A Study on the Dynamic Evolution and Prevention of Technological Innovation Risks in Major Railway Projects

Abstract

Technological innovation risk in major railway projects is high, and the risk factors in the innovation process are complex and variable. In order to successfully complete the goal of technological innovation in major railway projects. This paper identifies the risk factors of technological innovation in major railway projects, establishes a system dynamics model of technological innovation risk evolution, and carries out simulation analysis by taking the C railway project as an example, to understand the risk evolution process, clarify the key tasks of risk prevention, and finally draw the following research conclusions: (1) Technological innovation risk is categorized into innovation management risk, technology R&D risk, and technology application risk, of which innovation management risk has a greater impact on technological innovation risk. (2) During the construction process, the risk of technological innovation shows a general trend of increasing and then stabilizing, with the technology R&D stage being the centralized emergence stage of technological innovation risk. (3) Changes in the values of risk factors will have different degrees of impact on technological innovation risk, with changes in planning and implementation having the greatest impact on the technological innovation risk system, followed by the feasibility of results, and changes in the value of technological difficulties risk having a lesser impact on the system. The results of this study can provide lessons and references for the prevention of technological innovation risks in major railway projects.

1. Introduction

Railway projects are an important part of national infrastructure, the artery of the national economy and the backbone of the comprehensive transportation system [1,2,3]. The rapid development of the railway cannot be separated from the leadership and support of technological innovation. Under the innovation-driven development strategy, China has comprehensively pushed forward the research and industrialization of key railway technologies, and railway project technological innovation has undergone historic and global changes [4]. The technological innovation of major railway projects is a kind of technological innovation in the form of a project system, with the characteristics of high investment [5], multi-factor influence, multi-body participation, multi-stage construction [6], etc. The technological innovation of major railway project is often in a high-risk state, and it may be difficult for the innovation results to meet the expectations. Therefore, how to scientifically grasp the key factors influencing the risk of technological innovation, prevent and control the risk of technological innovation, and reduce the incidence of technological innovation accidents has become an important challenge facing the construction of major railway projects, and is also the focus of this paper.

Based on this, Cai Chaoxun et al. [7] identified 22 influencing factors in five dimensions for technological innovation in railway engineering, constructed a system dynamics model for simulation and sensitivity analysis, and concluded that technological innovation capability, technology adaptability to the environment, and research and development (R&D) capital investment are the key influencing factors. Wang Qing’e et al. [8,9] constructed a VAR model of technological innovation and management innovation in major construction projects, examined its synergistic relationship, and applied SNA to analyze the characteristics of the dual risk network, identify the core risk factors and risk links, and analyze the status and ability of the technological innovation subject in risk governance.

The above research mainly stays at the level of single-factor risk, limited to a single factor or simple risk superposition, and does not analyze the correlation and interaction strength between factors from a system perspective, ignoring the interaction between risk factors in the process of railway construction, and the technological innovation accidents of major railway projects are generally caused by the interaction between a variety of risk factors. Therefore, confronted with such a complex major project, this paper applies the system dynamics method, uses system thinking, starts from the uncertainty of various risk factors in the process of technological innovation of engineering projects, identifies the sources of risk brought by the internal and external environments, researches the dynamic evolution process of risk, and clarifies the focus of prevention and control. By studying the dynamic evolution of technological innovation risk for project managers to develop technological innovation, risk prevention and control strategies provide a guiding basis to improve the level of technological innovation risk management and efficiency, and to reduce the risk of technological innovation accidents, which are of great significance.

2. Literature Review

2.1. Construction Risk Management Research

Major railroad construction involves a number of projects in many fields, and there are many risk factors, wide influences, and great losses in the construction process. Efficient risk management plays a crucial role in ensuring the smooth progression of major railway construction projects and holds far-reaching significance for their overall success.

Generally, the risk management in major railway construction projects primarily centers on achieving key construction objectives, including quality, safety, schedule, and investment control. Wang X et al. [10] introduced a more efficient approach for accurately identifying railway track risks compared to traditional methods, leveraging risk prediction based on the track quality index of acceleration. Huang et al. [11] developed a tunnel collapse risk assessment model based on multi-morphological fuzzy Bayes, which is utilized to evaluate the probability of tunnel collapse and identify key risk factors. Wang G [12] employed the 5M theory to systematically recognize security risk factors, thereby more effectively determining the causes of security risks, potential hazards, and corresponding mitigation strategies. Yeh [13] employed the Work Breakdown Structure (WBS) decomposition method to address interface disputes, thereby enhancing progress in interface management. Timothy et al. [14] utilized Building Information Modeling (BIM) technology to identify interface conflicts and elucidated the critical factors for the successful implementation of BIM-based interface management. Guo et al. [15] developed an enhanced IM-DBSCAN algorithm, which integrates DBSCAN and rough set theory to assess railway investment risks.

In addition, recently, technological innovation has had a substantial and far-reaching impact on numerous industries. Consequently, a growing body of scholars have embarked on comprehensive studies regarding the risks associated with technological innovation in railway engineering. In the construction of large-scale infrastructures, such as railroad projects, the main bodies of technological innovation are always in a high-risk environment, and it is difficult to achieve the expected results of innovation [16]. Technological innovation plays a crucial role in the advancement of engineering construction and holds significant importance in enhancing engineering quality and safety [17], improving construction efficiency [18,19], reducing costs [20,21], promoting railway technology upgrades [22], and facilitating sustainable development [23,24]. Consequently, an increasing number of scholars have commenced investigations into the risks associated with technological innovation in railway engineering.

2.2. Railway Construction Technological Innovation Research

Currently, all industries across the globe are situated within the context of technological advancement. The role of technological innovation in railroad construction is growing, and it is crucial to enhance research on the risk management of technological innovation to foster the healthy and sustainable development of major railroad projects.

In recent years, academic research on the management of technological innovation has concentrated on application evaluation, influencing factors, development status, and solutions. Most studies on the risk management of technological innovation are conducted through factor identification and risk evaluation. Cai et al. [7] identified 22 influencing factors and analyzed them to pinpoint key factors such as technological innovation capability and technology adaptability to the environment, offering valuable insights for railway engineering and construction projects. Dou et al. [25] provided new perspectives on the risk management of technological innovation in railroad engineering and construction, analyzing the significance of technological innovation in building construction and the current state of innovation development, and proposing several strategies for implementing innovation in building construction technology. Mohammad et al. [26] investigated the obstacles hindering the development of photovoltaic (PV) technology in Iran by integrating the concepts of technological innovation systems and multilevel perspectives, using an explanatory structural model. Chuang P and BiK [27] assessed low-carbon technology innovation in China’s manufacturing industry within the global value chain, employing a combined hierarchical analysis and triangular fuzzy number assessment method. Gurrì Simona et al. [28] have driven the transition of the transportation mode on low-traffic routes towards rail through innovation. They employed the ELECTRE method to pinpoint the enhancements required for attaining an optimal solution that is convenient for both railway enterprises and end-users. In the context of the digital revolution, Garazi Carranza et al. [29] conducted a comprehensive analysis of existing technologies, delved into their functions within the innovation process, and put forward diverse solutions.

2.3. Evolution of Technological Innovation Risk Research

In recent years, risk events have become more complex, uncertain, and subject to rapid change. Consequently, traditional static methods of risk identification and assessment can no longer accurately and swiftly reflect the evolving risk situation [30,31,32]. Perrenoud et al. [33] have indicated that the nature of risk varies and dynamically changes at different stages of a project. Hence [34], there is a necessity to predict and categorize risks in a manner that evolves with the risks themselves and to propose more targeted and effective response strategies.

Büyüközkan G et al. [35] employed the risk assessment method based on the Choquet integral to explore the relationships between risk factors and engineering objectives. Feng et al. [36], grounded in complex network theory, investigated the risk interactions within key infrastructures of complex systems and large-scale engineering projects. Guided by system thinking, Brian H W [37] pointed out that the occurrence of an accident is like the collapse of dominoes. The accident leads to losses, and there exists a chain-like relationship among various factors. Wei Hu [38] utilized the entropy weight-TOPSIS model, the obstacle degree model, and the minimum variance method to systematically construct an evaluation index system for geopolitical risks, and analyzed the spatio-temporal evolution, obstacle factors, and risk types of geopolitical risks in countries along the Belt and Road.

Zeng et al. [39] utilized system dynamics to analyze the safety risk evolution law of railroad construction on existing lines. Perrenoud et al. [33] noted that risk varies and changes dynamically at different stages of a project. Fang and Marle [40], and Wang and Yuan [41] demonstrated that construction risks can be interconnected through causal loops, emphasizing the need to consider the dynamic interactions between project risks during risk management. Zhang et al. [42] combined fuzzy logic with system dynamics to study innovation in major engineering projects, while Suprun et al. [43] developed a model of the architectural innovation process based on system dynamics. Park [44] and Bajracharya [45] analyzed the dynamic process of architectural innovation using a system dynamics approach.

3. Materials and Methods

3.1. Analysis of Risk Factors for Technological Innovation in Major Railway Projects

Technological innovation risk refers to the major railway engineering construction projects to adapt to the ultra-long construction period, ultra-large-scale, ultra-complex construction environment [46], in the process of launching technological innovation by the uncertainty of the external environment, the difficulty and complexity of technological innovation project itself, the innovator’s own ability and strength of the limitations, which leads to the possibility of technological innovation activities not being able to reach the expected goal.

Due to the major railway project’s need to carry out a large number of technological innovations and scientific research work, the need for a number of industries and fields of the organization involved in technological innovation, coupled with the major railway project construction technology, and given the nature of construction equipment, the complexity of the process of technological innovation is bound to have a variety of risks. Major railway project construction technology innovation risk management, from the authors’ point of view, needs to focus on considering the role of management in the process of technological innovation and the existence of risk. In addition, based on the technological innovation process, technological innovation is mainly divided into two stages of R&D and application [47], without considering the market expansion of technology. After considering the case of technological innovation, on-site research data, related literature, the technological R&D, and application of the two links, supplemented by environmental and management factors, this paper will be divided into 3 risk subcategories: innovation management risk [48], technology R&D risk [49], and technology application risk; and 23 risk factors, with the risk breakdown shown in Figure 1.

3.2. Comparison of Risk Evolution Methods

3.2.1. Determination of Risk Evolution Method

Based on an in-depth literature review, prevalent risk evolution methods at present encompass the Interpretative Structural Modeling (ISM), Social Network Analysis (SNA), Markov Chain, Evolutionary Game Theory, Bayesian Network, and System Dynamics Simulation. Each of these methods is characterized by distinct advantages and limitations, and consequently, they possess specific applicable scopes. The detailed information is presented in

Table 1. Comparison of research methods.

Name	Advantages	Disadvantages	Application Scope
Interpretative Structural Modeling (ISM)	Facilitates clarifying risk factor relationships, shows hierarchical ones simply. Quantitatively expresses transmission relationships.	Cannot quantify risk values per factor, nor show time-based changes.	Constructing risk chains among risk factors.
Social Network Analysis (SNA)	Conducts statistical inferences on large samples. Qualitative-quantitative combo enables precise relationship analysis.	Cannot ensure finding all connected entities and is mainly static, with little dynamic analysis.	Exploring key risk factors and key risk relationships.
Markov Chain	Easy to calculate, shows risk factor occurrence probability in stages.	Lacking evolution dynamics.	Predicting the changes in risk factors in each period of time.
Evolutionary Game Theory	Realistically simulates human behavior, displays system evolution’s dynamic process.	Model simulation is highly affected by parameter settings.	Predicting or explaining the selection process of groups.
Bayesian Network	Features a flexible learning mechanism, with strong knowledge expression and reasoning.	Depends on assumptions like conditional independence.	Risk data identification and risk prediction, and can be used for the dynamic evolution of risks.
System Dynamics	Handles high-order, time-varying, non-linear complex problems well.	Needs a sound theoretical basis.	Simulating and emulating the internal structure of the system and its dynamic changes.

.

System dynamics models are particularly adept at handling difficult and complex problems that are high-order, prone to change over time, and exhibit non-linearity. They demonstrate good adaptability to long-term and cyclical problems, as well as issues that are difficult to quantify or lack a data foundation. Software such as Powersim, Vensim, Stella, and Ithink possess visualization functions and high operability, facilitating model simulation. However, its basic assumptions are the result of subjective abstraction and generalization of system problems, and their rationality when applied within a large-scale system remains debatable, necessitating a reasonable theoretical basis for support. In summary, system dynamics has superiority in clarifying the internal structure of a system, analyzing the causal relationships and action paths among its internal elements, and conducting simulations from the perspectives of system dynamic changes and internal structure to solve complex system problems.

Major railway engineering construction involves numerous technological innovation risk factors with complex relationships. Moreover, due to the extremely complex construction conditions of major railway projects, there are significant differences in risk factors and their corresponding risk values between major railway and ordinary railway projects. Thus, historical data from ordinary railway engineering construction risks cannot be used as a basis. Therefore, under the conditions of a complex system and lack of data, to ensure the feasibility and accuracy of the risk evolution model and fully reveal the risk evolution of technological innovation in major railway engineering construction, the system dynamics method can be utilized. System dynamics uses relevant model assumptions to simulate possible future scenarios of the system based on risk causal feedback. It has a relatively low data requirement for handling complex problems. Additionally, system dynamics can track and simulate the development and change trends of the system using computer simulation technology based on simulations of the system structure and feedback mechanisms, so it can be used to reveal the dynamic evolution of engineering construction risks when data accuracy is low.

3.2.2. System Dynamics Modeling Process

This paper explores the relationships among various risk factors in the technological innovation process of major railway engineering through system dynamics. Combining the risk factors encountered by innovation personnel in the actual innovation process, an SD model was constructed following steps such as model assumption, boundary determination, causal relationship analysis, characteristic equation determination, simulation testing, etc. First, through literature research and field survey data collation, the research purpose of this paper and the boundary of the constructed system are clarified, and relevant assumptions for this research are proposed. Based on theoretical research, the interactions between different risk factors are analyzed. Causal loop diagrams and stock-flow diagrams are drawn using Vensim PLE, initial values of factors are determined, model weights and characteristic equations are established, parameter variables are assigned values, and the model is simulated, tested, and optimized using Vensim PLE. The main steps are as follows:

(1)

Model Assumptions and System Boundary: To ensure the rigor and scientific nature of technological innovation risks in major railway engineering, this paper, combined with the characteristics of major railway engineering and railway technological innovation, makes assumptions about the model based on the basic theories and research methods of system dynamics, analyzes the system structure, clarifies specific problems, defines the reasonable boundary of the system, analyzes the main variables of the system, and clarifies the primary and secondary contradictions.

(2)

Drawing Causal Loop Diagrams: By analyzing the causes of problems and integrating system science, it is clear that there are interaction relationships among risk factors in different risk subsystems. The Vensim PLE 7.3.5 software is applied to establish a risk–relationship feedback model for technological innovation in major railway engineering, presenting system element indicators and their coupling relationships in the causal loop diagram.

(3)

Drawing Stock-Flow Diagrams: Based on the causal loop diagram, entity flows and information flows are distinguished, the causes of system changes are further refined, actual data to be obtained in variable equations and parameter settings are analyzed, constant coefficient parameters and element weights are determined, Dynamo equations are constructed, and the relationships among system elements, their feedback forms, and control laws are intuitively depicted.

(4)

Simulation: According to the progress of the project and combined with specific cases, the model is gradually simulated and modified based on the simulation results. Eventually, a model that conforms to logic and is close to the current situation is obtained, completing the verification and simulation of the model.

3.3. Construction of System Dynamics Model for Dynamic Evolution of Technological Innovation Risk in Major Railway Engineering

3.3.1. System Boundaries and Modeling Assumptions

This paper takes the risk of technological innovation in the construction of major railway projects as the research object, reveals the dynamic evolution law of risk over time in the construction process of technological innovation projects of major railroad projects, and analyzes the degree of influence of risk factors on technological innovation risk.

Before constructing the system dynamics model of major railway engineering technological innovation risk, it is necessary to delineate the system boundaries, exclude the unimportant elements that cannot effectively influence the system, and retain only the key factors, so as to facilitate the clearer expression of the main contents of the system. In this paper, the major railway engineering technological innovation project is defined as a closed system with a wide range and a complex operational process, which needs to determine the endogenous and exogenous variables in the system. The major railway engineering construction technological innovation process will face risk factors from different sources; with reference to the risk classification basis in the field of engineering construction technology innovation, this paper, from the authors’ point of view and technological innovation process, will be discussing a major railway engineering technological innovation risk system, which is divided into innovation management risk subsystems, technology R&D risk subsystems, and technology application risk subsystems. The innovation management risk subsystem includes risk factors such as central-local coordination, leadership coordination, plan implementation, schedule deviation, professional collaboration, information barriers, industry–university–research and application, personnel allocation, and conscious culture. The technology R&D risk subsystem includes risk factors such as lack of projects, empiricism, closed tendency, technical difficulties, construction method adaptability, equipment adaptability, and test verification. The technology application risk subsystem includes risk factors such as resource readiness, outcome feasibility, outcome interface association, construction equipment supporting, equipment mass production supporting, product application pollution, and intellectual property management. In summary, the boundaries of the major railway engineering technological innovation risk system dynamics model are innovation management risk subsystem, technology R&D risk subsystem, and technology application risk subsystem and its risk factors.

Meanwhile, in order to simplify the model and improve the relevance of the study, the following assumptions are made for the system:

Assumption 1. 

The technological innovation risk of railway engineering construction is a complex system, and only the role of innovation management risk, technology application risk, and technology application risk factors and their comprehensive impact on the technological innovation risk system are considered in the model.

Assumption 2. 

It is assumed that all the risk factors of technological innovation risk of major railway projects will neither be in a perfect nor imperfect state within the time limit, but will only change within a certain range, and will not suddenly undergo great changes.

Assumption 3. 

Only the risk factors from the decision-making and planning stage to the implementation and application stage of the project are considered.

3.3.2. Analyzing Model Causality

The causal loop diagram (CLD) [50] contains several variables linked by causal chains, which are indicated by arrows, and the causal relationship is transmitted with the direction of the causal arrows. The CLD can not only truly reflect the interaction between various factors in the technological innovation risk system of major railway engineering construction, but also truly reflect the internal operation of the system. Based on the system dynamics theory, this paper applies Vensim PLE 7.3.5 software to draw the system causality diagram of technological innovation risk, as shown in Figure 2.

3.3.3. Constructing System Dynamics Flow Diagrams

In order to more intuitively analyze and study the risk factors of technological innovation in major railway construction, on the basis of causality analysis, to further determine the nature of each factor, to more intuitively portray the logical relationship between system variables, to clarify the form of feedback and law, and to draw the corresponding stock flow diagram (Flow Diagram). Flow Diagram is the core content of system dynamics modeling, which usually includes five kinds of variables: state variables, rate variables, auxiliary variables, constants, and exogenous variables. Based on the causal loop diagram in Figure 2, three risk subsystems are designed to construct the stock flow diagram of the dynamic evolution of the risk of technological innovation in the construction of major railway projects, as shown in Figure 3.

3.3.4. Determine the Model Parameters and Equations

This model equation considers the different stages of each risk factor using the single pulse function as a constraint, and with reference to the actual progress of a technological innovation project of C railway engineering, it is set that the decision-making planning stage is 3 months, the technology R&D stage is 21 months, and the technology application stage is 12 months. According to whether the risk factors are affected by other risk factors in the system, they are divided into endogenous variables and exogenous variables; for the exogenous variables that are not affected by the internal system, the introduction of the risk evolution index describes the change in risk factors over time, specifically as shown in Equation (1), where Ri (t) is the value of the risk factor Ri at “t” time, r₀ is the value of the risk factor at t = t₀, a is the risk evolution index: a > 0 indicates that the risk is more and more serious, a = 0 indicates that the risk factor is constant and does not change over time, a < 0 indicates that the risk is gradually becoming smaller, (t₀, t₁) is the time of risk generation. For the endogenous variables to introduce the influence coefficient, specifically as shown in Equation (2), where w_ij is the degree of influence of risk factors R_j on risk factors R_i.

Ri (t) = r₀ × e^{a(t − t0)} × PULSE (t₀, t₁)

(1)

Ri(t) = (∑R_j × w_ij) × PULSE (t₀, t₁)

(2)

(1)

The determination of the initial value of a variable

After completing the construction of the technological innovation risk evolution model of major railway project construction, the initial value of the required parameters in the model variables should be determined first. This study uses the risk matrix method to determine the initial value of each risk factor, through expert scoring to determine the likelihood of the emergence of risk factors and the severity of the consequences of risk factors to determine the initial value of risk factors R = f (p, c).

The specific calculation formula is

$$\mathrm{R}=\mathsf{\rho }\times \mathrm{C}$$

(3)

The initial values of the risk factors in the system are shown in

Table 2. Preliminary table of risk factors for technological innovation in major railway projects.

Type of Risk	Risk Subcategory	Risk Factor	Starting Value
Technological innovation risk	Technological innovation risk	Central-local coordination	0.252
		Leadership coordination	0.09
		Plan implementation	0.240
		Schedule deviation	0.10
		Professional collaboration	0.15
		Information barrier	0.253
		Industry–university–research and application	0.09
		Personnel allocation	0.10
		Conscious culture	0.05
	Technology R&D risk	Lack of project	0.08
		Empiricism	0.15
		Closed tendency	0.182
		Technical problem	0.12
		Construction method adaptability	0.223
		Equipment adaptability	0.250
		Test verification	0.11
	Technology application risk	Resource readiness	0.11
		Outcome feasibility	0.20
		Outcome interface association	0.15
		Construction equipment supporting	0.14
		Equipment mass production support	0.16
		Product application pollution	0.09
		Intellectual property management	0.15

.

(2)

Calculation of risk factor weights

In this section, by adopting the comprehensive empowerment method, one can organically combine the subjective and objective empowerment methods, so that it can not only reflect the researcher’s subjective will for the research object, but also make up for the defects of poor objectivity brought about by insufficient data. To sum up, this study will adopt both the entropy value method and G1 method to empower the risk factors of technological innovation.

(1) Entropy

German scholar Rudolf Clausius put forward the concept of “entropy” in 1850, which is used to indicate the distribution of energy in space, the more uniform the distribution, the greater the entropy value. Claude Elwood Shannon introduced this concept into the information theory to judge the degree of disorder in the system. The connotative idea is to determine the weight of the corresponding indicator by the size of the information provided by each indicator. The main source of data for this study is the scoring of relevant experts and scholars in the field of railway construction. If there is a big difference between the data scored by experts for a certain indicator, the score cannot be deleted, and it should be considered that the information contained in this indicator is bigger, which may be more important for the study of the evolution of technological innovation risk in the construction of major railway projects. Therefore, the use of entropy value method of weight assignment can be pre-set indicators for experts to give subjective weight, and then use the entropy value method for weight correction, the specific steps are as follows. The questionnaire is shown in Appendix A.

① Create a matrix of expert scores for each risk indicator.

② Perform dimensionless processing of the data, using the standardization method in the non-linear method to calculate the mean and standard deviation:

$${\mathrm{Y}}_{\mathrm{i}\mathrm{j}}={\displaystyle \frac{{\mathrm{x}}_{\mathrm{i}\mathrm{j}}-\overline{{\mathrm{x}}_{\mathrm{j}}}}{\sqrt{\frac{1}{\mathrm{n}-1}{\sum }_{\mathrm{i}=1}^{\mathrm{n}} {\left({\mathrm{x}}_{\mathrm{i}\mathrm{j}}-\overline{{\mathrm{x}}_{\mathrm{j}}}\right)}^2}}}$$

(4)

③ Calculate the weight of each expert’s score under the jth risk or risk factor ${\mathrm{p}}_{\mathrm{i}\mathrm{j}}$:

$${\mathrm{p}}_{\mathrm{i}\mathrm{j}}={\displaystyle \frac{{\mathrm{Y}}_{\mathrm{i}\mathrm{j}}}{{\sum }_{\mathrm{i}=1}^{\mathrm{m}} {\mathrm{Y}}_{\mathrm{i}\mathrm{j}}}}(\mathrm{i}=1,2,\cdots ,\mathrm{m};\mathrm{j}=1,2,\cdots ,\mathrm{n})$$

(5)

④ Calculate the entropy of each risk and risk factor ${\mathrm{e}}_{\mathrm{j}}$:

$${\mathrm{e}}_{\mathrm{j}}=-{\displaystyle \frac{1}{\ln \mathrm{m}}}{\sum }_{\mathrm{i}=1}^{\mathrm{m}} {\mathrm{p}}_{\mathrm{i}\mathrm{j}}\ln {\mathrm{p}}_{\mathrm{i}\mathrm{j}}\left(0\le {\mathrm{e}}_{\mathrm{j}}\le 1\right)(\mathrm{i}=1,2,\cdots ,\mathrm{m};\mathrm{j}=1,2,\cdots \mathrm{n})$$

(6)

where “m” is the sample size, i.e., the number of experts involved in scoring.

⑤ Calculate the coefficient of variation for each risk and risk factor ${\mathrm{h}}_{\mathrm{i}}$:

$${\mathrm{h}}_{\mathrm{i}}=\left|1-{\mathrm{e}}_{\mathrm{i}}\right|$$

(7)

⑥ Calculate the weight of each risk and risk factor ${\mathrm{w}}_{\mathrm{i}}$:

$${\mathrm{w}}_{\mathrm{i}}={\displaystyle \frac{{\mathrm{h}}_{\mathrm{i}}}{{\sum }_{\mathrm{i}=1}^{\mathrm{n}} {\mathrm{h}}_{\mathrm{i}}}}$$

(8)

(2) G1 method

G1 method was first proposed by Prof. Guo Yajun of Northeastern University on the basis of hierarchical analysis method, which is a new method without constructing judgment matrix or consistency test. The principle of its assignment is to judge the importance of evaluation indexes by experts, and then assign the importance between them according to the scoring principle. The main steps of the method are as follows. The questionnaire is shown in Appendix B.

① Determination of the order relationship

When the degree of importance of the evaluation index a_i relative to a_j is large or equal, it is written as a_i ≥ a_j, when the evaluation index “${\mathrm{a}}_1, {\mathrm{a}}_2,$…$, {\mathrm{a}}_{\mathrm{n}}$” satisfy the relation “${\mathrm{a}}_1^{\prime }$≥${\mathrm{a}}_2^{\prime }$≥ … ≥${\mathrm{a}}_{\mathrm{n}}^{\prime }\mathrm{”}$, it represents that the evaluation indicators ${\mathrm{“}\mathrm{a}}_1, {\mathrm{a}}_2,$…$, {\mathrm{a}}_{\mathrm{n}}\mathrm{”}$ the order relationship is determined by “≥”. Here, ${\mathrm{a}}_{\mathrm{i}}^{\prime }$ denotes the ith evaluation indicator after {a_i} is ranked according to the ordinal relation “≥”, which is still denoted below. Denote the relation as

$${\mathrm{a}}_1\ge {\mathrm{a}}_2\ge \cdots \ge {\mathrm{a}}_{\mathrm{n}}$$

(9)

② Comparative judgment of the relative importance of the indicators is given

Set up a ratio ${\mathrm{W}}_{\mathrm{k}-1}$ to ${\mathrm{W}}_{\mathrm{k}}$ for the level of importance of the experts between the evaluation indicators ${\mathrm{a}}_{\mathrm{k}-1}$ and ${\mathrm{a}}_{\mathrm{k}}$:

$${\mathsf{\delta }}_{\mathrm{k}}={\displaystyle \frac{{\mathrm{W}}_{\mathrm{k}-1}}{{\mathrm{W}}_{\mathrm{k}}}}(\mathrm{k}=\mathrm{m},\mathrm{m}-1,\cdots ,3,2)$$

(10)

The assignment of ${\mathsf{\delta }}_{\mathrm{k}}$ can be referred to as shown in

Table 3. ${\mathsf{\delta }}_{\mathrm{k}}$ assignment reference table.

${\mathsf{\delta }}_{\mathbf{k}}$	Degree of Importance
1.0	Indicator ak-1 has the same level of importance as indicator ak.
1.2	Indicator ak-1 is marginally significant in relation to indicator ak.
1.4	Indicator ak-1 is clearly important in relation to indicator ak.
1.6	Indicator ak-1 is strongly associated with indicator ak.
1.8	Indicator ak-1 is extremely important in relation to indicator ak.
1.1, 1.3, 1.5, 1.7	The middle value of the two neighboring assignments, e.g., 1.1, is between the same level of importance and slightly more important.

.

③ Calculation of weighting factors

Based on the ${\mathsf{\delta }}_{\mathrm{k}}$ values given by the experts, it comes out ${\mathrm{w}}_{\mathrm{n}}$ to be

$${\mathrm{w}}_{\mathrm{n}}={\left(1+{\sum }_{\mathrm{k}=2}^{\mathrm{n}} {\prod }_{\mathrm{i}=\mathrm{k}}^{\mathrm{n}} {\mathsf{\delta }}_{\mathrm{k}}\right)}^{-1}$$

(11)

$${\mathrm{w}}_{\mathrm{k}-1}={\mathsf{\delta }}_{\mathrm{k}}{\mathrm{w}}_{\mathrm{k}}(\mathrm{k}=\mathrm{n},\mathrm{n}-1,\dots ,3,2)$$

(12)

④ Categorized discussion of cluster evaluations

When several experts were invited to evaluate the evaluation indicators, the findings that emerged were generally the following:

a. Cases with the same ordinal relationship

Assuming that L indicates that the experts are identical with respect to the ordinal relationship between indicators ${\mathrm{a}}_1,{\mathrm{a}}_2$, …, ${\mathrm{a}}_{\mathrm{n}}$, it can be recorded as follows:

$${\mathrm{a}}_1\ge {\mathrm{a}}_2\ge \cdots \ge {\mathrm{a}}_{\mathrm{n}}$$

(13)

Let the assignment of expert K to (i = n, n − 1, n − 2, …, 3, 2) be in the order

$${\mathrm{r}}_{\mathrm{k}2},{\mathrm{r}}_{\mathrm{k}3},\dots ,{\mathrm{r}}_{\mathrm{k}\left(\mathrm{m}-1\right)},{\mathrm{r}}_{\mathrm{k}\mathrm{m}}(\mathrm{k}=1,2,\dots ,\mathrm{L})$$

(14)

where ${\mathrm{r}}_{\mathrm{k}\mathrm{i}}$ satisfies ${\mathrm{r}}_{\mathrm{k}(\mathrm{i}-1)}\ge 1$ (i = n, n − 1, n − 2, …, 3, 2) yields:

$${\mathrm{r}}_{\mathrm{i}}^{\prime }={\displaystyle \frac{1}{\mathrm{L}}}{\sum }_{\mathrm{k}=1}^{\mathrm{j}} {\mathrm{r}}_{\mathrm{k}\mathrm{i}}(\mathrm{i}=2,3,\dots ,\mathrm{n})$$

(15)

$${\mathrm{w}}_{\mathrm{n}}={\left(1+{\sum }_{\mathrm{k}=2}^{\mathrm{n}} {\prod }_{\mathrm{i}=\mathrm{k}}^{\mathrm{n}} {\mathrm{r}}_{\mathrm{i}}^{\prime }\right)}^{-1}(\mathrm{i}=\mathrm{n},\mathrm{n}-1,\cdots ,3,2)$$

(16)

$${{\mathrm{w}}_{\mathrm{i}-1}}^{\prime }={\mathrm{r}}_{\mathrm{i}}^{\prime }{{\mathrm{w}}_{\mathrm{i}}}^{\prime }(\mathrm{i}=\mathrm{n},\mathrm{n}-1,\dots ,3,2)$$

(17)

b. The situation in the event of inconsistencies in ordinal relationships

Assuming that there are L0 (1 ≤ L0 ≤ L) experts to give the evaluation of the ordinal relationship between the indicator ${\mathrm{a}}_1,{\mathrm{a}}_2$, …, ${\mathrm{a}}_{\mathrm{n}}$ is the same, the corresponding weight of the indicator can be obtained as ${\mathrm{w}}_1,{\mathrm{w}}_2,\dots ,{\mathrm{w}}_{\mathrm{n}}$.

Let the L-L0 experts with different order relations give the order relations as

$${\mathrm{a}}_{\mathrm{k}1}\ge {\mathrm{a}}_{\mathrm{k}2}\ge \cdots \ge {\mathrm{a}}_{\mathrm{k}\mathrm{n}} (\mathrm{k}=1,2,\cdots ,\mathrm{L}-\mathrm{L}0)$$

(18)

where ${\mathrm{a}}_{\mathrm{k}\mathrm{i}}$ denotes the ith element in the set {${\mathrm{a}}_{\mathrm{i}}$} (i = 1, 2, $\dots$, n) of experts K arranged in the order relation “$\ge$”. Let the ratio of the importance of expert K to the ratio between ${\mathrm{a}}_{\mathrm{k}(\mathrm{i}-1)}$ and ${\mathrm{a}}_{\mathrm{k}\mathrm{i}}$ be ${\mathrm{r}}_{\mathrm{k}\mathrm{i}}$, when ${\mathrm{r}}_{\mathrm{k}\mathrm{i}}$ satisfies ${\mathrm{r}}_{\mathrm{k}(\mathrm{i}-1)}\ge 1/{\mathrm{r}}_{\mathrm{k}\mathrm{i}}$ then the weight coefficient of ${\mathrm{r}}_{\mathrm{k}\mathrm{i}}$:${\mathrm{w}}_{\mathrm{k}\mathrm{j}}$ (i = n, n − 1, n − 2, $\cdots$, 3, 2$;\mathrm{k}=1,2,\dots ,\mathrm{L}$ − L0) can be found.

For each k (1 ≤ k ≤ L-L0), the sets {${\mathrm{a}}_{\mathrm{k}\mathrm{i}}$} and {${\mathrm{a}}_{\mathrm{i}}$} correspond one to one. The weights (i = 1, 2, …, n) based on the judgments performed by each expert k (1 ≤ k ≤ L-L0) are used as the result of the combined calculation, denoted as ${{\mathrm{w}}_{\mathrm{i}}}^{″}$:

$${\mathrm{w}}_{\mathrm{i}}^{\prime }={\left({\prod }_{\mathrm{k}=1}^{\mathrm{L}-\mathrm{L}0} {\mathrm{w}}_{\mathrm{k}\mathrm{i}}^{\prime }\right)}^{{\displaystyle \frac{1}{\mathrm{L}}}-\mathrm{L}0}(\mathrm{j}=1,2,\cdots ,\mathrm{n})$$

(19)

$$\mathrm{or} {\mathrm{w}}_{\mathrm{i}}^{\prime }={\displaystyle \frac{1}{\mathrm{L}-\mathrm{L}0}}{\sum }_{\mathrm{k}=1}^{\mathrm{L}-\mathrm{L}0} {\mathrm{w}}_{\mathrm{k}\mathrm{i}}^{\prime }(\mathrm{j}=1,2,\cdots ,\mathrm{n})$$

(20)

The final synthesis yields a weight of ${\mathrm{w}}_{\mathrm{i}}={\mathrm{k}}_1{\mathrm{w}}_{\mathrm{i}}^{\prime }+{\mathrm{k}}_2{\mathrm{w}}_{\mathrm{i}}^{″}$ (${\mathrm{k}}_1,{\mathrm{k}}_2$ > 0), generally taken as ${\mathrm{k}}_1$=$\mathrm{L}0$/$\mathrm{L},{\mathrm{k}}_2$=$\left(\mathrm{L}-\mathrm{L}0\right)$/$\mathrm{L}$.

(3) Comprehensive Empowerment Method

After assigning weights to each risk and risk factor by the entropy value method and G1 method, respectively, the final weights are determined by comprehensive assignment method [51]. Let denotes the weight of each risk and risk factor, denotes the weight obtained by entropy value method, and denotes the weight obtained by G1 method.

$$\mathrm{w}=\mathsf{\alpha }{\mathrm{w}}_1+(1-\mathsf{\alpha }){\mathrm{w}}_2$$

(21)

where $\mathsf{\alpha }$ is the proportion of the objective preference coefficient weights to the combination weights. The objective function is established with the goal of minimizing the square sum of the deviations between the combined weights and the entropy weights and the deviation between the combined weights and the G1 weights, namely

$$\mathrm{m}\mathrm{i}\mathrm{n}\mathrm{z}={\sum }_{\mathrm{i}=1}^{\mathrm{n}} \left[{\left({\mathrm{w}}_{\mathrm{i}}-{\mathrm{w}}_{\mathrm{i}}^1\right)}^2+{\left({\mathrm{w}}_{\mathrm{i}}-{\mathrm{w}}_{\mathrm{i}}^2\right)}^2\right]$$

(22)

After substitution, the derivation about $\mathsf{\alpha }$ and the other first-order derivative is zero, solve the equation to obtain $\mathsf{\alpha }=0$.5, then the optimal composite weight results in half of the subjective and half of the objective weights when the composite weights minimize the sum of the squares of the two deviations from the subjective and objective weights, respectively.

(4) Weighting

Through the questionnaire data and entropy method steps for the objective weights of risk factors, the data were normalized, and the results of the calculations are shown in

Table 4. Weighting of risk factors for technological innovation in major railroad projects.

Type of Risk	Risk Subcategory	Risk Subcategory Weights	Risk Factor	Risk Factor Weights
Technological innovation risk	Technological innovation risk	0.42	Central-local coordination	0.12
			Leadership coordination	0.09
			Plan implementation	0.16
			Schedule deviation	0.10
			Professional collaboration	0.15
			Information barrier	0.13
			Industry–university–research and application	0.09
			Personnel allocation	0.10
			Conscious culture	0.05
	Technology R&D risk	0.30	Lack of project	0.08
			Empiricism	0.15
			Closed tendency	0.19
			Technical problem	0.12
			Construction method adaptability	0.18
			Equipment adaptability	0.17
			Test verification	0.11
	Technology application risk	0.28	Resource readiness	0.11
			Outcome feasibility	0.20
			Outcome interface association	0.15
			Construction equipment supporting	0.14
			Equipment mass production supporting	0.16
			Product application pollution	0.09
			Intellectual property management	0.15

.

(5) Determination of model parameter equations

Based on the above initial values of risk factors and risk factor weights, due to the limited space of the article, only the equations of the main variables of the model are listed, as shown in

Table 5. The equations for the main variables of the risk evolution model of technological innovation in major railroad projects.

Variable Name	Variable Type	Variable Expression
Technological innovation risk	Auxiliary variable	Innovation management risk × 0.42 + Technology R&D risk × 0.3 + Technology application risk × 0.28
Technology R&D risk	State variable	INTEG(Technology R&D risk variability, 0)
Technology R&D risk variability	Rate variable	Closed tendency × 0.19 + Construction method adaptability × 0.18 + Technical problem × 0.12 + Lack of project × 0.08 + Empiricism × 0.15 + Equipment adaptability × 0.17 + Test verification × 0.11
Conscious culture	Exogenous variable	0.146 × EXP(−0.2 × Time) × PULSE(0, 36)
Outcome interface association	Endogenous variable	(0.214 + Professional collaboration × 0.72) × PULSE(24, 6)

.

4. Results

4.1. Model Validity Test

In order to ensure the authenticity and accuracy of the model, the model is tested for validity and the test results are shown in Figure 4.

In this paper, Vensim-PLE software was used to construct the model, and “Check Units” was used to carry out the unit test and ensure it ran successfully, indicating that the model had passed the consistency test. The simulation results of the model show that the risk of technological innovation shows a rising and then stable trend, in the early stage, the risk appears and accumulates along with the work, and then stabilizes at 3.42 with the risk value of environmental and water protection after 30 months, and its evolution is in line with the actual project, so the system passes the realism test.

4.2. Simulation Scenario Analysis

In order to identify the priority stages and tasks for risk prevention and control, representative risk factors that can encompass all three aspects and cover all three stages need to be selected for analysis. Considering that exogenous variables may further affect endogenous variables, the greater the number of endogenous variables affected, the greater the impact on the system in general. In summary, considering the number of exogenous variables affecting endogenous variables, the endogenous variables only affected by exogenous variables, and the weight of the variables themselves, planning implementation, technical difficulties, and feasibility of results from the above three subcategories of risk were selected as representative risk factors to be analyzed, and keeping the equations of the other variables unchanged to analyze the degree of impact on the system of an increase of 30% in the risk value of a certain factor, and the results of the simulation are shown in Figure 5.

As illustrated in Figure 5, it becomes apparent that as the risk values of diverse risk factors ascend, the risk level of technological innovation correspondingly experiences varying degrees of augmentation. Among these factors, the fluctuation in the risk value of project planning implementation exerts the most significant impact on the technological innovation risk system. This is primarily because sound and effective planning implementation lay the groundwork for the entire innovation process, and any deviations therein can lead to a cascade of issues that heighten the overall risk [52]. Subsequently, the outcome feasibility occupies the second position in terms of its influence [53]. The assessment of whether a technological innovation is feasible in practical, economic, and other aspects serves as a critical checkpoint. If there are inaccuracies or uncertainties in this regard, it can introduce substantial risks, albeit to a lesser extent than those stemming from flawed plan implementation. Finally, technical problem, while undeniably posing challenges during the research and development stage, have a relatively minor impact on the system. This is attributable to the fact that, with appropriate resource allocation, such as increasing professional R&D personnel, ramping up scientific research funds, and fostering collaborative exchanges among different R&D groups, the negative consequences of technical difficulties can be effectively alleviated [54]. Comprehending the differential impacts of these factors is essential for devising targeted risk mitigation strategies and ensuring the smooth progression of technological innovation projects.

This phenomenon can be ascribed to the crucial roles that project planning and implementation play as prerequisites for project construction. In cases where the planning is insufficient and the implementation falls short, the overall realization of technological innovation within the project will be adversely affected. This, in turn, gives rise to a series of issues [55]. Firstly, it may lead to an inadequate allocation of personnel, causing a shortage of the required professionals in the project team. Secondly, deviations in project progress are likely to occur, disrupting the originally scheduled timelines. Thirdly, management loopholes in technological innovation may emerge, undermining the efficiency and effectiveness of the innovation process. Furthermore, the outcome feasibility constitutes the very core of technological innovation [56]. Any deviation in the feasibility assessment introduces latent risks that could potentially trigger the failure of the entire technological innovation endeavor. Given the profound interdependencies among these aspects, alterations in either the project planning implementation systems or the outcome feasibility evaluation framework will invariably exert a substantial impact on the overall system. Such impacts can permeate through multiple dimensions, ranging from resource allocation and schedule management to the ultimate success or failure of the technological innovation initiative. Technical problem predominantly center around the challenges encountered during the technology research and development process. Provided that there is an adequate number of scientific research personnel, these difficulties do not stem from other risk factors, nor do they augment the risk magnitudes of other associated factors. To a certain extent, such challenges can be mitigated by augmenting the pool of professional R&D staff, boosting the investment in scientific research funds [57], and intensifying the collaborative exchanges among diverse R&D groups. Consequently, alterations in technical difficulties exert a relatively minor impact on the overall system.

In conclusion, to ensure the successful realization of technological innovation projects, it is imperative for the project to bolster project planning implementation. This entails not only increasing the number of R&D professionals and the allocation of R&D funds but also placing a strong emphasis on the quality of preliminary work [58]. Strengthening the assessment of daily operations is equally crucial, as it enables the timely identification and resolution of potential hidden risks. Through these concerted efforts, the project can enhance its resilience and progress towards achieving its technological innovation goals.

5. Discussion and Recommendations

5.1. Discussion

According to the simulation results of technological innovation risk evolution in Figure 4, it can be analyzed that under the influence of given risk factors, the risk value of technological innovation risk grows slowly in the decision-making and planning stage, grows rapidly in the technology R&D stage, and grows faster in the early part of the application stage. Under the maturity of the technology, the risk value tends to be stable, and the maximum risk value is around 3.4 which will be reached within 30 months.

The characteristics of technological innovation risk evolution may be caused by the following reasons. Major railway project construction technology innovation is highly concerned by the decision-making level of the leadership, and the implementation of the application of the site puts forward high requirements, at this time, the complexity, the uncertainty, the difficulty and complexity of the technology innovation project itself, the innovator’s own ability and strength of the limitations of the technological innovation activities of the major railroad construction environment lead to the existence of a greater risk. Technological innovation risk in the decision-making planning stage will decide whether the project adopts the ‘four new technologies’ which include new technologies, new materials, new processes and new equipment. The decision-making information barriers between the various participating bodies, what is the proportion of their adoption, and whether the decision-making results are in line with the central and local policies, these questions lead to the generation of technological innovation risk; in the stage of planning and decision-making, the pre-existing risk factors on the layout of the technological innovation planning will continue to affect the R&D of ‘the four new technologies’, and the risk value of technological innovation risk continues to rise; in the research and R&D stage, the risk factors in terms of the feasibility and adaptability of the R&D results of ‘the four new technologies’ will continue to play a role in the risk of applying the technological innovation results, and when the technology is applied for a period of time, the R&D personnel will take risk control measures to smooth out the technological innovation risk to a reasonable level.

5.2. Recommendations

(1) Promoting dynamic synergies in technology and realizing multi-principal cooperation.

Under the guidance of the government, the innovation team of mega-railway engineering has formed a five-in-one technology system of government, industry, academia, research, and application, including construction units, design and survey units, construction units, research institutes and universities, etc. Based on the project information management platform, it improves the communication and synergy among all the parties involved in the construction and rationally synergizes the management of interfaces between different technologies according to the relevant systems and standards, so as to promote the synergistic cooperation among all the main parties in realizing the innovation of the project.

(2) Strengthen the coordination and linkage of all parties and break through information barriers.

This involves setting up an economic environment forecasting team to prevent changes in project implementation programs; setting up a progress monitoring team to ensure that the progress of the technological innovation project keeps up with the needs of construction; establishing a communication mechanism for the leading group of technological innovation to improve the degree of coordination and cooperation of the project team’s research and form the physical and technical interfaces between the various professional systems; creating a platform for data sharing; and implementing resource sharing. Other aspects include effectively applying the information system and the collaborative platform Building Information Modeling (BIM), big data, and other modern information technologies for dynamic collaborative management to achieve cross-organizational task identification, recording, and dynamic tracking, to solve the technical problems of the participating parties who need to deal with a large amount of complex information during the collaborative process, and to improve the information resource reserve of each professional system.

(3) Strengthen the talent team and enhance the awareness of technological innovation

This involves improving the mechanism of attracting talents, doing a good job in selecting and appointing excellent technological innovation talents, and establishing a technological innovation talent team that combines basic research, applied research, and technology development; establishing a sound management system for technological innovation, implement policies and measures such as rewards for the transformation of technological innovation achievements, autonomy of technological innovation, and open sharing of technological innovation resources; and improving the incentive mechanism for performance evaluation and income distribution. It will establish a scientific evaluation index system which focuses on the innovation ability and performance of major basic research, strengthening the orientation of categorized assessment and evaluation, and increasing the weight of assessment and evaluation of original and landmark achievements.

(4) Enhance the awareness of independent R&D, establish an information sharing mechanism.

It should strengthen the study and research on the theory of technological innovation in railroad construction, adhere to the position and viewpoints and methods therein conduct in-depth investigation and research, make field visits at the construction site, update the ideology and concepts in practice, and accumulate new experiences; establish an information sharing mechanism, strengthen the information sharing among the R&D staffs, open up the R&D process, enrich the theoretical data, and reduce the tendency of closure in the process of technology R&D.

(5) Set up the organization of technological innovation in water protection and environmental protection.

Improve the organization of technological innovation in environmental protection and water conservation, formulate environmental protection objectives and relevant rules and regulations, and take overall responsibility for technological innovation in environmental protection and water conservation in the jurisdiction area; be responsible for the formulation and implementation of environmental protection and water conservation measures and programs during the period of technological innovation; organize regular institutional discussions, research, and important environmental protection problem-solving.

(6) Strengthen intellectual property management.

The establishment of patent and soft-authorship decision-making management department and the training of professional management personnel for the management of patents, soft-authorship and other intellectual property rights must require the leadership to pay great attention and concern; to improve the professionalism, technicality and legality of the patent and soft-authorship management staff, and to improve the requirements for the recruitment of management personnel.

6. Conclusions and Limitations

6.1. Conclusions

(1) This paper identifies key risk factors associated with technological innovation from three perspectives: innovation management, technology R&D, and technology application. By utilizing the entropy value method and the G1 method, the study assigns weighted values to these risk factors, revealing that innovation management risks exert the greatest impact on overall technological innovation risk. Consequently, it is essential to enhance leadership and coordination in the pre-project phase, ensure the acquisition of real-time technological innovation information, and secure sufficient funding and resources throughout the project lifecycle. Additionally, fostering data sharing mechanisms and breaking down information silos between innovation stakeholders is critical.

(2) A simulation based on a case study of a major railroad project (Project C) demonstrates the dynamic evolution of technological innovation risks. The simulation results indicate that the risk of technological innovation during the construction phase of major railroad projects initially rises before stabilizing. Specifically, risk accumulates steadily during decision-making and planning, as well as technology R&D stages, eventually comprising over 80% of the total innovation risk. The technology R&D phase represents the peak of risk emergence. Therefore, it is crucial to establish preventive risk management practices early in the project and emphasize risk control throughout the technology R&D phase, thereby reducing the risk accumulation during the technology application stage.

(3) Planning and implementation of engineering projects as the basis for technological innovation work, planning and implementation of the risk value changes in the risk of technological innovation has the greatest impact; the feasibility of the results will affect the realization of the technological innovation goals, the risk value changes in the system has a relatively large impact; the technical difficulties did not cause the occurrence of other risk factors, and the impact of the risk value changes in the system has a relatively small impact. Therefore, the project should strengthen the leadership coordination, promote the dynamic synergy of technology, strengthen the talent team, and set up the organization of technological innovation in environmental and water conservation to reduce the pollution of the results of the application and improve the feasibility of the results. At the same time, more attention should be paid to the synergy between the preliminary R&D work and the on-site construction environment, and regular on-site investigation should be carried out to improve the accuracy of the work in the R&D stage.

6.2. Limitations

The risk system of technological innovation of major railway projects is a complex and huge system. This paper, on the basis of identifying and analyzing the factors, carries out the simulation research on the evolution of technological innovation risk of major railway projects with the help of the system dynamics method, and achieves certain research goals. In view of the fact that domestic and foreign research on the evolution of technological innovation risk of major railway projects still stays at the theoretical level, and its risk prevention and control presents a high degree of complexity, there are still issues that need to be discussed in depth:

(1) Major railway project technology innovation risk evolution research is still at the theoretical level; in practice, the risk involves a wide range, and this paper’s research is based on the authors’ perspective of the identification of risk factors and is not comprehensive enough. Future research needs to further explore the major railway project technology innovation risk factors to build a more scientific risk list.

(2) Due to the complexity of the interaction between risk factors, this paper may miss the key risk factors when combing the risks. In the sensitivity analysis of the various subsystems of the risk of technological innovation in major railway projects, even when considering the impact of single-factor changes on the risk, there is still a certain deviation from the actual situation; future studies need to further improve the SD model, to clarify the relationship between the risk factors and the degree of impact on the goal of technological innovation and the actual fit.

(3) As the focus of project management, risk prevention and control in the construction process of major railway projects requires active participation and active coordination of all parties, and in the future, it is necessary to closely contact the site to improve the operability of the coping strategies proposed in this paper.