

.

3.3.3. Comparison of Different Air Quality Evaluation Methods

A comparison of the air quality grade evaluation results obtained by the traditional Set Pair Analysis method and the AQI method with those from the Entropy-Weighted Set Pair Analysis method is presented in

Table 13. Comparison of evaluation results between Entropy-Weighted Set Pair Analysis method and other methods.

Date	Traditional Set Pair Analysis Method	AQI Method	Entropy-Weighted Set Pair Analysis Method
1.20	II	IV	IV
1.21	III	IV	IV
1.22	III	III	III
1.23	III	III	III
1.24	III	III	III
1.25	II	IV	IV
1.26	III	III	III
1.27	III	IV	III
1.28	III	IV	III
1.29	III	III	III
1.30	II	III	III
1.31	II	IV	III
2.1	II	III	III
2.2	II	III	III

. As can be seen from the table, the evaluation results of the traditional Set Pair Analysis method deviate significantly from the other two methods. This is because the traditional method does not account for the differences in concentrations of suspended particulate matter (PM_2.5, PM₁₀) and oxides (SO₂, NO₂) among air pollutants and applies equal weights in the air quality assessment. In contrast, the air quality rating results from the AQI method and the Entropy-Weighted Set Pair Analysis method are generally similar, which also demonstrates the rationality of the Entropy-Weighted Set Pair Analysis method in air quality evaluation. However, compared to the limitation of the AQI method, which determines the grade solely based on the maximum sub-index of a single pollutant, the Entropy-Weighted Set Pair Analysis model constructed in this study uses information entropy for objective weighting. It fully considers the synergistic effects of multiple pollutants and avoids deviations caused by subjective weight allocation, thereby making the evaluation process more scientific and comprehensive.

4. Conclusions

(1)

Combined with the air quality monitoring results of the BIPV project at Shenyang Jianzhu University, it was found that the concentrations of the four air pollutant indicators—PM_2.5, PM₁₀, SO₂, and NO₂—in the control group were significantly lower than those at the eight air quality monitoring points in the experimental group on the monitoring days (p < 0.05). This indicates that the BIPV system at Shenyang Jianzhu University may lead to an increase in the concentrations of air pollutant indicators, possibly due to the decomposition and deterioration of internal materials in solar cells under extreme environmental conditions, resulting in the release of pollutant gases. However, as the monitoring period of this study was relatively short and did not include analysis of seasonal variations, and since air pollutant concentrations are also influenced by urban traffic, industrial activities, and other pollution sources, subsequent studies should conduct spatially dispersed assessments of PM_2.5, PM10, SO2, and NO₂ to further analyze and verify the enhancing effect of BIPV on air pollutant concentrations.

(2)

Based on the analysis of air quality around the BIPV system at Shenyang Jianzhu University, it can be observed that the concentrations of SO₂ and NO₂ at the monitoring points correspond to Grade III and IV levels in the AQI, falling between “light pollution” and “moderate pollution.” This indicates that elevated SO₂ and NO₂ concentrations are the main reasons why the final AQI at Shenyang Jianzhu University is classified as “moderate pollution.” Therefore, during the construction of BIPV systems, the impact of oxides in air pollutants on air quality should be taken into account. At the BIPV system design stage, priority should be given to battery technologies with low or zero sulfur and nitrogen emissions, such as lithium iron phosphate batteries, and high-efficiency exhaust gas filtration devices should be equipped to reduce pollutant release from the source [51]. During operation and maintenance, a regular testing and battery replacement mechanism should be established to prevent aged batteries from generating air pollutants under extreme operating conditions. Additionally, vegetation with strong adsorption capacity for SO₂ and NO₂, such as Ginkgo biloba and Ligustrum lucidum, can be planted around the BIPV system to assist in air purification through ecological compensation methods [52].

(3)

The Entropy-Weighted Set Pair Analysis model can serve as an effective method for air quality evaluation. In this study, the air quality environment of the BIPV project at Shenyang Jianzhu University was evaluated using the Entropy-Weighted Set Pair Analysis method. The results were compared with those obtained from the traditional Set Pair Analysis method and the AQI method. The air quality evaluation grades were found to be similar to those derived from the AQI method. Moreover, compared to the traditional Set Pair Analysis method, which applies equal weighting, and the AQI method, which determines the grade solely based on the maximum sub-index of a single pollutant—both having certain limitations—the Entropy-Weighted Set Pair Analysis model objectively assigns weights through information entropy. This approach fully considers the synergistic effects of multiple pollutants and avoids biases associated with subjective weight allocation. Therefore, it demonstrates rationality in the air quality assessment of BIPV systems and can be regarded as an effective means for evaluating the air quality impacts of BIPV projects.