A railway line safety risk assessment method and system based on a two-dimensional cloud model

Abstract

The application discloses a railway line safety risk assessment method and system based on a two-dimensional cloud model, and the method comprises the following steps: constructing an index system for railway line safety risk assessment; determining the subjective weight of each index in the index system by using interval analytic hierarchy process, determining the objective weight of each index in the index system by using entropy weight method, and determining the combined weight by using improved game theory; for the railway line to be subjected to safety risk assessment, obtaining sequence data of each index including a plurality of sample observation values, and calculating two-dimensional cloud characteristic values of each index from two dimensions of risk occurrence probability and risk influence degree based on the sequence data of each index and the weight, and the two-dimensional cloud characteristic values are recorded as a cloud to be assessed; calculating the closeness between the cloud to be assessed and each level of risk standard cloud based on fuzzy distance, and determining the risk level with the largest closeness as the final assessment level of the railway line to be assessed. The application can simultaneously consider the risk occurrence probability and the consequence severity, and has high assessment accuracy and practicability.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, specifically relating to a method and system for assessing railway line safety risks based on a two-dimensional cloud model. Background Technology

[0002] Currently, although the railway industry has introduced various risk assessment methods, such as the Analytic Hierarchy Process (AHP), fuzzy comprehensive evaluation, and neural network prediction, many shortcomings still exist. For example, some methods rely too heavily on expert experience, and the weight allocation is highly subjective; some methods mainly focus on qualitative descriptions and lack intuitive presentation of risk quantification and visualization; at the same time, most traditional methods cannot simultaneously take into account the two key dimensions of "probability of risk occurrence" and "severity of consequences," resulting in fuzzy or uncertain risk level classifications that are difficult to meet the actual needs of the complex operating environment of railway lines.

[0003] Currently, there are numerous methods for railway safety risk assessment, with widely used ones including fuzzy comprehensive evaluation, analytic hierarchy process (AHP), and one-dimensional cloud modeling. Each of these methods has its advantages and has, to some extent, promoted quantitative analysis of safety risks in the railway industry. However, in practical applications, the following shortcomings still exist: (1) Most methods only deal with a single risk dimension and lack the ability to represent both "probability + severity"; (2) The weighting method relies heavily on subjective judgment and lacks an effective mechanism for integrating subjective and objective factors; (3) The risk assessment results are not intuitive enough, mostly consisting of static numerical values ​​or vague language, making them difficult to use directly for decision-making; (4) It has limited ability to differentiate risk levels, poor model interpretability, and insufficient scalability. Summary of the Invention

[0004] This invention provides a method and system for assessing railway line safety risks based on a two-dimensional cloud model. By constructing a multi-dimensional index system, adopting a weight calculation method that integrates subjective and objective factors, and introducing a two-dimensional cloud model for risk assessment, it can simultaneously consider the probability of risk occurrence and the severity of consequences, and has high assessment accuracy and practicality.

[0005] To achieve the above technical objectives, the present invention adopts the following technical solution: A method for assessing railway line safety risks based on a two-dimensional cloud model, comprising: Construct an indicator system for railway line safety risk assessment; The subjective weights of each indicator in the indicator system are determined by the interval hierarchical analysis method, the objective weights of each indicator in the indicator system are determined by the entropy weight method, and the combined weights are determined by the improved game theory method. For railway lines to be assessed for safety risks, score observation samples of each indicator in the two dimensions of probability of risk occurrence and degree of risk impact are obtained. Based on the score observation samples and weights of each indicator, the two-dimensional cloud feature values ​​of each indicator are calculated from the two dimensions of probability of risk occurrence and degree of risk impact, and are denoted as the cloud to be assessed. Based on fuzzy The proximity between the cloud to be evaluated for each indicator and the risk standard cloud at each level is calculated, and the risk level corresponding to the maximum proximity between the cloud to be evaluated for the comprehensive indicator of the indicator system and the risk standard cloud at each level is determined as the final evaluation level of the railway line to be evaluated.

[0006] Furthermore, the comprehensive index of the aforementioned indicator system is the railway line safety risk level. The primary indicators include environmental and geographical factors, meteorological factors, line factors, and social factors. Each primary indicator includes several independent secondary indicators, specifically: Environmental geographical factors include earthquakes, debris flows, landslides, collapses; and / or; Meteorological factors include wind speed, rainfall, snowfall, fog; and / or; Line factors include maintenance frequency, bridges and tunnels, signaling systems, curve radius, gradient; and / or; Social factors include the structure of the railway network, the degree of regional economic dependence, and the security system.

[0007] Furthermore, the subjective weights of each indicator in the indicator system are determined using the interval analytic hierarchy process, specifically including: Obtain the importance judgment matrix composed of all secondary indicators under each primary indicator and the importance judgment matrix composed of all primary indicators under the comprehensive indicator. ; where each element in the importance judgment matrix , indicating every two indicators , The relative importance of each factor to railway line safety risks, and all values ​​are range values. ; The importance judgment matrix will be represented using interval values, based on the upper limit of the interval values. and lower limit Divide into upper limit submatrices Lower bound submatrix , The number of indicators in the importance judgment matrix; Calculate the two submatrices separately The largest eigenvalue and the corresponding eigenvectors , , for The lower limit of the subjective weight of each indicator. for The upper limit of subjective weight for each indicator; Two feature vectors After normalizing each component, the components are combined to obtain the importance judgment matrix. The eigenvector corresponding to the largest eigenvalue ;in, and respectively, feature vectors , The normalized coefficient, , ; According to the consistency index and random consistency index Calculate the consistency ratio , Submatrix The largest eigenvalue The average value, i.e. If the consistency ratio is not met, the judgment matrix needs to be corrected. The pairwise comparison interval values ​​are performed until the consistency ratio meets the preset requirements; Calculate each indicator Subjective weight .

[0008] Furthermore, the combined weights are determined by improving game theory methods, specifically including: Each indicator Subjective weight and objective weight Linear combination is represented as: ; In the formula, As an indicator The combined weights, As an indicator The combination coefficient of subjective weight and objective weight; An improved game theory model is used to solve for the optimal weight combination coefficients. The optimization objective is to minimize the sum of squares of the Euclidean distances between the combined weights and the two initial weights. ; In the formula, This is the optimal solution for the weight combination coefficients, and , , ; Based on the first-order derivative formula, the combination coefficients in the above objective function are... Calculate the partial derivatives separately, and set them to zero to transform the problem into a system of linear equations for optimization. ; Find The optimal solution is obtained by combining the weights of each index based on improved game theory. : .

[0009] Furthermore, the method for obtaining the scoring observation samples of each secondary indicator is as follows: the raw data related to each secondary indicator is collected using a sensor monitoring system, and then the raw collected data is mapped to the level interval of risk occurrence probability and the level interval of risk impact degree according to a predefined level classification table. Several scoring observation samples of risk occurrence probability and risk impact degree are obtained by sampling within the mapped level interval. The scoring observation samples of the higher-level indicators are calculated by combining the scoring observation samples of the lower-level indicators with the weights of the corresponding indicators.

[0010] Furthermore, environmental and geographical factors are monitored by acquiring data through strong seismographs, GNSS displacement monitoring stations, deep displacement meters, and microseismic monitoring instruments. Real-time monitoring of ground motion, surface displacement, deep slippage, and rock mass fracture is achieved by utilizing inertial measurement, satellite positioning electromagnetic wave propagation delay, servo accelerometer inclinometer, and piezoelectric effect technologies, respectively. Meteorological factors were collected using ultrasonic anemometers, tipping bucket rain gauges, laser snow depth meters, and forward scattering visibility meters, based on ultrasonic time difference, gravity balance flipping, phase laser ranging, and Mie scattering principles, respectively, to collect data on wind speed and direction, rainfall intensity, snow depth, and atmospheric visibility. By utilizing the laser camera components and inertial measurement units of the track inspection vehicle, the bridge structural health monitoring system, the force-measuring wheelset and track circuit monitoring device, and employing laser triangulation, resistance strain effect, wheel spoke strain sensing and electromagnetic induction technology, the track geometry, structural stress and strain, wheel-rail dynamic response and signal equipment status are obtained. Social factors rely on the axle counters of the CTC dispatching system and the perimeter intrusion monitoring system, as well as RFID and vibration cables, to achieve real-time perception and integrated assessment of train operation density, passenger-freight ratio, and safety protection effectiveness.

[0011] Furthermore, based on the sequence data and weights of each indicator, two-dimensional cloud feature values ​​are calculated from two dimensions: the probability of risk occurrence and the degree of risk impact. Specifically: First, calculate the two-dimensional cloud feature value for each secondary indicator: , ; In the formula, and Indicators In the two dimensions of probability of risk occurrence and degree of risk impact, the first One score observation sample, The number of observation samples for scoring; , , The dimension of probability of risk occurrence is based on indicators The cloud feature values ​​correspond to expectation, entropy, and hyperentropy, respectively. , , Based on indicators in the dimension of risk impact. The cloud feature values ​​correspond to expectation, entropy, and hyperentropy, respectively; and Indicators In terms of both the probability of risk occurrence and the degree of risk impact The variance of the data; Then, by combining weights, the cloud feature values ​​of all secondary indicators under each primary indicator are weighted and summed to obtain the two-dimensional cloud feature value of each primary indicator in terms of the probability of risk occurrence and the degree of risk impact; and by combining weights, the cloud feature values ​​of all primary indicators under the comprehensive indicator are weighted and summed to obtain the two-dimensional cloud feature value of the comprehensive indicator in terms of the probability of risk occurrence and the degree of risk impact; wherein, the formula for calculating the weighted summation is: ; In the formula, The cloud feature value represents the probability of risk occurrence. Cloud feature values ​​representing the degree of risk impact. for The combined weights of each indicator.

[0012] Furthermore, based on fuzzy The formula for calculating the proximity between the cloud to be evaluated and the standard cloud for each risk level is as follows: ; , ; In the formula, and These represent the ambiguities between the cloud under evaluation and the standard clouds at each level of risk, specifically in the dimensions of probability of risk occurrence and degree of risk impact. distance, Standard cloud feature values ​​representing the probability of occurrence at each level of risk , , These correspond to expectation, entropy, and hyperentropy, respectively. Standard cloud feature values ​​representing the degree of risk impact at each level , , These correspond to their expected value, entropy, and hyperentropy, respectively. and Fuzzy logic in the two dimensions of probability of risk occurrence and degree of risk impact, respectively. Auxiliary angle parameters in distance calculation; This indicates the degree of similarity between the cloud to be evaluated and the risk standard clouds at each level.

[0013] A railway line safety risk assessment system based on a two-dimensional cloud model includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, causes the processor to implement any of the methods described above.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0015] (1) The rationality of indicator construction. Existing methods often design indicator systems based on experience or a few dimensions, which cannot fully cover the various risk sources involved in railway lines, such as the engineering structure status, environmental interference, and social factors, resulting in one-sided and distorted evaluation results. This invention proposes a multi-dimensional collaborative construction approach, which clearly emphasizes the systemic threats faced by railway lines as risk carriers, and emphasizes the structural integrity and content relevance of indicators, laying a data and structural foundation for scientific evaluation.

[0016] (2) The scientific nature of the integration of subjective and objective weighting. Traditional evaluation methods often rely on subjective scoring by experts or data-driven weight allocation alone, lacking a reconciliation mechanism. This results in evaluation results that vary greatly with expert opinions or cannot adapt to changes in data structure. This invention proposes a subjective and objective weighting integration mechanism to ensure that the weighting system is supported by professional judgment and can dynamically respond to actual changes, thereby fundamentally improving the scientific nature and persuasiveness of weight allocation.

[0017] (3) Clarity of risk level distinction. Existing risk level assessment methods often use static value judgment or membership degree classification, which can easily lead to blurred level boundaries and difficulty in interpreting assessment results, especially between adjacent levels where unclear boundaries and frequent misjudgments are common. This invention introduces a two-dimensional risk expression and similarity discrimination approach. Instead of classifying risk levels by a single value, it models risk through a two-dimensional model of "probability and severity" to achieve continuity, identifiability, and measurability between risk levels, significantly improving the clarity and operability of the classification effect.

[0018] (4) The intuitiveness of the graphical presentation of assessment results. Traditional assessment results are displayed in tabular or numerical form, which has limited information capacity and is difficult to support engineers' rapid reading and trend identification, which is not conducive to on-site decision-making or remote collaboration. This invention constructs a graphical and interactive assessment output mechanism, which supports the visual presentation of risk results in the form of images, expressing their uncertainty boundaries and core value trends, thereby improving the response efficiency of risk warning and the explanatory power of decision support. Attached Figure Description

[0019] Figure 1 This is a schematic diagram illustrating the steps of the railway line safety risk assessment method based on a two-dimensional cloud model as described in the embodiments of this application.

[0020] Figure 2 This is a Level 5 risk standard cloud map of an embodiment of this application, in which the sub- Figure 2 (a) Figure 2 (b) shows stereoscopic display and top view respectively.

[0021] Figure 3 This is a comparison chart of the subjective and objective weights and the combined weights described in this application.

[0022] Figure 4 This is the cloud map to be evaluated for implementing the primary indicator (environmental geographical factors) described in this application, wherein the sub- Figure 4 (a) Figure 4 (b) shows stereoscopic display and top view respectively.

[0023] Figure 5 This is the cloud map to be evaluated for implementing the primary indicator (meteorological factor) described in this application, wherein the sub- Figure 5 (a) Figure 5 (b) shows stereoscopic display and top view respectively.

[0024] Figure 6 This is the cloud map to be evaluated for implementing the primary indicator (route factor) described in this application, wherein the sub- Figure 6 (a) Figure 6 (b) shows stereoscopic display and top view respectively.

[0025] Figure 7 This is the cloud map to be evaluated for the implementation of the primary indicator (social factors) described in this application, wherein the sub- Figure 7 (a) Figure 7 (b) shows stereoscopic display and top view respectively.

[0026] Figure 8 This is the cloud map to be evaluated for implementing the comprehensive indicator (railway line safety risk) described in this application, wherein the sub- Figure 8 (a) Figure 8 (b) shows stereoscopic display and top view respectively. Detailed Implementation

[0027] The embodiments of the present invention will be described in detail below. These embodiments are based on the technical solutions of the present invention and provide detailed implementation methods and specific operation processes to further explain the technical solutions of the present invention.

[0028] Example 1

[0029] This embodiment provides a railway line safety risk assessment method based on a two-dimensional cloud model, referencing... Figure 1 As shown, it includes: Step 1: Construct an indicator system for railway line safety risk assessment.

[0030] Based on historical accident data mining and risk factor identification, and combined with the accident reports from the China State Railway Group over the past 10 years, primary indicators (such as environmental factors, equipment factors, meteorological factors, and social factors) were established. Each primary indicator includes several independent secondary indicators. For example, as shown in Table 1.

[0031] Table 1. Subjective and Objective Weights of Indicators at Each Level ; Note: *(1) Considering earthquake magnitude ≥ 5; *(2) Considering average wind force ≥ 6; *(3) Considering rain, snow, fog and sandstorms with visibility less than 200m. *(4) The design speeds of trunk railways are ≥ 350km / h, 350km / h, 300km / h, 250km / h, 200km / h, 160km / h, and < 160km / h. The higher the design speed, the larger the minimum curve radius.

[0032] Step 2: Use the interval hierarchical analysis method to determine the subjective weight of each indicator in the indicator system, use the entropy weight method to determine the objective weight of each indicator in the indicator system, and use the improved game theory method to determine the combined weight.

[0033] Step 2.1: Use Interval Analytic Hierarchy Process (IAHP) to determine the subjective weights of each indicator in the indicator system, specifically including: Step 2.1.1: Obtain the importance judgment matrix composed of all secondary indicators under each primary indicator and the importance judgment matrix composed of all primary indicators under the comprehensive indicator.

[0034] For each importance judgment matrix Each element Indicates every two indicators , In this embodiment, the importance of two indicators to railway line safety risks is represented by an interval, expressed as an interval value. Therefore, the importance judgment matrix is ​​as follows: ; In the formula, , The number of indicators in the importance judgment matrix.

[0035] Step 2.1.2 involves using an importance judgment matrix represented by interval values, based on the upper limit of the interval values. and lower limit Divide into upper limit submatrices Lower bound submatrix .

[0036] Step 2.1.3: Calculate the two submatrices respectively. The largest eigenvalue and the corresponding eigenvectors , , for The lower limit of the subjective weight of each indicator. for The upper limit of subjective weight for each indicator.

[0037] Step 2.1.4, combine the two feature vectors After normalizing each component, the components are combined to obtain the importance judgment matrix. The eigenvector corresponding to the largest eigenvalue ;in, and respectively, feature vectors , The normalized coefficient, , .

[0038] Step 2.1.5, Consistency check of importance judgment matrix.

[0039] The purpose of the consistency test is to verify the rationality of the importance judgment matrix. Based on the consistency test formula, the consistency ratio of the indicators is calculated. When <0.1, the consistency requirement is met. As a consistency indicator, This is a random consistency index. When the consistency ratio is not met, adjustments are needed until the consistency requirement is met. The consistency ratio is calculated using the following formula: When n=4, =0.9; when n=3, =0.58; when n=5, =1.12.

[0040] Step 2.1.6, calculate each indicator. Subjective weight .

[0041] Step 2.2: Use the entropy weight method to determine the objective weight of each indicator in the indicator system.

[0042] It has Each sample source, Several evaluation indicators form the original judgment matrix. ,in, For the first The sample source for the first The evaluation value of each indicator. The steps for determining the objective weight of the indicators using the entropy method are as follows.

[0043] Step 2.2.1, Normalize the judgment matrix Obtain the standard matrix : .

[0044] Step 2.2.2: Calculate the first value based on the normalized judgment value in the standard matrix. Information entropy value of each indicator : .

[0045] Step 2.2.3, calculate the first [item] based on the information entropy value. Objective weight of each indicator : .

[0046] Step 2.3 involves determining the combined weights using an improved game theory approach, specifically including: Step 2.3.1, set each indicator Subjective weight and objective weight Linear combination, expressed as: ; In the formula, As an indicator The combined weights, As an indicator The combination coefficient of subjective weight and objective weight; Step 2.3.2: The optimal weight combination coefficients are solved using an improved game theory model. The optimization objective is to minimize the sum of squares of the Euclidean distances between the combined weights and the two initial weights. ; In the formula, This is the optimal solution for the weight combination coefficients, and , , ; Step 2.3.3: According to the first-order derivative formula, calculate the combination coefficients in the above objective function. Calculate the partial derivatives separately, and set them to zero to transform the problem into a system of linear equations for optimization. ; Step 2.3.4, calculate The optimal solution is obtained by combining the weights of each index based on improved game theory. : .

[0047] Based on the above IAHP-entropy weighting method, the importance of the evaluation indicators is judged according to the two dimensions of the possibility and severity of railway safety risks, and the weight results of all evaluation indicators of railway safety risks are obtained.

[0048] Step 3: For the railway line to be assessed for safety risks, obtain the score observation samples of each indicator in the two dimensions of risk occurrence probability and risk impact degree. Based on the score observation samples and weights of each indicator, calculate the two-dimensional cloud feature value of each indicator in the two dimensions of risk occurrence probability and risk impact degree, and record it as the cloud to be assessed.

[0049] Step 3.1: For the railway line to be assessed for safety risks, obtain the score observation samples of each indicator in the two dimensions of probability of risk occurrence and degree of risk impact.

[0050] First, raw data related to each secondary indicator are collected using a sensor monitoring system. Environmental and geographical factors are monitored using strong-motion seismometers, GNSS displacement monitoring stations, deep displacement gauges, and microseismic monitoring instruments. Real-time monitoring of ground motion, surface displacement, deep slippage, and rock mass fracturing is achieved using inertial measurement, satellite positioning electromagnetic wave propagation delay, servo accelerometer inclinometer, and piezoelectric effect technologies, respectively. Meteorological factors are monitored using ultrasonic anemometers, tipping bucket rain gauges, laser snow depth meters, and forward scattering visibility meters. Based on ultrasonic time difference, gravity balance reversal, phase-based laser ranging, and Mie scattering principles, data on wind speed and direction, rainfall intensity, snow depth, and atmospheric visibility are collected. Track factors are monitored using laser camera components and inertial measurement units on track inspection vehicles, bridge structural health monitoring systems, force-measuring wheelsets, and track circuit monitoring devices. Laser triangulation, resistance strain effect, wheel spoke strain sensing, and electromagnetic induction technologies are used to obtain track geometry, structural stress and strain, wheel-rail dynamic response, and signal equipment status. Social factors rely on the axle counters of the CTC dispatching system and the perimeter intrusion monitoring system, as well as RFID and vibration cables, to achieve real-time perception and integrated assessment of train operation density, passenger-freight ratio, and safety protection effectiveness.

[0051] To ensure the authenticity and random uncertainty of the observation samples, based on the level definition standards of each risk indicator under specific geographical environment and operational conditions, the Monte Carlo simulation method is used to randomly sample within the corresponding score interval to construct an input sequence containing q sample observations. .

[0052] Specifically, based on the predefined grading table, as shown in Table 2, the corresponding grading intervals are mapped from the original collected data. Then, the Monte Carlo simulation method is used to randomly sample within the grading intervals to construct an observation sample containing q rating values.

[0053] Table 2 Risk Level Classification Table for Each Indicator ; In Table 2, PGA refers to peak ground acceleration (unit: g), TQI represents track quality index (unit: mm), L / 2000 is the deflection limit, L is the span (unit: mm), and the derailment factor is the ratio of lateral force to vertical force.

[0054] Step 3.2: Calculate the two-dimensional cloud feature values ​​of the railway line to be evaluated based on each indicator from two dimensions: the probability of risk occurrence and the degree of risk impact.

[0055] The characteristics of a visualized cloud map can be represented by the digital features of the cloud, such as the expected value (…). ),entropy( ), hyperentropy These are three numerical features that comprehensively describe the characteristics of the cloud model. The central point representing the distribution of cloud droplets in the number domain space is the point that best represents this qualitative concept. Uncertainty representing a qualitative concept can reflect the fuzziness (range) and randomness (dispersion) of a qualitative concept. The larger the value, the wider the range of the score. The uncertainty of entropy, or "entropy of entropy," can reflect the dispersion and thickness of cloud droplets.

[0056] First, calculate the two-dimensional cloud feature value for each indicator: , ; In the formula, and Indicators In the two dimensions of probability of risk occurrence and degree of risk impact, the first One score observation sample, The number of observation samples for scoring; , , The dimension of probability of risk occurrence is based on indicators The cloud feature values ​​correspond to expectation, entropy, and hyperentropy, respectively. , , Based on indicators in the dimension of risk impact. The cloud feature values ​​correspond to expectation, entropy, and hyperentropy, respectively; and Indicators In terms of both the probability of risk occurrence and the degree of risk impact The variance of the data.

[0057] Then, by combining weights, the cloud feature values ​​of all secondary indicators under each primary indicator are weighted and summed to obtain the two-dimensional cloud feature value of each primary indicator in terms of the probability of risk occurrence and the degree of risk impact; and by combining weights, the cloud feature values ​​of all primary indicators under the comprehensive indicator are weighted and summed to obtain the two-dimensional cloud feature value of the comprehensive indicator in terms of the probability of risk occurrence and the degree of risk impact; the formula for the weighted summation is: ; In the formula, The cloud feature value represents the probability of risk occurrence. Cloud feature values ​​representing the degree of risk impact. for The combined weights of each indicator.

[0058] Step 4, based on fuzzy The distance between the cloud to be evaluated for each indicator and the risk standard cloud at each level is calculated, and the risk level corresponding to the closest distance between the cloud to be evaluated for the comprehensive indicator of the indicator system and the risk standard cloud at each level is determined as the final evaluation level of the railway line to be evaluated.

[0059] The standard cloud is a cloud model generated from evaluation criteria. The probability of risk occurrence and the degree of risk impact are each categorized into five levels and quantified. , Representing the upper and lower bounds of the universe of discourse, respectively, the expected value, entropy, and hyperentropy of the two-dimensional standard cloud are calculated using the improved golden section method, serving as the standard cloud characteristics for each level interval: ; In the formula, The expected levels of risk standards for clouds are Level 1, Level 2, Level 3, Level 4, and Level 5, respectively. These are the entropy values ​​of cloud environments at risk levels one, two, three, four, and five, respectively. These are the hyperentropy values ​​for risk standards of Level 1, Level 2, Level 3, Level 4, and Level 5, respectively.

[0060] Let the universe of discourse (i.e., the range of values ​​for the risk assessment results) be... , Taking a value of 0.1, the digital characteristics of the risk standard cloud at each level were calculated, and the results are shown in Table 3. The risk cloud map is shown below. Figure 2 As shown. The golden ratio method is used to determine the risk level (Level I to V) boundaries; digital feature triples for each level of standard cloud are set. .

[0061] Table 3. Risk Level Corresponding Standard Cloud Digital Characteristics ; Specifically, based on fuzzy The distance between the cloud to be evaluated and the risk standard clouds at each level is calculated using the following formula: ; , ; In the formula, and These represent the ambiguities between the cloud under evaluation and the standard clouds at each level of risk, specifically in the dimensions of probability of risk occurrence and degree of risk impact. distance, Standard cloud feature values ​​representing the probability of occurrence at each level of risk , , These correspond to expectation, entropy, and hyperentropy, respectively. Standard cloud feature values ​​representing the degree of risk impact at each level , , These correspond to their expected value, entropy, and hyperentropy, respectively. and Fuzzy logic in the two dimensions of probability of risk occurrence and degree of risk impact, respectively. Auxiliary angle parameters in distance calculation This indicates the degree of similarity between the cloud to be evaluated and the standard cloud.

[0062] Ultimately, based on the principle of maximum proximity, the risk level with the highest proximity to the comprehensive indicator of the indicator system is determined as the final evaluation level of the railway line to be evaluated.

[0063] Example 2

[0064] This embodiment provides a railway line safety risk assessment system based on a two-dimensional cloud model, including a memory and a processor. The memory stores a computer program, and when the computer program is executed by the processor, the processor implements the method described in Embodiment 1.

[0065] Experimental data: Based on the railway line safety risk assessment index system, the subjective and objective weights of the indicators were calculated according to the steps of IAHP and entropy weight method, and the subjective and objective weights were combined using improved game theory. The comprehensive weights determined for each primary and secondary indicator are shown in Table 4.

[0066] Table 4. Subjective and Objective Weights of Indicators at Each Level ; As can be seen from the table above, the weights of the secondary indicators differ significantly under each primary indicator. The improved game theory combined weighting method described in this embodiment of the invention makes the calculated weight values ​​more reliable. A comparison of risk weights obtained by different methods is provided below. Figure 3 As shown.

[0067] By combining the numerical characteristics of the probability of risk occurrence and the degree of risk impact with the weight values ​​in Table 4, the digital characteristics of the primary indicators are calculated based on the numerical characteristics of the secondary indicators. Based on the numerical characteristic values ​​of the primary indicators and their weights, the digital characteristics of the comprehensive indicators can be further obtained. The numerical characteristic values ​​of the primary indicators and the comprehensive indicators are shown in Table 5.

[0068] Table 5 Numerical Characteristic Values ​​of Primary Indicators and Comprehensive Indicators ; Input the feature values ​​of the primary indicators for cloud evaluation from the table above into the positive cloud generator in MATLAB to generate a two-dimensional comprehensive cloud map, and compare it with a two-dimensional standard cloud map, such as... Figures 4-7 As shown.

[0069] As can be seen from the above two-dimensional cloud evaluation cloud map, the safety risks of the primary indicators U1, U2, U3, and U4 fall between levels IV and V, with U1, U2, and U3 closer to level IV and U4 closer to level V. The comprehensive indicator two-dimensional evaluation cloud is as follows: Figure 8 As shown.

[0070] The two-dimensional evaluation cloud map of the comprehensive indicators shows that the railway line safety risk is closer to Level IV. To more accurately obtain the line risk level, the closeness between each evaluation cloud and the standard cloud is calculated to accurately determine the railway line safety risk level. The calculation results of the closeness between the evaluation cloud of each indicator and the standard cloud of each risk level are shown in Table 6.

[0071] Table 6. Two-dimensional assessment of the proximity of railway lines to the risk cloud ; Based on the proximity calculation results, the risk levels of the primary indicators can be obtained, with environmental and geographical factors, meteorological factors, and route factors all classified as Level IV risk, while social factors are classified as Level V risk. The overall indicator is close to Level IV risk. In actual risk management, greater emphasis should be placed on social factors, as even minor risks can have significant impacts in busy route areas.

[0072] The risk assessment results of this invention not only provide a scientific basis for railway safety management, but can also be directly applied to practical decision-making.

[0073] Real-time monitoring and data collection: The evaluation indicators are monitored and collected in real time. The measured data of each indicator are used to generate the evaluation results of the probability of risk occurrence and the severity of consequences corresponding to the evaluation level.

[0074] Dynamic risk assessment: Input real-time monitoring data into a two-dimensional cloud evaluation model to conduct real-time or periodic risk assessments.

[0075] Risk Response and Adjustment: Develop train adjustment plans based on the current risk level, or identify risk trends by comparing historical data with the current status and provide timely warnings of potential risks.

[0076] For example, when the risk level is Level I (low risk), the mapping mechanism triggers the "normal operation" control strategy: the train runs at full speed at the line's designed maximum speed (e.g., 350 km / h), maintains the standard tracking interval (3 minutes), the backup line is closed to reduce energy consumption, and the emergency response is in a blue alert standby state without activating additional measures; when the risk level is Level II, the system limits the speed to 90%, increases the tracking interval by 10%, shortens the inspection cycle by 30%, the backup line is in hot standby, and the blue alert is activated; at Level III, the speed is limited to 80%, the interval is increased by 30%, special inspection and diversion are initiated by 20%, and the yellow alert is upgraded; at Level IV, the speed is limited to 60%, the interval is doubled, on-site duty is implemented, diversion is carried out by 50%, and the orange alert is upgraded; at Level V, emergency braking or shutdown is implemented, the section is closed, the entire network is diverted, and the red alert is automatically triggered. Each level converts the output of the two-dimensional cloud model into specific speed, interval, path, and emergency control parameters through the mapping mechanism, achieving precise matching between the risk level and the operating status.

[0077] The above embodiments are preferred embodiments of this application. Those skilled in the art can make various changes or improvements based on them. Without departing from the overall concept of this application, these changes or improvements should fall within the scope of protection claimed in this application.

Claims

1. A method for assessing railway line safety risks based on a two-dimensional cloud model, characterized in that, include: Construct an indicator system for railway line safety risk assessment; The subjective weights of each indicator in the indicator system are determined by the interval hierarchical analysis method, the objective weights of each indicator in the indicator system are determined by the entropy weight method, and the combined weights are determined by the improved game theory method. For railway lines to be assessed for safety risks, score observation samples of each indicator in the two dimensions of probability of risk occurrence and degree of risk impact are obtained. Based on the score observation samples and weights of each indicator, the two-dimensional cloud feature values ​​of each indicator are calculated from the two dimensions of probability of risk occurrence and degree of risk impact, and are denoted as the cloud to be assessed. Based on fuzzy The proximity between the cloud to be evaluated for each indicator and the risk standard cloud at each level is calculated, and the risk level corresponding to the maximum proximity between the cloud to be evaluated for the comprehensive indicator of the indicator system and the risk standard cloud at each level is determined as the final evaluation level of the railway line to be evaluated.

2. The railway line safety risk assessment method based on a two-dimensional cloud model according to claim 1, characterized in that, The comprehensive index of the aforementioned indicator system represents the safety risk level of railway lines. The primary indicators include environmental and geographical factors, meteorological factors, line-related factors, and social factors. Each primary indicator comprises several independent secondary indicators, specifically: Environmental geographical factors include earthquakes, debris flows, landslides, collapses; and / or, Meteorological factors include wind speed, rainfall, snowfall, fog; and / or, Line factors include maintenance frequency, bridges and tunnels, signaling systems, curve radius, gradient; and / or, Social factors include the structure of the railway network, the degree of regional economic dependence, and the security system.

3. The railway line safety risk assessment method based on a two-dimensional cloud model according to claim 1, characterized in that, The subjective weights of each indicator in the indicator system are determined using the interval analytic hierarchy process, specifically including: Obtain the importance judgment matrix composed of all secondary indicators under each primary indicator and the importance judgment matrix composed of all primary indicators under the comprehensive indicator. ; where each element in the importance judgment matrix , indicating every two indicators , The relative importance of each factor to railway line safety risks, and all values ​​are range values. ; The importance judgment matrix will be represented using interval values, based on the upper limit of the interval values. and lower limit Divide into upper limit submatrices Lower bound submatrix , The number of indicators in the importance judgment matrix; Calculate the two submatrices separately The largest eigenvalue and the corresponding eigenvectors , , for The lower limit of the subjective weight of each indicator. for The upper limit of subjective weight for each indicator; Two feature vectors After normalizing each component, the components are combined to obtain the importance judgment matrix. The eigenvector corresponding to the largest eigenvalue ;in, and respectively, feature vectors , The normalized coefficient, , ; According to the consistency index and random consistency index Calculate the consistency ratio , Submatrix The largest eigenvalue The average value, i.e. If the consistency ratio is not met, the judgment matrix needs to be corrected. The pairwise comparison interval values ​​are performed until the consistency ratio meets the preset requirements; Calculate each indicator Subjective weight .

4. The railway line safety risk assessment method based on a two-dimensional cloud model according to claim 1, characterized in that, Determining the combination weights through improved game theory methods specifically includes: Each indicator Subjective weight and objective weight Linear combination is represented as: ； In the formula, As an indicator The combined weights, As an indicator The combination coefficient of subjective weight and objective weight; An improved game theory model is used to solve for the optimal weight combination coefficients. The optimization objective is to minimize the sum of squares of the Euclidean distances between the combined weights and the two initial weights. ； In the formula, This is the optimal solution for the weight combination coefficients, and , , ; Based on the first-order derivative formula, the combination coefficients in the above objective function are... Calculate the partial derivatives separately, and set them to zero to transform the problem into a system of linear equations for optimization. ； Find The optimal solution is obtained by combining the weights of each index based on improved game theory. : 。 5. The railway line safety risk assessment method based on a two-dimensional cloud model according to claim 2, characterized in that, The method for obtaining the scoring observation samples of each secondary indicator is as follows: the original data related to each secondary indicator is collected using a sensor monitoring system, and then the original collected data is mapped to the level interval of risk occurrence probability and the level interval of risk impact degree according to the predefined level classification table. Several scoring observation samples of risk occurrence probability and risk impact degree are obtained by sampling within the mapped level interval. The scoring observation samples of the higher-level indicators are calculated by combining the scoring observation samples of the lower-level indicators with the weights of the corresponding indicators.

6. The railway line safety risk assessment method based on a two-dimensional cloud model according to claim 5, characterized in that, Environmental and geographical factors are monitored by acquiring data from strong-motion seismometers, GNSS displacement monitoring stations, deep displacement meters, and microseismic monitoring instruments. Real-time monitoring of ground motion, surface displacement, deep slippage, and rock mass fracture is achieved by utilizing inertial measurement, satellite positioning electromagnetic wave propagation delay, servo accelerometer inclination measurement, and piezoelectric effect technology, respectively. Meteorological factors were collected using ultrasonic anemometers, tipping bucket rain gauges, laser snow depth meters, and forward scattering visibility meters, based on ultrasonic time difference, gravity balance flipping, phase laser ranging, and Mie scattering principles, respectively, to collect data on wind speed and direction, rainfall intensity, snow depth, and atmospheric visibility. By utilizing the laser camera components and inertial measurement units of the track inspection vehicle, the bridge structural health monitoring system, the force-measuring wheelset and track circuit monitoring device, and employing laser triangulation, resistance strain effect, wheel spoke strain sensing and electromagnetic induction technology, the track geometry, structural stress and strain, wheel-rail dynamic response and signal equipment status are obtained. Social factors rely on the axle counters of the CTC dispatching system and the perimeter intrusion monitoring system, as well as RFID and vibration cables, to achieve real-time perception and integrated assessment of train operation density, passenger-freight ratio, and safety protection effectiveness.

7. The railway line safety risk assessment method based on a two-dimensional cloud model according to claim 1, characterized in that, Based on the sequence data and weights of each indicator, two-dimensional cloud feature values ​​are calculated from two dimensions: the probability of risk occurrence and the degree of risk impact. Specifically: First, calculate the two-dimensional cloud feature value for each secondary indicator: ， ； In the formula, and Indicators In the two dimensions of probability of risk occurrence and degree of risk impact, the first One score observation sample, The number of observation samples for scoring; , , The dimension of probability of risk occurrence is based on indicators The cloud feature values ​​correspond to expectation, entropy, and hyperentropy, respectively. , , Based on indicators in the dimension of risk impact. The cloud feature values ​​correspond to expectation, entropy, and hyperentropy, respectively; and Indicators In terms of both the probability of risk occurrence and the degree of risk impact The variance of the data; Then, by combining weights, the cloud feature values ​​of all secondary indicators under each primary indicator are weighted and summed to obtain the two-dimensional cloud feature value of each primary indicator in the two dimensions of risk occurrence probability and risk impact degree. Furthermore, by combining weights, the cloud feature values ​​of all primary indicators under the comprehensive indicator are weighted and summed to obtain the two-dimensional cloud feature values ​​of the comprehensive indicator in terms of both the probability of risk occurrence and the degree of risk impact; the formula for calculating the weighted sum is: ； In the formula, The cloud feature value represents the probability of risk occurrence. Cloud feature values ​​representing the degree of risk impact. for The combined weights of each indicator.

8. The railway line safety risk assessment method based on a two-dimensional cloud model according to claim 1, characterized in that, Based on fuzzy The formula for calculating the proximity between the cloud to be evaluated and the standard cloud for each risk level is as follows: ； ， ； In the formula, and These represent the ambiguities between the cloud under evaluation and the standard clouds at each level of risk, specifically in the dimensions of probability of risk occurrence and degree of risk impact. distance, Standard cloud feature values ​​representing the probability of occurrence at each level of risk , , These correspond to expectation, entropy, and hyperentropy, respectively. Standard cloud feature values ​​representing the degree of risk impact at each level , , These correspond to their expected value, entropy, and hyperentropy, respectively. and Fuzzy logic in the two dimensions of probability of risk occurrence and degree of risk impact, respectively. Auxiliary angle parameters in distance calculation; This indicates the degree of similarity between the cloud to be evaluated and the risk standard clouds at each level.

9. A railway line safety risk assessment system based on a two-dimensional cloud model, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the computer program is executed by the processor, the processor causes the processor to implement the method as described in any one of claims 1 to 8.

Images