Toward Low-Carbon and Cost-Efficient Prefabrication: Integrating Structural Equation Modeling and System Dynamics

Abstract

Against the backdrop of the ongoing implementation of the “dual-carbon” strategy and green building policies, this study concentrates on the production stage of precast concrete (PC) components. A composite analytical framework that integrates the structural equation model (SEM) with the system dynamics model (SD) is developed, through which a systematic and dynamically responsive model for the joint optimization of carbon emissions and costs is proposed. The results demonstrate that (1) when investment in green policies is maintained within the range of 10–20%, a 4.2% reduction in carbon emissions can be achieved by 2030, while costs remain optimized; (2) under the scenario of moderate green policy investment (10–20%) combined with a carbon tax of CNY 100/ton, carbon emissions can be reduced by 7.52%, with costs also reaching an optimal level; and (3) among the multi-path emission reduction strategies, the technology optimization pathway and energy structure optimization pathway achieve reductions of 9.68% and 8.97%, respectively. These findings provide theoretical support for the coordinated control of carbon emissions and costs during the production stage of PC components, while also offering empirical evidence and practical guidance for governments in formulating green building policies and for enterprises in advancing low-carbon transitions.

1. Introduction

1.1. Research Background

Amid escalating global climate change and increasing pressure from resource and environmental constraints, pursuing the dual objectives of carbon peaking and carbon neutrality has become a central priority in China’s ecological civilization strategy. The construction sector, as one of the largest sources of national energy use and carbon emissions [1], is estimated to contribute nearly 40% of energy-related carbon emissions worldwide [2]. Data from the China Building Energy Consumption Research Report (2023), issued by the China Building Energy Efficiency Association [3], indicate that in 2021, the life cycle carbon emissions of residential construction in China reached 4.07 billion tons of CO₂, accounting for 38.2% of the country’s total energy-related emissions. Consequently, advancing a green and low-carbon transition in the construction industry is essential to achieving the national “dual-carbon” goals.

Against the macro background of the continuous advancement of green building policies, prefabricated buildings have developed rapidly [4]. Among them, PC components, due to their advantages of industrialized production, standardization, and energy efficiency, have gradually become an important means of promoting construction industrialization and emission reduction. However, during their production stage, energy-intensive processes and raw material manufacturing are still involved, leading to a heavy carbon emission burden [5]. At the same time, while enterprises face the pressure of carbon reduction, they also encounter uncertainties in cost increases, and the synergy between carbon reduction and cost control remains unclear. The typical characteristics of the production stage include high energy-consuming processes (such as cement preparation, steel processing, concrete mixing, and curing), a strong dependence on materials (with significant differences in the carbon emission factors of cement, steel, and aggregates), a high degree of industrialization and standardization (which facilitates process optimization and technological upgrading), and batch production and economies of scale (where the carbon emissions and costs per unit component are significantly affected by production scale and technological level).

Given these characteristics, this study focuses on the production stage of prefabricated components as the primary research object. This stage accounts for a significant proportion of life cycle carbon emissions, while also allowing enterprises to exert direct influence over both costs and emissions through technological innovation, process refinement, and management practices. Such controllability provides considerable potential for mitigation and efficiency gains. Concentrating on this stage, thus, enables a clearer examination of the dynamic coupling between carbon emissions and costs, while offering quantitative insights that can inform policy interventions and support low-carbon practices within enterprises.

Research on the accounting of carbon emissions and reduction pathways for PC components has been conducted [6]; however, most efforts have remained focused on life cycle assessment (LCA) and static calculations. Systematic modeling and scenario-based simulation of the dynamic relationship between emissions and costs during the production stage have been largely absent. Moreover, studies on influencing mechanisms and policy intervention effects have often relied on single modeling approaches with limited feedback mechanisms, which constrain their ability to capture complex causal relationships and temporal evolutionary processes.

Based on this, the production stage of PC components is chosen as the focus of this study. By combining SEM with the system dynamics (SD) model, a joint optimization model for carbon emissions and costs is built with both explanatory and dynamic feedback functions. Using multi-scenario simulations, this study examines how key factors such as green policy investment, carbon taxes, and new technology options affect both emission reduction and cost changes. The aim is to provide practical evidence and quantitative support for governmental policies on green buildings and for enterprises seeking low-carbon transformation.

1.2. Literature Review

1.2.1. Research Status of Carbon Emissions of PC Components

In recent years, extensive research has been carried out by scholars both in China and abroad on carbon emission accounting [7,8,9]. In the field of prefabricated buildings, carbon emissions are commonly calculated using the life cycle assessment (LCA) method, which mainly includes process-based LCA (P-LCA), economic input–output LCA (EIO-LCA), and hybrid LCA (H-LCA). Bian et al. [10] established an inventory model for the production stage of PC components based on P-LCA and combined it with BIM to conduct a full life cycle evaluation. Chen et al. [11] applied the EIO-LCA method to develop an evaluation framework for low-carbon buildings, providing a systematic accounting process across different construction stages. Zhang et al. [12] compared the carbon emissions of three buildings, identified the limitations of both P-LCA and EIO-LCA, and proposed a hybrid approach to improve accuracy and consistency. However, most existing studies are still based on static, code-driven calculations, making it difficult to capture the impacts of policy interventions, feedback mechanisms, and temporal dynamics [13,14].

1.2.2. Research Progress on the Collaborative Optimization of Carbon Emissions and Costs

In the construction industry, efforts to reduce carbon emissions are often accompanied by rising costs. As a result, balancing environmental sustainability with economic feasibility has become a central issue in green building research. Existing studies have commonly adopted multi-objective optimization models and joint evaluations of life cycle cost (LCC) and life cycle assessment (LCA) to explore pathways for simultaneous cost control and carbon reduction. For example, Eleftheriadis et al. [15] proposed a multi-objective optimization approach based on BIM, integrating LCA with a genetic algorithm to achieve joint optimization of cost and emissions in reinforced concrete structures, and found that the two are often linearly correlated. Chen et al. [16] developed an LCSA model that incorporates environmental, economic, and social dimensions, providing a quantitative assessment of sustainability performance across the production, transportation, and storage stages of PC components. Heydari et al. [17] introduced a BIM-based integrated design framework in which carbon and cost assessments are embedded in the early design phase, thereby enhancing carbon–cost trade-offs at the decision-making stage. Although these studies have produced valuable insights into the coordination of carbon emissions and costs, limitations remain. Most work has focused on macro-level coordination across the building life cycle or at the project level, while micro-level modeling, especially for the production stage of prefabricated components, has been overlooked. Given the standardized and quantifiable nature of PC components, which enables reliable data collection on both carbon emissions and costs, this stage should serve as a key entry point for advancing research on coordinated optimization.

1.2.3. Application of SEM and SD Models in Building Carbon Emission Research

In recent years, SEM has been widely employed to analyze green building policies and identify key drivers of carbon emissions. SD, known for its strength in capturing temporal dynamics within complex systems, has frequently been used for policy scenario analysis and decision support. For example, Ping et al. [18] examined carbon emission drivers through an extended STIRPAT model and subsequently developed an SD framework to further explore the main influencing factors. Li et al. [19] utilized SEM to assess determinants of carbon emissions during the materialization phase of prefabricated buildings and provided theoretical validation. Although SEM and SD each offer distinct benefits, their combined application in the construction sector remains limited. In particular, systematic modeling and empirical studies addressing the joint optimization of carbon emissions and costs in prefabricated components are still scarce.

To facilitate comparison and method selection, this study summarizes commonly used approaches in existing research on carbon emissions in PC components and prefabricated buildings (

Table 1. Methods used in carbon emission studies.

Method	Main Features	Scope of Application	Advantages	Limitations	References
P-LCA	Inventory model constructed based on material and energy consumption data	Life cycle or stage-based carbon emission accounting for prefabricated buildings	Data are transparent and traceable, suitable for micro-level analysis	High data acquisition cost, static, and limited in capturing policy effects or temporal dynamics	[10]
EIO-LCA	Carbon emissions estimated based on economic input–output tables and sectoral emission coefficients	Macro-level accounting of building carbon emissions	Allows rapid estimation of carbon emissions for large-scale buildings or regions	Accuracy is affected by macro-level data and lacks detailed granularity	[11]
H-LCA	Integrates the advantages of P-LCA and EIO-LCA	Multi-level life cycle analysis	Improves the accuracy and comprehensiveness of accounting	Data integration is complex, and calculations remain primarily static	[12]
Multi-objective optimization/LCC-LCA collaboration	Considers both cost and carbon emissions as optimization objectives	Building life cycle or design stage	Enables joint optimization of carbon emissions and costs	Most studies focus on the macro level, lacking detailed modeling of the production stage	[15,17]
SEM	Causal relationship modeling based on statistical analysis	Identification of carbon emission drivers and policy analysis	Quantifies direct and indirect relationships, identifying key driving factors	Static analysis, limited in simulating temporal dynamics	[19]
SD	Simulates the dynamic behavior of systems over time	Policy scenario simulation for carbon emissions and decision making in complex systems	Captures feedback mechanisms and dynamic evolution, supporting forecasting	Difficult to directly quantify causal relationships and relies on empirical parameters	[18]

) and clarifies the rationale for adopting SEM and SD models.

This study, therefore, adopts an integrated framework that combines SEM and SD. SEM is applied to quantify both direct and indirect effects on carbon emissions during the production stage of prefabricated components, identify key drivers, and provide a rigorous basis for theoretical analysis and managerial decision making. SD, in contrast, is employed to simulate temporal variations in carbon emissions and system responses under alternative policy scenarios, thereby capturing feedback loops and complex interactions. Integrating SEM with SD allows not only the identification of causal pathways but also the execution of dynamic scenario simulations, offering a robust methodological foundation for establishing a comprehensive, systematic, and predictive framework aimed at optimizing carbon emissions and costs in the production stage of prefabricated components.

1.3. Research Contributions

To overcome existing limitations in the research on the coordinated optimization of carbon emissions and economic costs in the construction sector—such as the static nature of prevailing models, the absence of system feedback mechanisms, and insufficiently structured policy variables—this study centers on the production stage of prefabricated components and establishes a coupled SEM–SD model that integrates the strengths of structural identification with dynamic simulation. The overall framework of this study is illustrated in Figure 1.

The main contributions of this study are summarized as follows:

(1)

The SEM-SD approach is applied for the first time to the coordinated analysis of carbon emissions and costs in the production stage of PC components;

(2)

Key driving factors of carbon emissions and costs during the production stage are systematically identified and quantified;

(3)

Multi-scenario simulation strategies are developed, providing practical solutions for carbon reduction and cost control;

(4)

In the broader context of green building policies, the study offers scientific decision-making support and quantitative references for policymakers and enterprises to promote low-carbon transformation in the production of PC components.

2. Research Methods

2.1. Definition of Research Boundaries

Given that the production of PC components involves large-scale operations, numerous processes, complex construction, and a wide variety of materials and energy sources, this study defines the research boundaries of the production stage from a temporal perspective to enhance the accuracy and practicality of carbon accounting. Based on this, a staged emission analysis framework is developed. As shown in Figure 2, the production stage of PC components can be divided into the following three main phases from a temporal perspective:

(1)

The material production stage: This stage includes all processes related to the extraction of resources, industrial processing, manufacturing, and packaging of key raw materials, such as cement, steel, and aggregates, up until their delivery to the precast factory.

(2)

The material transportation stage: This stage covers the entire transportation process of raw materials from production sites or suppliers to the precast factory, considering factors such as transportation mode, distance, and fuel type.

(3)

The component fabrication stage: This stage refers to the forming, curing, storage, and assembly preparation of components within the precast factory, involving extensive operation of machinery and energy consumption.

2.2. SEM-SD Coupled Modeling Method

First, SEM is used to identify the key factors influencing carbon emissions, clarify the causal relationships between variables, and quantify the path coefficients. Based on this, a structural model is developed to determine the main drivers of carbon emissions during the production stage of PC components and to calculate their respective weights, providing quantitative input for the subsequent construction of the SD model equations. SEM effectively handles complex relationships between multiple variables and demonstrates strong reliability and validity [20], which helps to identify the significant paths and the strength of their effects.

Building on the SEM results, the identified key factors are incorporated into the SD framework to develop a system dynamics model that captures feedback mechanisms and dynamic evolution. The SD approach is particularly advantageous for simulating temporal changes, assessing the impacts of policy interventions, and performing scenario forecasting [21]. It enables the dynamic representation of interactions between carbon emissions and costs, while facilitating the evaluation of emission reduction pathways, cost optimization ranges, and long-term coordinated control under alternative policy scenarios.

The coupling of SEM and SD models integrates static causal identification with dynamic system simulation, providing a solid theoretical and technical foundation for thoroughly exploring the formation mechanisms and optimization pathways of carbon emissions during the production stage of PC components. The research roadmap is shown in Figure 3.

3. SEM

3.1. Development of Influencing Factor System

Through a review of relevant studies on carbon emissions and national policies related to energy conservation and emission reduction in prefabricated buildings, the key factors influencing carbon emissions during the production stage of PC components are identified using a literature-based analytical approach. Particular attention is given to the carbon footprint of each production activity and its associated emission sources. Five latent variables are selected: carbon emissions from material production (A), material transportation (B), component fabrication (C), production cost (D), and total carbon emissions during the production stage of prefabricated components (E). In total, 15 observed variables are defined, as presented in

Table 2. System of influencing factors.

Latent Variable	Observation Variable	Factor Explanation	References
Material production (A)	Type of raw material (A1)	Various raw materials used in the material production stage, such as steel and cement	[22,23,24,25,26]
	Type of energy consumption (A2)	Energy consumed during production, such as electricity and coal, which directly affects carbon emission intensity
	Material production process (A3)	The technical level and advancement of production processes determine energy consumption and emission efficiency
	Operational level of personnel (A4)	The skill level of operators affects production efficiency and resource waste rate
Material transportation (B)	Mode of transportation (B1)	Energy consumption varies across different transportation modes, such as heavy trucks and light trucks	[22,23,24,27]
	Fuel type (B2)	Different types of energy used by transportation vehicles, such as diesel and electricity
Component fabrication (C)	Types of mechanical equipment (C1)	The type, age, and technological level of production line equipment affect energy consumption	[9,22,23,24,28]
	Types of energy consumption (C2)	Types of energy consumed during component fabrication, such as electricity and fuel
	Personnel operation level (C3)	Operator skills affect equipment efficiency and the rate of defective products
	Capacity utilization rate (C4)	The ratio of actual output to designed capacity, influencing carbon emissions per unit of product
	Equipment efficiency (C5)	The energy efficiency of the equipment itself, with higher efficiency indicating lower emissions
Production cost (D)	Application of new technology (D1)	The application of BIM and low-carbon technologies, which directly affects costs	[29,30]
	Digital management (D2)	Refined management achieved through information systems, which may lead to increased costs
Total carbon emissions (E)	Policy norms and standards (E1)	National or local carbon reduction policies, regulations, and industry standards	[26,31]
	Climate, environmental, and geographical factors (E2)	The impact of natural conditions, such as temperature, humidity, and terrain, on carbon emissions

.

3.2. Basic Hypotheses for Model Construction

To examine the influence of latent variables on carbon emissions during the production stage of PC components, a hypothetical path model was developed based on relevant theories, prior studies, and the characteristics of PC component production. Accordingly, the following hypotheses are proposed:

H1. 

Carbon emissions from material production are positively associated with carbon emissions during the production stage of PC components.

H2. 

Carbon emissions from material transportation are positively associated with carbon emissions during the production stage of PC components.

H3. 

Carbon emissions from component fabrication are positively associated with carbon emissions during the production stage of PC components.

H4. 

The production costs of PC components are positively associated with carbon emissions during the production stage of PC components.

3.3. Analysis of Model Evaluation Indicators

The survey was conducted through a combination of online and offline methods. A total of 350 questionnaires were distributed to construction enterprises, prefabricated component manufacturers, building material suppliers, and relevant researchers. Among them, 326 responses were collected, of which 315 were valid, resulting in an effective response rate of over 90.0%. Invalid questionnaires were excluded mainly due to incomplete information or low consistency in responses. The overall sample was balanced in terms of respondent type, industry distribution, job position, and educational background, thereby ensuring a representative reflection of both industry practitioners and the academic community, as shown in

Table 3. Basic characteristics of the surveyed sample.

Project	Category	Count (n = 315)	Percentage (%)	Project	Category	Count (n = 315)	Percentage (%)
Type of organization	Construction	108	34.3	Education level	Doctorate or above	21	6.7
	Prefabricated component production	92	29.2		Master’s degree	95	30.2
	Building material supply	58	18.4		Bachelor’s degree	141	44.8
	Universities/research institutes	57	18.1		Associate degree or below	58	18.4
Position	Managers	87	27.6	Years of work experience	3 years or less	74	23.5
	Technical staff	119	37.8		3–5 years	83	26.3
	Researchers	71	22.5		5–10 years	82	26.0
	Others	38	12.1		More than 10 years	76	24.1

. The detailed structure of the questionnaire and examples of item design are provided in Appendix A (

Table A1. Survey of influencing factors system.

Module	Item Number	Item Content	Rating (1 = Very Unimportant, 5 = Very Important)
Carbon Emissions from Material Production (A)	A1	Impact of Raw Material Type on Carbon Emissions	1□ 2□ 3□ 4□ 5□
	A2	Impact of Energy Consumption Type on Carbon Emissions	1□ 2□ 3□ 4□ 5□
	A3	Impact of Material Production Process on Carbon Emissions	1□ 2□ 3□ 4□ 5□
	A4	Impact of Personnel Operation Level on Carbon Emissions	1□ 2□ 3□ 4□ 5□
Carbon Emissions from Material Transportation (B)	B1	Impact of Transportation Mode on Carbon Emissions	1□ 2□ 3□ 4□ 5□
	B2	Impact of Energy Consumption Type on Carbon Emissions	1□ 2□ 3□ 4□ 5□
Carbon Emissions from Component Manufacturing (C)	C1	Impact of Machinery and Equipment Type on Carbon Emissions	1□ 2□ 3□ 4□ 5□
	C2	Impact of Energy Consumption Type on Carbon Emissions	1□ 2□ 3□ 4□ 5□
	C3	Impact of Personnel Operation Level on Carbon Emissions	1□ 2□ 3□ 4□ 5□
	C4	Impact of Capacity Utilization on Carbon Emissions	1□ 2□ 3□ 4□ 5□
	C5	Impact of Equipment Efficiency on Carbon Emissions	1□ 2□ 3□ 4□ 5□
Production Cost (D)	D1	Impact of New Technology Adoption on Cost and Carbon Emissions	1□ 2□ 3□ 4□ 5□
	D2	Impact of Digital Management on Cost and Carbon Emissions	1□ 2□ 3□ 4□ 5□
Total Carbon Emissions (E)	E1	Impact of Policies, Regulations, and Standards on Carbon Emissions	1□ 2□ 3□ 4□ 5□
	E2	Impact of Climatic and Geographical Factors on Carbon Emissions	1□ 2□ 3□ 4□ 5□

).

3.3.1. Reliability Test

Reliability analysis was conducted to assess the stability and consistency of the measurement scale. As a key indicator for evaluating the validity of the questionnaire, it was crucial to perform a thorough analysis. In this study, the SPSS 27.0 was employed to systematically evaluate the reliability of the questionnaire. Cronbach’s α coefficient, widely recognized as a standard measure of internal consistency, was used to assess the reliability of the five latent variables. Higher α-values, approaching 1, indicate greater reliability of the collected data. The detailed results are presented in

Table 4. Reliability test indicators.

Latent Variable	Cronbach’s α	Testing Conditions	Items
Material production	0.875	A Cronbach’s α coefficient below 0.6 is generally considered indicative of insufficient internal consistency. Values between 0.7 and 0.8 suggest acceptable reliability, while coefficients ranging from 0.8 to 0.9 indicate a high level of scale reliability [32,33].	4
Material transportation	0.837		2
Component fabrication	0.893		5
Production cost	0.847		2
Total carbon emissions	0.871		2
Overall	0.868		15

.

Based on the reliability analysis, the Cronbach’s α-values for material production, material transportation, component fabrication, production cost, and total carbon emissions were 0.875, 0.837, 0.893, 0.847, and 0.871, respectively, with an overall questionnaire Cronbach’s α of 0.868. All values fell within the range of 0.8 to 0.9, indicating good reliability for each variable.

3.3.2. Validity Test

Validity analysis evaluates the degree to which a measurement instrument accurately captures the intended constructs. Observed values are meaningful only if the items within a given set adequately represent the same theoretical construct from the respondents’ perspective. To assess construct validity, this study employed both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA).

(1)

Exploratory Factor Analysis

The questionnaire data were subjected to the Kaiser–Meyer–Olkin (KMO) measure and Bartlett’s test of sphericity to assess the suitability for factor analysis. The results are presented in

Table 5. KMO and Bartlett’s test.

KMO and Bartlett Values			Testing Conditions
KMO value		0.820	Construct validity is considered acceptable when the KMO measure exceeds 0.6 and Bartlett’s test of sphericity yields a p-value below 0.05 [34,35,36].
Bartlett’s sphericity test	Approximate chi-square	2528.784
	df	105
	p-value	0.000

. The KMO value was 0.820, and Bartlett’s test produced an approximate chi-square value of 2528.784 with 105 degrees of freedom and a significance level of 0.000, which is below 0.05. These results indicate that the measured variables met the requirements for KMO and Bartlett’s tests, demonstrating suitability for factor analysis and confirming their statistical significance.

(2)

Confirmatory Factor Analysis (CFA)

Confirmatory factor analysis (CFA) was performed to assess the extent to which the latent variables were represented by their observed indicators and to verify whether the measurement items satisfied the standards of the structural equation model. The measurement model was evaluated using AMOS software for structural equation modeling.

Two key indices were examined in the CFA: composite reliability (CR) and average variance extracted (AVE). CR reflects the internal consistency of each construct, whereas AVE measures the average proportion of variance in the observed variables explained by the latent construct. Higher AVE values indicate stronger convergent validity. A CR value above 0.7 is generally deemed acceptable, with higher values indicating better reliability. Factor loadings are expected to fall between 0.6 and 0.95 [37]. As presented in

Table 6. Convergence validity test.

Variable	Measurement Item	Standard Loading Coefficient	SE	CR	p	CR	AVE
Material production (A)	A1	0.861				0.877	0.642
	A2	0.770	0.057	15.496	***
	A3	0.811	0.060	16.632	***
	A4	0.759	0.065	15.205	***
Material transportation (B)	B1	0.854				0.838	0.721
	B2	0.844	0.112	9.202	***
Component fabrication (C)	C1	0.813				0.893	0.626
	C2	0.828	0.064	16.289	***
	C3	0.764	0.061	14.701	***
	C4	0.766	0.058	14.748	***
	C5	0.782	0.058	15.142	***
Production cost (D)	D1	0.864				0.847	0.734
	D2	0.850	0.096	10.122	***
Total carbon emissions (E)	E1	0.852				0.873	0.775
	E2	0.908	0.092	12.588	***

Note. *** p <0.001 [37].

and Figure 4, all standardized factor loadings exceeded 0.60 and were statistically significant (p < 0.05), indicating robust measurement relationships. Furthermore, all AVE values were above 0.50, and all CR values surpassed 0.70, confirming satisfactory convergent validity of the constructs.

Discriminant validity assesses the degree to which measures of distinct constructs are truly independent, reflecting their capacity to capture unique concepts. As reported in

Table 7. Discriminant validity test.

Variable	Material Production	Material Transportation	Component Fabrication	Production Cost	Total Carbon Emissions
Material production	0.801
Material transportation	0.397	0.849
Component fabrication	0.265	0.31	0.791
Production cost	0.414	0.22	0.29	0.857
Total carbon emissions	0.465	0.356	0.315	0.42	0.88

, the diagonal values (i.e., the square roots of each construct’s AVE) exceed the correlations between that construct and all other constructs, indicating that the measurement scales exhibit satisfactory discriminant validity.

3.4. Path Coefficient and Hypothesis Testing

3.4.1. SEM Fit Analysis

The structural equation model was evaluated using the AMOS 29 software to assess model fit among the five variables. As presented in

Table 8. Results of the goodness-of-fit test for the structural equation model.

Common Indicators	χ2/df	GFI	AGFI	IFI	TLI	CFI	RMSEA
Statistical value	1.432	0.953	0.930	0.986	0.982	0.986	0.037
Reference value	<3	>0.8	>0.8	>0.9	>0.9	>0.9	<0.08
Achievement status	Meets the standards	Meets the standards	Meets the standards	Meets the standards	Meets the standards	Meets the standards	Meets the standards

, all fit indices met or exceeded recommended thresholds: χ²/df = 1.432 (<3), GFI = 0.953 (>0.9), AGFI = 0.930 (>0.9), IFI = 0.986 (>0.9), TLI = 0.982 (>0.9), CFI = 0.986 (>0.9), and RMSEA = 0.037 (<0.08). These results collectively demonstrate excellent model fit according to established standards.

3.4.2. Path Coefficient Test

To examine the hypothesized relationships, we evaluated path coefficients (β) and their statistical significance (see

Table 9. Test results of the structural equation model.

Variable Relationship	Unstandardized Regression Coefficient	Standardized Regression Coefficient	SE	CR	p
Material production → carbon emissions	0.270	0.272	0.069	3.908	***
Material transportation → carbon emissions	0.142	0.157	0.060	2.364	0.018
Component fabrication → carbon emissions	0.116	0.126	0.056	2.055	0.040
Production cost → carbon emissions	0.222	0.237	0.063	3.512	***
Material production → A1	1.000	0.861
Material production → A2	0.891	0.770	0.057	15.496	***
Material production → A3	1.002	0.811	0.060	16.632	***
Material production → A4	0.983	0.759	0.065	15.205	***
Material transportation → B1	1.000	0.854
Material transportation → B2	1.035	0.844	0.112	9.202	***
Component fabrication → C1	1.000	0.813
Component fabrication → C2	1.044	0.828	0.064	16.289	***
Component fabrication → C3	0.898	0.764	0.061	14.701	***
Component fabrication → C4	0.852	0.766	0.058	14.748	***
Component fabrication → C5	0.871	0.782	0.058	15.142	***
Production cost → D1	1.000	0.864
Production cost → D2	0.969	0.850	0.096	10.122	***
Carbon emissions → E1	1.000	0.852
Carbon emissions → E2	1.162	0.908	0.092	12.588	***

Note. *** p <0.001.

and Figure 5). The analysis yielded four key findings:

Material production emissions significantly increased PC component production emissions (β = 0.272; p < 0.001), supporting H1.

Material transportation emissions positively influenced component production emissions (β = 0.157; p = 0.018), confirming H2.

Component manufacturing emissions had a significant positive influence on production emissions (β = 0.126; p = 0.040), validating H3.

Production cost reduction was associated with higher production emissions (β = 0.237; p < 0.001), substantiating H4.

3.4.3. Analysis of SEM Results

Based on the analysis of the model evaluation metrics, path coefficients, and the summary of hypothesis tests, the structural equation model developed in this study demonstrates a high level of fit, with the overall adequacy considered satisfactory. All four hypotheses proposed in Section 3.2 are supported and closely align with the observed data.

The results show that, among the primary indicators, the path coefficient for the material production stage is markedly higher than those of the other stages, dominating the path analysis and underscoring its critical influence on carbon emissions during PC component production. At the level of individual observed variables, carbon emissions in this stage primarily stem from the types of raw materials used, energy consumption patterns, and the efficiency of production processes. These factors directly determine the energy required per unit of component output and the corresponding carbon emission factors. As a result, the material production stage not only accounts for the largest share of total carbon emissions but also exhibits the highest carbon intensity and the greatest potential for emission reduction throughout the component’s lifecycle.

At the level of secondary indicators, policy standards, climate and geographic conditions, types of raw materials, and the application of new technologies were identified as the core drivers of carbon emissions. Policy standards impose constraints on enterprise behavior by regulating emission limits and guiding technology choices. Climate and geographic conditions affect raw material transportation routes, energy use patterns, and building performance requirements. The type of raw materials determines the baseline level of carbon emissions and imposes strong structural constraints, while the application of new technologies provides key opportunities for improving energy efficiency and reducing emissions. The combined effects of these core variables shape the system characteristics of carbon emissions during PC component production and provide theoretical support and parameter references for simulating emission reduction pathways under different scenarios.

3.5. Impact Factor Weighting Assessment

Using the results of the SEM (Table 9), the estimated path coefficients were employed to determine objective weights for each factor via the correlation coefficient method. The path coefficients were first normalized, and the resulting weights for the factors influencing carbon emissions during the PC component production stage were calculated. The detailed findings are summarized in

Table 10. Weighting table of influencing factors.

Latent Variable	Assigned Value	Observation Variable	Assigned Value
Material production	0.343	A1	0.092
		A2	0.083
		A3	0.087
		A4	0.081
Material transportation	0.198	B1	0.100
		B2	0.098
Component fabrication	0.159	C1	0.033
		C2	0.033
		C3	0.031
		C4	0.031
		C5	0.031
Production cost	0.299	D1	0.151
		D2	0.148

.

(1)

Weight Assignment for Latent Variables

$$P_i={\displaystyle \frac{X_i}{{\displaystyle {\sum }_{i=1}^4{X}_i}}}$$

(1)

In the formula, the weight of the latent variable is denoted as $P_i$, and the overall influence degree is represented by $X_i$.

(2)

Weight Assignment for Observed Variables

$$q_{ij}={\displaystyle \frac{x_{ij}}{{\displaystyle {\sum }_{j=1}^n{x}_{ij}}}}$$

(2)

$$p_{ij}=P_i \times q_{ij}$$

(3)

In the formula, the path coefficient after normalization is denoted by $q_{ij}$, the path coefficient of the observed variable to the latent variable is represented by $x_{ij}$, and the weight of the observed variable is indicated by $p_{ij}$.

The latent and observed variables influencing carbon emissions during the PC component production stage were systematically assigned and analyzed. This study quantitatively assessed the specific contributions of each stage to overall carbon emissions, providing parameter settings and data support for the construction of equations in the subsequent SD model.

4. SD Model

4.1. Construction of Causal Loop Diagram

By systematically examining the influencing factors and interaction pathways between carbon emissions and costs in the production stage of PC, a system dynamics causal loop diagram was developed. From a systems perspective, this diagram reveals the causal linkages and feedback mechanisms among variables. It primarily comprises five core subsystems: carbon emissions from material production, carbon emissions from material transportation, carbon emissions from component manufacturing, total production cost, and total carbon emissions during production. In addition, external driving factors, such as policy interventions and carbon taxes, were incorporated. This causal loop diagram provides the structural foundation for subsequent system simulation modeling and offers theoretical support for identifying key leverage variables and optimizing policy pathways.

Vensim 10.2.2 was used to develop the causal loop diagram illustrating the synergistic optimization mechanism of carbon emissions and costs in PC production, as depicted in Figure 6. The arrows indicate the direction of causal influence, while “+” and “−” denote positive and negative feedback relationships, respectively.

4.2. Construction of System Flow Diagram

The causal loop diagram, a qualitative analysis tool, is employed to identify and depict the interactive relationships between internal variables within a system. Nevertheless, it lacks the capabilities for quantitative analysis and dynamic simulation, making it difficult to be used for predicting system behavior and simulating decision-making processes. To more clearly illustrate the logical relationships, feedback structures, and control rules among the sub-elements in the PC component carbon emission system, and to take into account the operability of system simulation analysis as well as the accessibility of carbon emission and cost data, a corresponding system flow diagram model was constructed in this study based on the causal loop diagram, as shown in Figure 7. Boxes represent state variables, which describe the accumulated effects of the system; triangles represent rate variables, which depict the changes in the system’s accumulated effects. Other symbols serve as auxiliary variables or constants, connecting state and rate variables to ensure smooth information flow and transformation. Arrows are used to indicate the connections between these elements.

4.3. System Variable Setting and Equation Construction

4.3.1. System Variable Setting

Among the various types of PC components, prefabricated composite slabs were selected as the focus of this study due to their wide range of applications and typical production processes. Studies by Zhang et al. [38] indicate that the carbon emissions per unit volume of prefabricated composite slabs are higher than those of other components. In this study, data were collected through on-site investigation at a prefabrication plant in Wuhan, covering the period of 2014–2024. The data included raw material consumption (e.g., cement, sand, stone, and steel), production energy use (electricity, water, and fuel), transportation distances and modes, labor, and production costs.

In accordance with the design standards “Technical Standard for Prefabricated Concrete Buildings” GB/T 51231-2016 and “Code for Design of Concrete Structures” GB 50010-2010 [39,40], a prefabricated composite slab measuring 3000 mm × 1200 mm × 60 mm (length × width × height) was selected as the representative study unit. This size is typical for prefabricated composite slabs, and the detailed parameters are provided in

Table 11. Main parameter table.

Stage	Kind	Consumption Rate	Unit
Material production stage	Cement	82.1	kg
	Medium sand	155.5	kg
	Gravel (5–20 mm)	226.8	kg
	Water	41	kg
	Reinforcing steel	25.92	kg
	Admixture	1.23	kg
	Release agent	0.06	kg
	Energy	1.5	kwh
	Manual labor	0.1	person-day
Material transportation stage	Diesel fuel	0.07	kg
Component fabrication stage	Electricity	1.2	kwh
	Diesel fuel	0.07	kg
	Methanol	2.68	kg
	Manual labor	0.13	person-day

. The table lists the required materials, labor, and energy per slab based on standard values, which may be adjusted to reflect actual conditions.

The selection of carbon emission factors for the production stage should adhere to the principles of carbon accounting, component-type differentiation, and regional energy structures, and must be based on reliable data sources. In this study, carbon emission factors for materials, energy, machinery, and labor were drawn from the “IPCC Guidelines for National Greenhouse Gas Inventories,” the GB/T 51366-2019 “Standard for Calculating Building Carbon Emissions,” the “Study on CO₂ Emission Factors of Regional Power Grids in China (2023),” the CLCD database, and other relevant sources [41,42,43,44]. The detailed values are summarized in

Table 12. Major carbon emission factors.

Type	Kind	Carbon Emission Factor	Unit
Material	Cement	735	kgCO2e/t
	Sand	2.51	kgCO2e/t
	Gravel (5–20mm)	2.18	kgCO2e/t
	Water	0.168	kgCO2e/t
	Reinforcing steel	2340	kgCO2e/t
	Admixture	1164	kgCO2e/t
	Release agent	2081	kgCO2e/t
Machinery	Heavy-duty diesel truck	0.129	kgCO2e/(t·km)
	Light-duty diesel truck	0.286	kgCO2e/(t·km)
Energy	Electricity	0.608	kgCO2e/(kW·h)
	Diesel fuel	3.16	kgCO2e/kg
	Methanol	2.54	kgCO2e/kg
Manual labor	Manual labor	6.64	kgCO2e/(person-day)

.

4.3.2. Construction of Equations

To quantitatively simulate the dynamic evolution of carbon emissions and costs, it is necessary to define the mathematical relationships among the relevant variables, thereby establishing a foundational set of system equations. These equations provide the core logical framework for the simulation and capture the causal relationships and feedback mechanisms among the subsystems. The primary formulas are presented as follows:

(1)

Carbon Emission System

$$\begin{array}{l}{E}_{\mathrm{material}}={\displaystyle \sum _{i=1}^n{M}_i}{D}_i+{\displaystyle \sum _{i=1}^j{F}_i}{J}_i{P}_i+{\displaystyle \sum _{i=1}^m{A}_{i}S}\\ E_{\mathrm{transport}}={\displaystyle \sum _{i=1}^n{M}_i}{Y}_i{V}_i+{\displaystyle \sum _{i=1}^j{F}_i}{J}_i{P}_i+{\displaystyle \sum _{i=1}^m{A}_{i}S}\\ E_{\mathrm{production}}={\displaystyle \sum _{i=1}^j{F}_i}{J}_i{P}_i+{\displaystyle \sum _{i=1}^m{A}_{i}S+{\displaystyle \sum _{m=1}^k{\displaystyle \sum _{i=1}^h\left(T_{mi}{L}_i\right)}}}\end{array}$$

(4)

In the formula, $E_{\mathrm{material}}$ represents the carbon emission volume during the material production stage; $E_{\mathrm{transport}}$ represents the carbon emission volume during the material transportation stage; $E_{\mathrm{production}}$ represents the carbon emission volume during the component-manufacturing stage; $M_i$ represents the usage quantity of raw material $i$; $D_i$ represents the carbon emission factor of raw material $i$ $[\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_2\mathrm{e}/{\mathrm{m}}^3]$; $n$ represents the types of raw materials; $Y_i$ represents the transportation distance of raw material $i$; $V_i$ represents the carbon emission factor of the transportation vehicle $[\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_2\mathrm{e}/(\mathrm{t}\cdot \mathrm{k}\mathrm{m})]$; $F_i$ represents the energy consumption per unit shift of the construction machinery of raw material $i$; $J_i$ represents the number of shift units of the construction machinery of raw material $i$; $P_i$ represents the carbon emission factor of the energy consumed by the construction machinery of raw material $i$ $[\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_2\mathrm{e}/\mathrm{k}\mathrm{g}]$; $j$ represents the type of construction machinery; $j$ represents the carbon emission factor per person-hour of labor $[\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_2\mathrm{e}/(\mathrm{person-day})]$; $A_i$ represents the labor hours consumed by the $i$ process; $m$ represents each single unit of each equipment; $T_{mi}$ represents the number of shift units of the $i$ type of machinery required for the production of $m$ components; and $L_i$ represents the carbon emission factor of each type of machinery $[\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_2\mathrm{e}/\mathrm{shift}\mathrm{unit}]$.

$$R\mathrm{a}teE(t)=\left[E_{material}+E_{transport}+E_{production}\right]\times Y(t)$$

(5)

In the formula, $R\mathrm{a}teE(t)$ represents the total carbon emission rate, and $Y(t)$ represents the annual output.

$$TotalE(t)={\displaystyle {\int }_{t_0}^{t}R\mathrm{a}teE(t)dt}+Total{E}_0$$

(6)

In the formula, $TotalE(t)$ represents the cumulative total carbon emission value, and $Total{E}_0$ represents the initial value.

(2)

Cost System

$$R\mathrm{a}teC(t)=\left[C_{material}+C_{transport}+C_{production}\right]\times Y(t)$$

(7)

In the formula, $R\mathrm{a}teC(t)$ represents the total cost rate, and $Y(t)$ represents the annual output.

$$TotalC(t)={\displaystyle {\int }_{t_0}^{t}R\mathrm{a}teC(t)dt}+Total{C}_0$$

(8)

In the formula, $TotalC(t)$ represents the cumulative total cost value, and $Total{C}_0$ represents the initial value.

To comprehensively account for the multidimensional factors influencing carbon emissions and costs, the model incorporates observed variables such as equipment efficiency, operational level, and policy environment. Detailed definitions, parameter settings, and procedures are provided in Appendix A,

Table A2. Main equations list.

Variable	Equation	Unit
Total carbon emissions during the production stage	INTRG (total carbon emission change rate, 9.89791 × 106)	kgCO2e
Total carbon emission change rate	(Carbon emissions during the material production stage + carbon emissions during the material transportation stage + carbon emissions during the component manufacturing stage) × production volume × (1 − carbon tax) × EXP (-green building policy)	kgCO2e
Change rate–CO2e (material production)	(Raw materials-CO2e + manual labor-CO2e (production) + energy-CO2e (production)) × (1 − production process × 0.087-personnel operational proficiency (production) × 0.081 − types of raw materials × 0.092 − types of energy consumption (production) × 0.083)	kgCO2e
Raw materials–CO2e	Admixtures-CO2e + cement-CO2e + sand-CO2e + gravel-CO2e + release agent-CO2e + steel bar-CO2e	kgCO2e
Manual labor–CO2e (production)	Manual labor-EF × person-day (production)	kgCO2e
Electricity–CO2e (production)	Electricity consumption (production) × electric-EF
Change rate–CO2e (material transportation)	(Energy-CO2e (transportation) + vehicle-CO2e) × (1 − types of energy consumption (transportation) × 0.098 − mode of transportation × 0.1)	kgCO2e
Energy–CO2e (transportation)	Diesel-EF × fuel consumption	kgCO2e
Vehicle–CO2e	(Light-duty truck-EF × the carrying capacity of a light truck + heavy-duty truck-EF × the carrying capacity of heavy trucks) × transportation distance	kgCO2e
Change rate–CO2e (component fabrication)	(1-personnel operational proficiency (fabrication) × 0.031 − capacity utilization rate × 0.031 − equipment efficiency × 0.031-type of mechanical equipment × 0.033 − types of energy consumption (fabrication) × 0.033) × (manual labor-CO2e (fabrication) + energy-CO2e (fabrication))	kgCO2e
Manual labor–CO2e (fabrication)	Manual labor-EF × person-day (fabrication)	kgCO2e
Energy–CO2e (fabrication)	Electricity-CO2e (fabrication) + diesel-CO2e + methanol-CO2e	kgCO2e
Total production cost	INTRG (cost change rate, 3.11292 × 107)	CNY
Cost change rate	(1 + digital management × 0.148 + application of new technologies × 0.151) × production cost × (1 + (1 − EXP(-green building policy))) × production volume + total carbon emissions during the production stage × carbon tax	CNY

Note: In the figure, CO₂e represents the abbreviation for carbon emissions, and EF represents the abbreviation for carbon emission factor.

. The coefficient values in the table were determined based on the assignment of latent and observed variables for the influencing factors presented in Table 10.

4.4. Model Validation

4.4.1. Historical Testing

Historical validation is primarily used to assess a model’s ability to reproduce actual historical behavior, that is, whether the model can accurately simulate the operation of a real system over past periods. In this study, the carbon emissions and cost data for composite slab production from 2018 to 2024 at a PC component plant in Wuhan were used as a reference. The comparison shows that the model outputs closely follow historical trends, with the mean absolute percentage error of key output variables kept below 5%, confirming the model’s effectiveness and accuracy in representing the historical behavior of the real system, as shown in

Table 13. Historical test table.

Year	Total Carbon Emissions/kgCO2e			Total Cost/CNY
	True Value	Simulated Value	Absolute Error	True Value	Simulated Value	Absolute Error
2014	9,898,394.57	9,897,910	0.00%	31,129,600	31,129,200	0.00%
2015	20,099,805.46	19,795,820	1.51%	62,891,210	62,258,436	1.01%
2016	31,187,849.14	29,996,764	3.82%	97,117,220	94,020,248	3.19%
2017	42,978,790.19	41,084,256	4.41%	133,505,800	128,246,480	3.94%
2018	55,631,552.97	52,874,688	4.96%	172,855,600	164,634,880	4.76%
2019	68,728,620.51	65,586,824	4.57%	213,912,630	203,984,832	4.64%
2020	81,347,472.97	78,683,224	3.28%	254,251,990	245,041,984	3.62%
2021	93,421,201.36	91,301,440	2.27%	293,430,190	285,380,800	2.74%
2022	104,978,938.25	103,374,544	1.53%	331,625,690	324,559,136	2.13%
2023	116,221,968.81	114,931,600	1.11%	369,365,210	362,754,816	1.79%
2024	127,430,272.81	126,173,984	0.99%	407,980,650	400,493,984	1.84%

.

4.4.2. Sensitivity Testing

To evaluate the model’s responsiveness to changes in input parameters [45] and to examine whether the variables exhibit stable dynamic evolution, thereby ensuring the rationality of the model structure and the absence of abnormal feedback paths [46,47], a sensitivity analysis was conducted on selected key parameters. The analysis was performed by adjusting parameter values within defined ranges and varying the simulation step size to observe the trends of critical state variables. Initial state variables were assigned the baseline values specified in the model. The simulation step size was set to 1, 0.5, and 0.25, respectively. Raw material types, transportation modes, and capacity utilization were selected as observation variables, and their reassigned values were uniformly increased by 50% to examine the corresponding changes in total carbon emissions. As shown in Figure 8, the results indicate that when input parameters vary within a certain range, the output remains reasonably stable without significant fluctuations or unrealistic trends, thereby confirming the credibility of the model structure.

4.5. Simulation and Result Analysis

4.5.1. Baseline Scenario Prediction

Based on the calibrated and validated system dynamics model, a baseline scenario was established to forecast carbon emission trends. This scenario assumes that, in the absence of additional interventions or technological upgrades, all stages maintain their current trajectories. As shown in Figure 9, the model simulates the dynamic evolution of carbon emissions and production costs from 2014 to 2030. The results indicate that during the period of 2020–2024, overall carbon emissions and total costs experienced a temporary decline due to market fluctuations and a short-term reduction in PC component output. However, from a long-term perspective, despite intermittent fluctuations in the early years, both carbon emissions and production costs exhibited a steady upward trend throughout the simulation period.

4.5.2. Scenario Simulation

This study explores optimal control strategies for carbon emissions and production costs during the production stage under multiple scenarios, including green building policies, carbon tax adjustments, material optimization, energy optimization, new technological solutions, and digital management. The specific scenarios are summarized in

Table 14. Scenario simulation table.

Simulation Plan	Situation	Parameter Adjustment	References
Green policies	P1	Green policies have an investment of 10% to 20%	[48]
	P2	Green policies have an investment of 20% to 30%
Carbon tax mechanism	P3	Carbon tax: CNY 50 per ton	[49,50]
	P4	Carbon tax: CNY 100 per ton
	P5	Carbon tax: CNY 150 per ton
Policy–carbon tax synergistic approach	P6	Green policy investment: 10–20%; carbon tax: CNY 50 per ton	[48,49]
	P7	Green policy investment: 10–20%; carbon tax: CNY 100 per ton
	P8	Green policy investment: 10–20%; carbon tax: CNY 150 per ton
Material structure	P9	Raw material assignment: 3	[50]
Energy structure	P10	Energy consumption assignment: 3	[50]
Technical structure	P11	Material production process, transportation method, type of machinery equipment, capacity utilization rate, equipment efficiency, and application of new technologies: 3	[48,50]
Management structure	P12	Personnel operational proficiency and digital management score: 3	[50]

. Each scenario is grounded in a solid theoretical basis and supported by the relevant literature to ensure the design is both reasonable and scientifically robust.

Specifically, different carbon tax scenarios (e.g., CNY 50, 100, and 150 per ton) represent various policy stages, progressing from low initial rates to medium and high levels to gradually achieve emission reduction targets. Stern et al. [48] indicated that the design of carbon taxes should consider not only the economic capacity of a country or region but also the urgency of emission reduction goals, with tax rates being increased progressively. In addition, the level of investment in green building policies (10–30%) and adjustments to other factors, such as material optimization, energy structure improvements, and the application of new technologies, have varying impacts across scenarios. These settings are supported by the relevant literature and informed by practical implementation considerations.

(1)

Adjustment of Green Building Policies

To investigate the impact of green policy investment on carbon emission control and production costs, simulations were conducted under different scenarios representing varying levels of policy intervention. Figure 10 illustrates the projected trends in carbon emissions and costs under these policy scenarios in 2030, with detailed numerical results provided in

Table A3. Simulation results of total carbon emissions under different scenarios.

Situation	2025	2026	2027	2028	2029	2030
P	137,382,000	149,121,000	161,393,000	174224,000	187,614,000	201,573,000
P1	137,382,000	148,004,000	159,108,000	170718,000	181,681,000	193,109,000
P2	137,382,000	146,993,000	157,040,000	167546,000	177,465,000	187,806,000
P3	136,821,000	147,974,000	159,632,000	171822,000	184,542,000	197,803,000
P4	136,261,000	146,826,000	157,871,000	169419,000	181,47,0000	194,033,000
P5	13,5701,000	145,679,000	156,110,000	167017,000	178,398,000	190,263,000
P6	136,821,000	146,913,000	157,461,000	168491,000	178,905,000	189,762,000
P7	136,261,000	144,698,000	153,517,000	162738,000	173,918,000	185,574,000
P8	135,701,000	144,289,000	153,267,000	162654,000	171,086,000	179,876,000
P9	135,162,000	144,580,000	154,429,000	164729,000	175,478,000	186,685,000
P10	134,693,000	143,617,000	152,946,000	162701,000	172,881,000	183,493,000
P11	134,481,000	143,182,000	152,277,000	161788,000	171,713,000	182,059,000
P12	135,376,000	145,017,000	155,099,000	165642,000	176,645,000	188,115,000

and

Table A4. Simulation results of production costs under different scenarios.

Situation	2025	2026	2027	2028	2029	2030
P	439,109,000	479,988,000	523,155,000	568,638,000	616,462,000	666,653,000
P1	439,109,000	481,933,000	527,154,000	574,802,000	626,960,000	681,700,000
P2	439,109,000	487,398,000	538,390,000	592,118,000	652,337,000	715,536,000
P3	445,418,000	493,137,000	543,704,000	597,168,000	653,583,000	713,001,000
P4	451,727,000	506,231,000	564,082,000	625,352,000	690,117,000	758,455,000
P5	458,035,000	519,269,000	584,289,000	653,188,000	726,064,000	803,015,000
P6	445,418,000	497,028,000	551,649,000	609,333,000	674,251,000	742,485,000
P7	451,727,000	508,509,000	568,551,000	631,920,000	696,951,000	765,513,000
P8	458,035,000	523,159,000	592,078,000	664,879,000	745,770,000	830,722,000

in Appendix A. The results indicate the following.

Scenario 1: When green policy investment is adjusted within the range of 10–20%, carbon emissions are projected to decrease by 4.2% by 2030, while costs increase by 2.21%. This indicates that moderate policy intervention can achieve a certain level of emission reduction with relatively mild cost implications, demonstrating acceptable economic feasibility.

Scenario 2: When green policy investment is adjusted within the range of 20–30%, carbon emissions are projected to decrease by 6.83%, with costs increasing by 6.83% by 2030. Compared with Scenario 1, the emission reduction effect improves significantly, but the marginal benefits of reduction begin to diminish, and the economic burden increases noticeably. This suggests that as policy intensity rises, the pressure on enterprise operating costs also increases.

Overall, the intensity of green policy intervention exhibits a positive correlation and trade-off effect on carbon emissions and costs. Moderate policy measures can achieve an initial balance between emission reduction and cost control, whereas high-intensity interventions can accelerate the reduction process but also impose additional cost burdens. Therefore, when designing green development strategies, it is essential to carefully balance environmental benefits with economic capacity and to identify the optimal level of policy investment.

(2)

Adjustment of Carbon Tax Mechanism

To evaluate the effectiveness of the carbon tax mechanism in controlling carbon emissions during the PC component production stage and its impact on production costs, three simulation scenarios with different tax rates were established. The simulation results are presented in Figure 11.

Scenario 3: With a carbon tax of CNY 50/ton, carbon emissions are projected to decrease by 1.87% by 2030 compared with the baseline scenario, while production costs increase by 6.5%.

Scenario 4: With a carbon tax of CNY 100/ton, the projected reduction in carbon emissions rises to 3.74%, and costs increase to 12.1%, indicating an increasing effectiveness of emission reduction under higher tax rates.

Scenario 5: With a carbon tax of CNY 150/ton, carbon emissions are expected to decrease by 5.61%, but production costs rise further to 16.98%.

Overall, the carbon tax mechanism is effective in curbing carbon emissions, but its impact is primarily transmitted through increased production costs, resulting in limited emission reduction and a significant economic burden. Relying solely on carbon taxes may impose substantial financial pressure on enterprises, potentially reducing their motivation for green transformation. Therefore, carbon taxes should be implemented as a price-based regulatory tool in combination with guiding measures, such as green policies, to achieve a dynamic balance between emission reduction targets and economic feasibility.

(3)

Collaborative Path of Policy–Carbon Tax

To further investigate the synergistic effects of green policies and the carbon tax mechanism on emission reduction, three combined scenarios (Scenarios 6–8) were established. In these scenarios, green building policy investment was maintained at 10–20%, while different carbon tax rates (CNY 50, 100, and 150/ton) were applied to examine the dynamic interactions between emission control and cost variations. The simulation results are presented in Figure 12 and

Table 15. Simulation results of policy–carbon tax synergistic pathway.

Situation	Policy Mix	Carbon Emission Range	Cost Range	Note
P6	Policy: 10–20% + carbon tax: CNY 50 per ton	−5.86%	+10.21%	Moderate emission reduction, with good cost control
P7	Policy: 10–20% + carbon tax: CNY 100 per ton	−7.94%	+12.91%	Significant emission reduction, with acceptable costs
P8	Policy: 10–20% + carbon tax: CNY 150 per ton	−10.76%	+19.75%	The strongest in terms of emission reduction, but faces excessive cost pressure

.

Through a comparative analysis of the simulation results of the above-mentioned collaborative regulation scenarios, it can be seen that in Scenario 7, a relatively strong carbon emission reduction effect (7.94%) is expected to be achieved in 2030, while the cost increase is controlled within 12.91%, taking into account both environmental goals and economic affordability. In contrast, although the cost in Scenario 6 is lower, the carbon emission reduction range is limited. In Scenario 8, although the emission reduction is the most significant, the cost burden rises significantly, which may have a greater impact on enterprise operations. Therefore, considering both carbon emission reduction benefits and economic constraints, Scenario 7 can be regarded as the optimal policy combination path, providing a feasible basis for subsequent green transformation policies.

(4)

Comparative Analysis of Carbon Emission Impacts of the “Materials–Energy–Technology–Management” Four-Mechanism Approach

To further clarify the carbon emission reduction contributions of different emission reduction mechanisms during the prefabricated component production stage, this paper selects four typical mechanism paths: material structure optimization, energy structure adjustment, new technology application, and digital management. The simulation results are shown in Figure 13.

Scenario 9 (material structure optimization): By optimizing raw material selection and structural design, carbon emissions are projected to decrease by 1.85%, indicating that this pathway has a limited impact on emission reduction. However, it offers advantages such as low implementation difficulty and high adaptability, making it suitable as a foundational mitigation measure.

Scenario 10 (energy structure adjustment): By increasing the proportion of clean energy, carbon emissions are projected to decrease by 8.97%, suggesting that energy structure transformation is one of the more effective reduction pathways at the current stage, particularly in the context of industrialized construction, where it holds strong practical significance.

Scenario 11 (technology structure adjustment): By comprehensively enhancing material production processes, transportation methods, equipment efficiency, and capacity utilization, this pathway achieves the most significant emission reduction among all mechanisms, with carbon emissions projected to decrease by 9.68%. These results indicate that technological measures play a critical role in supporting the transition to low-carbon production.

Scenario 12 (management structure adjustment): By improving operational skills and digital scheduling capabilities, carbon emissions are projected to decrease by 6.68%. Although slightly lower than the reductions achieved through energy structure adjustment and technological pathways, digital management offers continuous optimization and managerial empowerment, representing a key approach for long-term carbon control.

Overall, the performance of different mechanism pathways varies significantly. Technology application and energy structure optimization demonstrate the highest effectiveness, whereas digital management and material structure optimization, while less impactful, offer strong potential for widespread adoption and integration. Future strategies should consider multi-mechanism integration to create a synergistic emission reduction system combining structural optimization, energy substitution, technological advancement, and management enhancement.

5. Discussion

5.1. Result Analysis and Discussion

5.1.1. Analysis and Discussion of SEM Results

This study conducted SEM analysis across four dimensions—material production, material transportation, component manufacturing, and production costs—to systematically identify the main pathways of carbon emissions during the PC component production stage. Reliability and validity tests confirmed that the variable settings were reasonable, and the model structure was stable. The results indicate that the material production stage has the most significant direct impact on carbon emissions, highlighting the decisive role of raw material selection, which is consistent with the findings of Bao et al. [51]. At the secondary indicator level, policy standards, climatic and geographic conditions, raw material types, and new technology applications were identified as the core factors influencing carbon emissions. Raw material attributes determine unit emission factors, while technological advancements contribute to reductions through process optimization and efficiency improvements. These findings suggest that future carbon reduction strategies should focus on green material substitution and technological innovation to enhance emission reduction potential and cost-effectiveness during the PC component production stage.

5.1.2. Analysis and Discussion of SD Model Results

Based on the system simulation results of twelve policy scenarios, this study examined single-policy pathways, coordinated policy mechanisms, and four types of mitigation mechanisms, including material, energy, technology, and management, to systematically reveal the dynamic trade-offs between emission reduction benefits and cost constraints, while further exploring their stage-specific characteristics and long-term trends.

Firstly, under single-policy pathways, green policies with an investment intensity of 10% to 20% are projected to achieve approximately 4.2% carbon reduction by 2030, with limited cost increases and high marginal benefits, indicating strong practical feasibility. However, when the investment exceeds 30%, the rate of emission reduction slows while costs continue to rise, suggesting that excessive policy intensity may lead to diminishing marginal returns, highlighting the need to balance policy strength and emission reduction effectiveness [52]. In comparison, carbon tax pathways offer higher theoretical reduction potential, reaching up to 5.61%, but result in substantial cost increases, up to 16.98%, which can adversely affect short-term corporate profits and return on investment. These findings demonstrate the tension between carbon pricing intensity and economic capacity, indicating that a single pathway is unlikely to achieve a long-term balance between emission reduction and cost control.

Secondly, under the coordinated mechanism of green policies and carbon taxation, simulation results indicate that when green policy investment is maintained at 10% to 20% combined with a carbon tax of CNY 100 per ton, a 7.94% reduction in emissions and a 12.91% increase in costs can be achieved by 2030, representing an optimal balance between emission reduction effectiveness and economic feasibility. The coordinated mechanism of green policies and carbon taxation effectively promotes the balance between carbon reduction and cost control [53]. Moreover, long-term trends suggest that this mechanism can effectively delay carbon rebound, demonstrating higher robustness and potential for broader application.

Thirdly, among the four types of emission reduction mechanisms of “materials–energy–technology–management”, the technology structure adjustment exhibits the highest emission reduction potential at 9.68%, highlighting the central role of intelligent manufacturing and equipment upgrades in improving carbon efficiency [54]. Energy structure optimization follows with an 8.97% reduction, indicating that energy restructuring is a key support for long-term emission mitigation [55]. Management structure adjustment achieves a 6.68% reduction; although its short-term performance is lower than that of the technology and energy pathways, it demonstrates sustained optimization potential over the long term [56], making it suitable as a supplementary measure to enhance overall system efficiency. Material optimization yields a relatively small reduction of 1.85%, but it plays a foundational role in improving raw material utilization, component lightweighting, and the development of green supply chains. Overall, the optimal emission reduction strategy should establish a multi-dimensional low-carbon transition system, with technology upgrades and energy optimization as the core, supported by management enhancement and material optimization, to achieve both stage-specific breakthroughs and long-term sustainability.

Finally, from the perspective of policy and corporate practice, the government should emphasize differentiated incentives and structural coordination in policy formulation. On the one hand, low-carbon behavior should be guided through green subsidies and tax adjustments to avoid excessive cost pressure from a single policy. On the other hand, the integration of green finance, carbon markets, and digital platforms should be promoted to provide multi-level support for the industrial chain. Meanwhile, enterprises should prioritize investment in technology and energy pathways that offer significant emission reduction and economic benefits. On this basis, management and material optimization measures can be gradually incorporated to form a phased and incremental low-carbon transition strategy, thus achieving a dynamic balance between carbon control and cost management under the “dual-carbon” strategy.

5.1.3. Practical Implications

The SD simulation results of this study hold not only theoretical significance but also provide actionable decision-making guidance for both government authorities and stakeholders in the PC industry chain. For policymakers, the findings reveal the trade-off dynamics between emission reduction and cost under different policy pathways, supporting the design of more precise and differentiated mitigation policies. For example, the synergy between green subsidies and carbon taxes has been shown to balance emission reduction effectiveness with economic feasibility, allowing the government to optimize fiscal expenditures, promote the integration of green finance and carbon markets, and enhance the robustness of policy implementation. For PC component enterprises, the model indicates that technology upgrades and energy optimization represent the most cost-effective emission reduction pathways. Enterprises can prioritize investments in smart manufacturing equipment, energy structure improvements, and digital management platforms to reduce carbon emissions while maintaining profitability. Additionally, the results enable construction companies and developers to assess carbon and cost risks at the project planning stage, fostering green supply chain development and enhancing market competitiveness. Overall, this study provides quantitative evidence and actionable pathways for both government policy formulation and corporate green transition.

5.2. Limitations of This Research and Future Outlook 25-262

Despite achieving certain outcomes in theoretical analysis and simulation modeling, this study has several limitations that warrant further investigation:

(1)

Uncertainty in data samples and parameter settings: Some variables involve internal enterprise operational data and regional policy implementation, which may suffer from data lag and sample limitations, potentially affecting the precision and applicability of the findings.

(2)

Insufficient representation of social-level feedback and external factors: The current analysis primarily focuses on production-side variables and does not fully incorporate social mechanisms such as government behavior, consumer green preferences, or enterprise green investment intentions. Additionally, fluctuations in energy prices, policy adjustments, and climate change may also introduce deviations to the model predictions.

(3)

Limitations in case representativeness and generalizability: This study is based solely on a prefabricated composite slab in Wuhan, chosen for its representativeness of production processes. However, the applicability of the findings to other component types or regional contexts should be interpreted with caution.

(4)

Conditional applicability of policy scenarios: Under the combined policy and carbon tax scenario (P7), although a relatively favorable emission reduction effect was achieved, the associated cost increase (12.91%) may be too high for some small and medium-sized enterprises to bear in the short term, indicating that the conclusions are condition-dependent.

Future research could proceed in several directions: (1) developing a parameter database grounded in field surveys and industry practices to improve the accuracy and applicability of model predictions; (2) employing multiple scenario designs and simulation methods to systematically address external uncertainties and to explore policy–enterprise–market interactions in depth, thereby enhancing the robustness of the findings; (3) incorporating advanced techniques such as artificial intelligence and big data analytics to enable real-time monitoring, forecasting, and optimization of carbon emissions in the construction sector, expanding the technological avenues for research; and (4) performing empirical and sensitivity analyses across a broader range of component types and regional settings, complemented by social surveys and employment impact assessments, to validate the robustness and generalizability of the model results.

These extensions would facilitate a deeper and more nuanced understanding of the driving mechanisms behind carbon emissions during the production phase of PC components, thereby improving both the scientific rigor and practical relevance of this study.

6. Conclusions

This study combines SEM and SD to systematically identify and quantify the key factors affecting carbon emissions and costs during the production of prefabricated concrete components, while also simulating the impacts of various policy measures and technological interventions on emission reductions and cost dynamics.

(1)

The SEM results indicate that among the primary indicators, the material production stage serves as the core driver of carbon emissions. At the secondary indicator level, policy standards, climatic and geographical conditions, types of raw materials, and the adoption of new technologies are identified as the key variables affecting carbon emissions.

(2)

The SD model further elucidates the dynamic regulatory mechanisms of different pathways on carbon emissions and costs. The main findings include the following: when the proportion of green policy investment is maintained within 10–20%, carbon emissions can be reduced by 4.2% while keeping costs at an optimal level; under the combined effect of moderate green policy (10–20%) and a carbon tax of CNY 100 per ton, the emission reduction rate increases to 7.52%, achieving a favorable balance between emission control and cost management; in multi-pathway strategies, technological adjustment and energy structure adjustment result in emission reductions of 9.68% and 8.97%, respectively, demonstrating significant carbon reduction potential; and while material structure optimization and management structure adjustment show relatively limited emission reduction effects, they provide continuous improvement potential and foundational support.

Overall, the findings provide theoretical and quantitative support for coordinated governance of carbon emission control and production cost optimization and offer empirical data and practical guidance for policymakers in designing green building strategies and for enterprises in implementing low-carbon transformation initiatives.