Risk domain control method and system based on train formation operation risk cause analysis

Abstract

The application provides a risk domain control method and system based on train formation operation risk cause analysis, belongs to the technical field of train operation control, and realizes real-time perception of formation state information, realizes adjacent train state information interaction based on vehicle-vehicle communication technology, takes low brake anti-skid system efficiency and communication link establishment failure between formation trains as a risk source, takes train formation collision as an accident result, constructs a risk situation propagation model by using a dynamic Bayesian network, and performs sensitivity analysis and sorting on each risk node, uses nonlinear vibration theory and Rayleigh equation to describe vibration characteristics and a micro evolution process in the change of the risk situation, takes the sensitivity weight value of the node most influencing the collision accident occurrence probability as a damping coefficient of the Rayleigh equation, and explores the relationship between the control variable and the damping coefficient based on damping control, so that the risk domain converges to an initial safe state. The application can realize risk domain control for train formation operation from a system level.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of train operation control, in particular to a risk domain control method and system based on train formation operation risk cause analysis. BACKGROUND

[0002] Urban rail transit operation network plays an important role in relieving traffic congestion and accelerating the process of carbon neutralization, however, with the development of urban rail transit network operation, the phenomenon of passenger flow-power mismatch caused by the change of passenger flow over time gradually becomes obvious.

[0003] Inspired by the idea of vehicle formation operation, U.Block, a German scholar, first proposed the concept of virtual reconnection in 1999, hoping to achieve extremely short running intervals through formation, so that freight trains can not rely on physical couplings. It has been applied in heavy haul and urban rail transit in China. At the same time, by integrating positioning, communication and control technologies, the virtual reconnection-full automatic operation system improves the overall automation level of the rail transit control system while improving the efficiency of line operation. With the compression of running intervals, compared with traditional single train autonomous operation, trains running in formation mode are affected by potential risks such as the state of their own equipment, operating environment, personnel operation and organizational management form, which means that the risk of any train may have a chain reaction on the formation. In addition, due to the close coupling relationship between trains, mutual cooperation, frequent transmission of information and instructions, any potential risk will spread faster, wider and more difficult to control. If not intervened in time before the accident, it is easy to cause serious accidents, even large-scale interference with line operation, and seriously affect the operation efficiency. Therefore, it is urgent to improve the risk prevention and control and safety management capabilities of train formation operation in the field of rail transit.

[0004] Currently, the research on the mechanism of train formation operation has become mature, but the related operation safety research is relatively scattered, mainly focusing on risk analysis, or implicitly in the constraints or fault tolerance in interval control or optimization algorithms, and the risk analysis and prevention methods are disconnected; and most of them are for the operation of a single or a few non-formation trains, focusing on using the constantly developing safety analysis methods to call for a large number of human investigations, reviews and analyses of the whole process of the accident after the accident, and there is no perfect formation operation and timely safety management system, and there is a lack of in-depth research on formation operation and timely safety management methods and whole-link risk prevention methods. Therefore, for rail transit train formation operation, based on the update of precise positioning, train-to-train communication and other technologies, it is of great significance and application value to carry out research on risk cause analysis and active risk prevention methods, which can provide theoretical and technical support for the safe, smooth and efficient operation of train formation.

[0005] Currently, there is no in-depth research on risk domain control methods based on the risk causal analysis of train formation operation in existing technologies. Summary of the Invention

[0006] The purpose of this invention is to provide a risk domain control method and system based on the risk causal analysis of train formation operation, so as to solve at least one of the technical problems existing in the background art.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] In a first aspect, the present invention provides a risk domain control method based on the risk causal analysis of train formation operation, comprising:

[0009] In view of the multifactorial and cumulative nature of risks during train formation operation, a risk situation propagation model is constructed, taking the low efficiency of the braking anti-skid system under abnormal weather conditions and the high communication latency between the train formation workshops as risk sources and the collision of trains in the formation as the accident result.

[0010] By using dynamic Bayesian networks to simulate the occurrence of risk scenarios, sensitivity analysis is performed on each node in the risk situation propagation model and the nodes are ranked according to their sensitivity to identify the risk nodes that have the greatest impact on the probability of collision accidents.

[0011] Based on the applicability of the Rayleigh equation to describe the changes in the risk situation of the VC-FAO system, a risk state equation for the VC-FAO system is constructed. The sensitivity weight value of the node that has the greatest impact on the probability of collision accidents is used as the damping coefficient of the Rayleigh equation to describe the change process of the risk situation.

[0012] Based on the risk state equation and damping control, by introducing control variables, the magnitude of the limit cycle of the risk domain is derived, so that the risk domain constructed by the limit cycle is as small as possible or converges to the initial safe state.

[0013] Furthermore, a risk situation propagation model is constructed, including:

[0014] Based on the causes of risks and the consequences of accidents, a risk scenario is assumed: the automatic monitoring system issues formation instructions to two trains operating in moving block mode through the area controller and trackside communication equipment; based on the risk scenario, the evolution process of risk-accident is deduced; the deduced evolution process of risk-accident is transformed into a risk situation propagation model based on the risk-causing factors.

[0015] Furthermore, based on the risk scenario, the evolution of the risk-accident process is deduced, including: both trains confirm and start their car-to-car communication equipment, but due to the abnormal occupation of the rear car's car-to-car communication equipment and the abnormal weather conditions, the front car's car-to-car communication equipment experiences high latency, and the car-to-car communication link in the formation fails to be established; at the same time, the environmental monitoring system detects that the front car is slipping due to the influence of the rail surface adhesion coefficient, reports to ZC and ATS, and sets a temporary speed limit to implement braking; however, due to the failure to establish the car-to-car communication link, the rear car cannot obtain the change in the running status of the front car in time, still judges it to be running on the predetermined trajectory, and accelerates to catch up according to the formation instructions, which leads to a reduction in the running interval and the occurrence of a collision accident.

[0016] Furthermore, identify the risk nodes that have the greatest impact on the probability of a collision, including:

[0017] Based on logical behavior and sequence, the situation propagation model based on risk-causing factors is transformed into DBN, and the state monitoring data of the VC-FAO system is continuously input into the DBN model as evidence information;

[0018] Assuming that each node is independent of its parent node, the joint probability of all nodes is obtained by multiplying the conditional probabilities of each node.

[0019] Assuming that each time node must satisfy the Markov assumption, the current state is only related to the previous state, in order to reduce the complexity of DBN and improve analysis efficiency;

[0020] Assuming that each temporal node is stable within a single time slice, its conditional probability remains unchanged;

[0021] Based on the assumptions, a DBN risk situation propagation model is constructed to continuously calculate the probability of each risk node in the network, thereby realizing the dynamic Bayesian network visualization update of the probability of collision accidents in the train formation operation scenario.

[0022] Furthermore, since the parameters of each node in the DBN change over time, it is necessary to dynamically add evidence nodes at different times and consider the situational changes in adjacent time slices. In this way, the sensitivity weight of each risk node in the DBN can be calculated, and the risk nodes with the greatest impact on the probability of train formation collision accidents and their sensitivity weight values ​​can be obtained by sorting them.

[0023] Furthermore, based on the applicability of the Rayleigh equation to describe the changes in the risk situation of the VC-FAO system, a risk state equation for the VC-FAO system is constructed. The sensitivity weight of the node with the greatest impact on the probability of collision is used as the damping coefficient of the Rayleigh equation to describe the change process of the risk situation. This includes: using nonlinear vibration theory and the Rayleigh equation to describe the vibration characteristics and micro-evolution process of the VC-FAO system during the change process of the risk situation, and constructing the risk state equation based on this; considering that phase plane technology can intuitively observe the interrelationships between state variables and the changes of state variables over time; and transforming the risk state equation into a first-order differential equation containing only state variables. The system damping term in the equation is negative when it is small, but positive when it is large enough. This means that when using the Rayleigh equation to describe changes in the system's risk state, there is a moment when the damping term rapidly changes from positive to negative or vice versa, i.e., there is an abrupt change. Using the Lenard plotting method, a zero-slope isochoric line can be obtained in the phase plane, thus proving that the Rayleigh equation, while satisfying abrupt changes, is also applicable to describing changes in the risk state of the VC-FAO system, and that its phase trajectory will maintain periodic motion after a period of time. In addition, since each point in the phase trajectory contains the system's current risk state and the changing trend of the risk state, it can constitute the system's risk domain.

[0024] Furthermore, based on the risk state equation and damped control, by introducing control variables, the magnitude of the limit cycle amplitude of the risk domain is derived, so that the risk domain constructed by the limit cycle is as small as possible or converges to the initial safe state. This includes: introducing a risk control function into the risk state equation based on damped control, introducing small parameters, and using the multi-scale method to obtain the spatial approximate solution of the risk control function; synthesizing the original time phase using different time scales; obtaining the limit cycle amplitude of the risk domain when the Rayleigh equation describes the changes in the system risk situation; and by adjusting the relationship between the control variables and the damping coefficient of the Rayleigh equation, the risk domain of the queuing trains can be controlled from the system level.

[0025] Secondly, the present invention provides a risk domain control system based on risk causal analysis of train formation operation, comprising:

[0026] The module is designed to address the multifactorial and cumulative nature of risks during train platooning operations. It takes the low efficiency of the braking and anti-skid system under abnormal weather conditions and the high communication latency between the platooning workshops as risk sources, and the collision of trains in the platooning as the accident result, and constructs a risk situation propagation model.

[0027] The ranking module is used to simulate the occurrence of risk scenarios using dynamic Bayesian networks, perform sensitivity analysis on each node in the risk situation propagation model and rank them according to their sensitivity to determine the risk nodes that have the greatest impact on the probability of collision accidents.

[0028] The simulation module is used to simulate the changes in the risk situation of the VC-FAO system, construct the risk state equation of the VC-FAO system, and use the sensitivity weight value of the node that has the greatest impact on the probability of collision accidents as the damping coefficient of the Rayleigh equation to describe the process of risk situation changes.

[0029] The control module is used to control the risk domain of queuing vehicles from a system-level perspective by combining the risk state equation and damping control. By introducing control variables, the magnitude of the limit cycle of the risk domain is derived so that the risk domain constructed by the limit cycle is as small as possible or converges to the initial safe state.

[0030] Thirdly, the present invention provides a non-transitory computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the risk domain control method based on the risk causal analysis of train formation operation as described in the first aspect.

[0031] Fourthly, the present invention provides a computer device including a memory and a processor, wherein the processor and the memory communicate with each other, the memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the risk domain control method based on the risk causal analysis of train formation operation as described in the first aspect.

[0032] Fifthly, the present invention provides an electronic device, comprising: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions for implementing the risk domain control method based on risk causal analysis of train formation operation as described in the first aspect.

[0033] The beneficial effects of this invention are as follows: It uses multi-sensor fusion technology to perceive multi-dimensional train operation status information in real time, and vehicle-to-vehicle communication technology to achieve real-time interaction of status information between adjacent trains; addressing the multi-factor and cumulative nature of risks during train convoy operation, it constructs a risk situation propagation model, taking low efficiency of the braking and anti-skid system under abnormal weather conditions and failure to establish communication links between convoy trains as risk sources, and a collision in the convoy as the accident outcome; based on the risk-accident evolution process, it uses DBN to simulate the occurrence of risk scenarios and performs sensitivity analysis and ranking of each risk node; considering that the risk situation of train convoy operation is constantly changing, it uses nonlinear vibration theory and Rayleigh equations to describe the vibration characteristics and micro-evolution process of the VC-FAO system during the risk situation change process, and constructs a risk state equation accordingly, using the sensitivity weight value of the node with the greatest impact on the probability of collision accidents as the damping coefficient of the Rayleigh equation; based on the derivation of the limit loop amplitude of the risk domain, it explores the relationship between the control variables and the damping coefficient of the Rayleigh equation, minimizing the limit loop amplitude and converging the risk domain to the initial safe state. This invention enables risk domain control for train formation operation from a system-level perspective, reducing the reliance of risk studies on accident data or expert knowledge, and providing assurance for the intelligent, reliable, and safe operation of flexible train formation.

[0034] The advantages of additional aspects of the invention will be set forth more clearly in the following description or will be learned by practice of the invention. Attached Figure Description

[0035] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 This is a schematic diagram illustrating the specific process of the risk domain control method based on the risk causal analysis of train formation operation as described in an embodiment of the present invention.

[0037] Figure 2 This refers to various internal and external risks that may exist during the autonomous operation of a single vehicle and the collaborative formation operation of multiple vehicles as described in the embodiments of the present invention.

[0038] Figure 3 This is the risk situation propagation model based on risk causal factors described in the embodiments of the present invention.

[0039] Figure 4 This describes the risk-accident evolution process under train formation operation as described in the embodiments of the present invention. Detailed Implementation

[0040] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0041] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0042] It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with their meanings in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as here.

[0043] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, and / or groups thereof.

[0044] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0045] To facilitate understanding of the present invention, the present invention will be further explained and described below with reference to the accompanying drawings and specific embodiments. However, the specific embodiments do not constitute a limitation on the embodiments of the present invention.

[0046] Those skilled in the art should understand that the accompanying drawings are merely schematic diagrams of embodiments, and the components in the drawings are not necessarily essential for implementing the present invention.

[0047] Example 1

[0048] In this embodiment 1, a risk domain control system based on the risk causal analysis of train formation operation is first provided, including: a construction module, used to construct a risk situation propagation model to address the multi-factor and cumulative nature of risks occurring during train formation operation, taking the low efficiency of the braking anti-skid system under abnormal weather and the high communication latency between the train formation and the train collision as the accident result; a ranking module, used to simulate the occurrence process of risk scenarios using a dynamic Bayesian network, perform sensitivity analysis on each node in the risk situation propagation model and rank them according to their sensitivity to determine the risk node with the greatest impact on the probability of collision accidents; a deduction module, used to deduce the changes in the risk situation of the VC-FAO system, construct the risk state equation of the VC-FAO system, and use the sensitivity weight value of the node with the greatest impact on the probability of collision accidents as the damping coefficient of the Rayleigh equation to describe the change process of the risk situation; and a control module, used to combine the risk state equation and damping control from a system-level perspective to realize the control of the risk domain of the train formation, and derive the magnitude of the limit cycle of the risk domain by introducing control variables, so that the risk domain constructed by the limit cycle is as small as possible or converges to the initial safe state.

[0049] In this embodiment, the above-described system is used to implement a risk domain control method based on the risk causal analysis of train formation operation. This includes: considering the multi-factor and cumulative nature of risks occurring during train formation operation, a risk situation propagation model is constructed, taking the low efficiency of the braking and anti-skid system under abnormal weather conditions and the high communication latency between the train formation and the train collision as the accident outcome; a dynamic Bayesian network is used to simulate the occurrence of risk scenarios, and sensitivity analysis is performed on each node in the risk situation propagation model, ranking them by sensitivity to determine the risk node with the greatest impact on the probability of a collision accident; based on the applicability of the Rayleigh equation to describe the risk situation changes of the VC-FAO system, a risk state equation for the VC-FAO system is constructed, using the sensitivity weight value of the node with the greatest impact on the probability of a collision accident as the damping coefficient of the Rayleigh equation to describe the change process of the risk situation; based on the risk state equation and damping control, by introducing control variables, the magnitude of the risk domain limit cycle is derived, making the risk domain constructed by the limit cycle as small as possible or converging to the initial safe state.

[0050] In this embodiment, considering the multifactorial and cumulative nature of risks during train platooning operations, a risk situation propagation model is constructed, taking the low efficiency of the braking and anti-skid system under abnormal weather conditions and the high communication latency between the platooning units as risk sources, and the collision of trains in the platoon as the accident outcome. This model includes: assuming a risk scenario based on risk causes and accident outcomes: assuming the automatic monitoring system issues platooning instructions to two trains operating in moving block mode via the area controller and trackside communication equipment; deducing the risk-accident evolution process based on the risk scenario; and transforming the deduced risk-accident evolution process into a risk situation propagation model based on risk-causing factors. Specifically, this includes: based on risk causes and accident outcomes, assuming a risk scenario: to improve line operating efficiency, assume the automatic monitoring system issues formation commands to two trains operating in moving block mode via the area controller and trackside communication equipment; deduce the risk-accident evolution process: both trains confirm and activate their car-to-car communication equipment, but due to abnormal occupation of the rear train's car-to-car communication equipment and abnormal weather conditions, the front train's car-to-car communication equipment experiences high latency, resulting in failure to establish a car-to-car communication link between the trains in the formation; simultaneously, the environmental monitoring system detects that the front train is slipping due to the rail adhesion coefficient, reports to ZC and ATS, and sets a temporary speed limit to implement braking; however, due to the failure to establish the car-to-car communication link, the rear train cannot obtain the change in the front train's operating status in time, still judges it to be running on the predetermined trajectory, and accelerates to catch up according to the formation command, thus leading to a reduction in the running interval and a collision accident. The above risk-accident evolution process is transformed into a risk situation propagation model based on risk cause factors, such as... Figure 3 As shown.

[0051] In this embodiment, a dynamic Bayesian network is used to simulate the occurrence of risk scenarios, and sensitivity analysis is performed on each node in the DBN and sorted according to its sensitivity to determine the risk nodes that have the greatest impact on the probability of collision accidents. This includes: transforming the situation propagation model based on risk causal factors into a DBN based on logical behavior and sequence; continuously inputting the state monitoring data of the VC-FAO system as evidence information into the DBN model; assuming that each node is independent of its parent node, and obtaining the joint probability of all nodes by multiplying the conditional probabilities of each node; assuming that each temporal node must satisfy the Markov assumption that the current state is only related to the previous state, in order to reduce the complexity of the DBN and improve the analysis efficiency; assuming that each temporal node is stable within a single time slice and its conditional probability does not change; constructing a DBN risk situation propagation model based on these assumptions, continuously calculating the probability of each risk node in the network, and realizing the dynamic Bayesian network visualization update of the probability of collision accidents in the train formation operation scenario.

[0052] Specifically, based on logical behavior and sequence, the situation propagation model based on risk-causing factors is transformed into DBN, and the state monitoring data of the VC-FAO system is continuously input into the DBN model as evidence information.

[0053] Assuming that each node is independent of its parent node, the joint probability of all nodes can be obtained by multiplying the conditional probabilities of each node:

[0054] (1)

[0055] Assuming that each time node must satisfy the Markov assumption, the current state is only related to the previous state, as shown in equation (2), in order to reduce the complexity of DBN and improve analysis efficiency.

[0056] (2)

[0057] Assume that each temporal node is stable within a single time slice, and its conditional probability remains unchanged.

[0058] (3)

[0059] Based on the assumptions, a DBN risk situation propagation model is constructed to continuously calculate the probability of each risk node in the network, realizing a dynamic Bayesian network visualization update of the probability of collision accidents in train convoy operation scenarios. However, since the parameters of each node in the DBN change over time, it is necessary to dynamically add evidence nodes at different times, while considering the situation changes in adjacent time slices, defined as:

[0060] (4)

[0061] This allows us to calculate the sensitivity weight of each risk node in the DBN, and sort them to obtain the risk nodes that have the greatest impact on the probability of train formation collisions, along with their sensitivity weight values. .

[0062] In this embodiment, based on the applicability of the Rayleigh equation to describe the risk situation changes of the VC-FAO system, a risk state equation for the VC-FAO system is constructed. The sensitivity weight value of the node with the greatest impact on the probability of collision is used as the damping coefficient of the Rayleigh equation to describe the risk situation change process. This includes: using nonlinear vibration theory and the Rayleigh equation to describe the vibration characteristics and micro-evolution process of the VC-FAO system during the risk situation change process, and constructing the risk state equation based on this; considering that phase plane technology can intuitively observe the interrelationships between state variables and the changes of state variables over time; and transforming the risk state equation into a first-order differential equation containing only state variables. The system damping term in the equation is negative when it is small, but positive when it is large enough. This means that when using the Rayleigh equation to describe the change in the risk state of the system, there is a moment when the damping term rapidly changes from positive to negative or vice versa, i.e., there is an abrupt change. Using the Lenard plotting method, a zero-slope isochoric line can be obtained in the phase plane, thus proving that the Rayleigh equation, while satisfying the abrupt change, is also applicable to describing the risk state changes of the VC-FAO system, and that its phase trajectory will maintain periodic motion after a period of time. In addition, since each point in the phase trajectory contains the current risk state of the system and the changing trend of the risk state, it can constitute the system risk domain.

[0063] Specifically, nonlinear vibration theory and Rayleigh equations are used to describe the vibration characteristics and microscopic evolution of the VC-FAO system during the risk state change process, and a risk state equation is constructed based on this.

[0064] (5)

[0065] Considering that phase plane techniques can intuitively observe the relationships between state variables and how state variables change over time; let Transformed into a system containing only state variables First-order differential equations:

[0066] (6)

[0067] The second term in the Rayleigh equation is the system damping term. When it is small, this damping term is negative; however, when it is sufficiently large, it is positive. This means that when using the Rayleigh equation to describe changes in the system's risk state, there exists a moment when the damping term rapidly changes from positive to negative or vice versa, i.e., there is an abrupt change. Using the Lenard plotting method, let... The phase plane can then be obtained. The isochoric line with zero slope within the region.

[0068] (7)

[0069] It can then be proven that while satisfying abrupt changes, the Rayleigh equation is also applicable to describing the risk state changes of the VC-FAO system, and its phase trajectory will maintain periodic motion after a period of time; in addition, since each point in the phase trajectory contains the current risk state of the system and the changing trend of the risk state, it can constitute the system risk domain.

[0070] In this embodiment, since the size of the limit cycle depends on its amplitude, based on the risk state equation and damping control, the amplitude of the risk domain limit cycle is derived by introducing control variables, so that the risk domain constructed by the limit cycle is as small as possible or converges to the initial safe state. This includes: introducing a risk control function into the risk state equation based on damping control, introducing small parameters, and using a multi-scale method to obtain a spatial approximate solution of the risk control function; synthesizing the original time phase using different time scales; obtaining the amplitude of the limit cycle of the risk domain when the Rayleigh equation describes the changes in the system risk situation; and by adjusting the relationship between the control variables and the damping coefficients of the Rayleigh equation, the risk domain of the queuing trains can be controlled from a system-level perspective.

[0071] Specifically, a risk control function is introduced into equation (5) based on damping control:

[0072] (8)

[0073] Introducing small parameters and using the multi-scale method, we assume that the spatial approximate solution of equation (8) is:

[0074] (9)

[0075] The original temporal phase can be synthesized using different time scales; furthermore, by introducing operators... , The derivative of the time variable is expressed as:

[0076] (10)

[0077] Substituting equations (9) and (10) into equation (8), we get:

[0078] (11)

[0079] By simplifying and comparing the powers, we can obtain:

[0080] (12)

[0081] Since the zero-order term is used, it can be expanded into a complex number using Euler's formula, as shown in equation (13), and then substituted into equation (11). The right side of the equation yields:

[0082] (13)

[0083] (14)

[0084] Since the perpetual term exists, in order to eliminate resonance and ensure that the result does not diverge, let:

[0085] (15)

[0086] Combining Euler's formula, Expressed using equation (16), and substituting it into equation (15), we get:

[0087] (16)

[0088] (17)

[0089] Based on this, the limiting cycle amplitude of the risk domain when the Rayleigh equation describes changes in the system's risk situation can be derived. By adjusting the relationship between the control variables and the damping coefficients of the Rayleigh equation, the risk domain of queuing can be controlled from a system-level perspective.

[0090] Example 2

[0091] In this embodiment 2, a risk domain control method based on train formation operation risk causation analysis is provided, including the following steps:

[0092] Based on multi-sensor fusion technology, train multi-dimensional operation status information is perceived in real time, and vehicle-to-vehicle communication is established between trains in train formation to realize real-time interaction of status information between adjacent trains.

[0093] In response to the higher requirements for reliability and safety brought about by autonomous perception of train operation status, data command exchange and interaction, and formation collaborative decision-making and control, this paper conducts an in-depth analysis of various internal and external risks that may exist in the autonomous operation of a single train and the collaborative formation operation of multiple trains.

[0094] In view of the multifactorial and cumulative nature of risks during train formation operation, a risk situation propagation model is constructed, taking the low efficiency of the braking anti-skid system under abnormal weather conditions and the failure to establish communication links between the train formation workshops as risk sources and the collision of trains in the formation as the accident result.

[0095] Based on the risk-accident evolution process, Dynamic Bayesian Networks (DBN) are used to simulate the occurrence process of the above risk scenarios, and the sensitivity analysis of each risk node in the Dynamic Bayesian Network is performed and ranked to determine the risk node with the greatest impact on the probability of collision accidents.

[0096] Since the risk situation of train formation operation is constantly changing, nonlinear vibration theory and Rayleigh equations are used to describe the vibration characteristics and micro-evolution process of the Virtual Coupling-Fully Automatic Operation (VC-FAO) system during the risk situation change process. Based on this, a risk state equation is constructed, and the sensitivity weight value of the node with the greatest impact on the probability of collision is used as the damping coefficient of Rayleigh equation to describe the change process of risk situation.

[0097] Because it satisfies self-excited oscillation, its phase trajectory will exhibit a stable limit cycle after a period of time, constituting the system risk domain. If no intervention is made, the system risk domain will not change, and the risk and its risk resistance capacity will reach a balance, remaining in a high-risk state. Based on damping control, by introducing control variables into the risk state equation, the magnitude of the limit cycle amplitude of the risk domain is derived.

[0098] Based on the limit cycle amplitude of the risk domain, we explore the relationship between the control variables and the damping coefficient of the Rayleigh equation, so that the limit cycle amplitude is minimized and the risk domain converges to the initial safe state. This enables risk domain control for train formation operation from a system-level perspective.

[0099] In this embodiment, multi-sensor fusion technology is used to perceive multi-dimensional train operation status information in real time, and vehicle-to-vehicle communication is established between trains in convoy, specifically including:

[0100] Utilizing data from multiple sensors, including BeiDou, inertial navigation, speed sensors, transponders, lidar, and cameras, and supported by vehicle-to-vehicle / vehicle-to-ground networks, a train positioning and perception algorithm based on multi-sensor information fusion is used to calculate train operating status information in real time. Train operating environment information is acquired through a train operating environment collaborative perception method based on information interaction, achieving high-precision positioning and continuous perception of train operating status. This provides sensing and communication technology support for train tracking interval control and train group operation planning adjustments in platooning mode, ensuring safe and efficient collaborative operation control of trains in platooning mode.

[0101] Formation mode offers high operating speed and short relative distances, which enhances the flexibility and efficiency of formation and dismantling. However, it also faces various risks from both inside and outside the formation, such as... Figure 2As shown. A detailed analysis of the various risks that may exist in single-vehicle autonomous operation and multi-vehicle collaborative platooning operation is conducted, specifically:

[0102] The risk analysis process should consider the main elements constituting the system and its environment, including four categories: equipment, capabilities, procedures, and personnel. Specifically, it can be divided into internal risks caused by internal elements of the system or equipment and their spatiotemporal changes, such as traction and braking system failures, communication delays or interruptions, etc., and external risks caused by external environmental interference and their spatiotemporal changes, such as extreme weather effects, foreign object intrusion, etc. The occurrence of any risk may cause changes in system state, equipment failure, or malfunction, reducing system reliability and preventing operation according to normal or pre-defined operating modes. If proactive maintenance or timely correction of the system status is not carried out, it will directly affect system safety, and in severe cases, may even lead to systemic or functional safety accidents, or even large-scale disruption of line operation, seriously affecting operational efficiency.

[0103] Using the low efficiency of the braking and anti-skid system under abnormal weather conditions and the failure to establish communication links in the queuing workshop as risk sources, and a collision as the accident outcome, a risk situation propagation model is constructed. Specifically, this includes:

[0104] To improve line operating efficiency, assume that the Automatic Train Supervision (ATS) issues formation commands to two trains operating in moving block mode via the Zone Controller (ZC) and trackside communication equipment. Both trains confirm and activate their car-to-car communication equipment. However, due to abnormal occupation of the rear train's car-to-car communication equipment and abnormal weather conditions, the leading train's car-to-car communication equipment experiences high latency, resulting in the failure to establish a car-to-car communication link between the trains in the formation. Simultaneously, the environmental monitoring system detects that the leading train is slipping due to the rail adhesion coefficient, reports this to the ZC and ATS, and sets a temporary speed limit to apply brakes. However, due to the failure to establish the car-to-car communication link, the rear train cannot promptly obtain the change in the leading train's operating status, still judges it to be running on the predetermined trajectory, and accelerates to catch up according to the formation command, leading to a reduction in the running interval and a collision. This risk evolution process is transformed into a risk situation propagation model based on risk causal factors, as follows: Figure 3 As shown.

[0105] Based on the aforementioned risk-causing scenarios, and considering the risk-accident evolution process, a dynamic Bayesian network is used to simulate the occurrence of the risk scenario. Sensitivity analysis is performed on each risk node, and they are ranked to determine the risk nodes that have the greatest impact on the probability of a collision accident. Specifically, these include:

[0106] Due to differences in operating environments and the low reliability of system components, deviations from the intended operating state can occur. Therefore, the control system for implementing train platooning can be divided into a component layer consisting of hardware and software equipment, environment, and personnel; a subsystem layer consisting of individual trains; and a system layer consisting of the entire train platoon. Each layer has a corresponding function. The system's physical structure is based on components, and the realization of system functions also benefits from each component being able to perform its respective function, such as... Figure 4 As shown.

[0107] When the system is in operation, as risk sources emerge and their development progresses, component failures may occur. This could lead to individual vehicle malfunctions at the subsystem level of the system's functional structure, potentially causing platooning disruptions at the system level. Without predictive preventative measures, functional failures of varying degrees at different levels of the functional structure could lead to critical safety events, and in severe cases, accidents.

[0108] DBN is an extension of Bayesian networks incorporating the time dimension t. Based on the logical behavior and sequence of each risk node, the situation propagation model based on risk causative factors is transformed into DBN. State monitoring data from the VC-FAO system is continuously input into the DBN model as evidentiary information, and the following assumptions are made:

[0109] Assumption 1: Each node is independent of its parent node, therefore the joint probability of all nodes can be obtained by multiplying the conditional probabilities of each node:

[0110] (1)

[0111] In the formula, Indicate each risk node; It is a conditional probability distribution; As the parent node, when it is an empty set, it represents It is the root node, and Let it be its prior probability.

[0112] Assumption 2: Each time node must satisfy the Markov assumption, and the current state is only related to the previous state, as shown in equation (2), in order to reduce the complexity of DBN and improve analysis efficiency.

[0113] (2)

[0114] In the formula, express Time-based child node set; For each risk node at The joint probability at time t.

[0115] Assumption 3: Each time node is stable within a single time slice, and its conditional probability does not change.

[0116] (3)

[0117] In the formula, and They represent Time and The set of parent nodes at any given moment.

[0118] Based on this, a DBN risk situation propagation model can be constructed to continuously calculate the probability of each risk node in the network, realizing a dynamic Bayesian network visualization update of the probability of collision accidents in train convoy operation scenarios. However, since the parameters of each node in the DBN change over time, it is necessary to dynamically add evidence nodes at different times, while considering the situation changes in adjacent time slices, defined as:

[0119] (4)

[0120] In the formula, This represents the total number of time slices. For the time slice to which the target node belongs, For the first Nodes, For the first One evidence node.

[0121] Taking the probability of a train collision in a platoon as the target node, and setting the node probability when no collision occurs as the prior probability, and the node probability when a collision occurs as the posterior probability, then the sensitivity weight of each risk node in the DBN can be calculated using equation (4). The risk nodes with the greatest impact on the probability of train platoon collision accidents and their sensitivity weight values ​​can be obtained by sorting them. .

[0122] The vibration characteristics and micro-evolution process of the VC-FAO system during the risk situation change process are described using nonlinear vibration theory and Rayleigh equations. Based on this, a risk state equation is constructed, and the sensitivity weight values ​​of the risk nodes with the greatest impact on the probability of train formation collision accidents are obtained. As the damping coefficient in the Rayleigh equation, it describes the changing process of the risk situation. Specifically, it includes:

[0123] Differential equations are commonly used to describe the vibrational characteristics of complex systems, and are used to analyze the system's stability, resonance phenomena, and energy transfer properties. Because risks are constantly changing and accumulate with the spatiotemporal operational state of the train platoon, they can be mitigated through... describe The risk status of the time system, utilizing Describe the changes in the risk situation, and then analyze the vibration characteristics of the system during the process of changing risk situation;

[0124] The Rayleigh equation, as shown in equation (5), has the advantage of describing complex nonlinear systems while satisfying the order of the risk state equation of the VC-FAO system. Therefore, it is necessary to first prove the applicability of the Rayleigh equation for describing the risk state change process of the VC-FAO system;

[0125] (5)

[0126] Considering that phase plane techniques can intuitively observe the relationships between state variables and how state variables change over time; let The equation can be transformed into one containing only state variables. First-order differential equations:

[0127] (6)

[0128] The second term in the Rayleigh equation is the system damping term. When it is small, this damping term is negative; however, when it is sufficiently large, it is positive. This means that when using the Rayleigh equation to describe changes in the system's risk state, there exists a moment when the damping term rapidly changes from positive to negative or vice versa, i.e., there is an abrupt change. Using the Lenard plotting method, let... The phase plane can then be obtained. The zero-slope isoclipt lines within the equation (7) can then be used to prove that the Rayleigh equation, while satisfying abrupt changes, is also applicable to describing the risk situation changes of the VC-FAO system.

[0129] (7)

[0130] The isoclimax line with zero slope in the y-direction occurs in the first and third quadrants near the origin, with phase points diverging outwards; however, far from the origin, the isoclimax line with zero slope in the y-direction occurs in the second and fourth quadrants, with phase points contracting inwards. Therefore, these two opposing phase points must tend towards a stable limiting cycle, indicating that when using the Rayleigh equation to describe the risk state change process of the VC-FAO system, it satisfies self-excited oscillation, and its phase trajectory will maintain periodic motion after a period of time. In addition, since each point in the phase trajectory contains the current risk state of the system and the changing trend of the risk state, it can constitute the system's risk domain.

[0131] Based on the constructed risk state equation and damped control, the magnitude of the limit cycle amplitude of the risk domain is derived by introducing control variables; specifically including:

[0132] As the spatiotemporal operating state changes cumulatively, the system risk situation deviates from the initial safe state and generally shows an upward trend; and after a certain period of time, the system risk situation presents a stable limit cycle; if no intervention is made, the risk domain of the system will not change, the risk and its risk resistance capacity will reach a balance, and the system will continue to be in a high-risk state, which is contrary to the principle of prioritizing rail transit safety. It is urgent to control the risk domain so that the region formed by its limit cycle is as small as possible or converges to the initial safe state, and the size of the limit cycle depends on the amplitude of the limit cycle; therefore, based on damping control, a risk control function is introduced in equation (5):

[0133] (8)

[0134] In the formula, For control variables.

[0135] First, introduce small parameters. And using the multi-scale method, the spatial approximate solution of equation (8) is assumed to be:

[0136] (9)

[0137] In the formula, For time scale.

[0138] The original temporal phase can be synthesized using different time scales; furthermore, by introducing operators... , Equation (8) is transformed into:

[0139] (10)

[0140] By simplifying and comparing the powers, we can obtain:

[0141] (11)

[0142] Due to the zeroth-order term Expanding using Euler's formula, the result is:

[0143] (12)

[0144] In the formula, It does not include The amplitude of the term is a slowly varying process; substituting it into equation (11) The right side of the equation yields:

[0145] (13)

[0146] In the formula, This represents all conjugate terms.

[0147] Since the perpetual term exists, in order to eliminate resonance and ensure that the result does not diverge, let:

[0148] (14)

[0149] in addition, It can be represented as:

[0150] (15)

[0151] In the formula, and All are time scales Real functions;

[0152] Therefore, equation (14) can be transformed into:

[0153] (16)

[0154] Based on this, the limiting cycle amplitude of the risk domain when the Rayleigh equation describes changes in the system's risk situation can be derived:

[0155] (17)

[0156] Based on the risk domain limit cycle amplitude obtained from the above derivation, adjust the control variables. Damping coefficients in Rayleigh equations By understanding the relationship between them, the risk domain of train queuing can be controlled from a system-level perspective. The specific implementation process of damping control based on Rayleigh's equations is as follows: Figure 1 As shown.

[0157] Example 3

[0158] In this embodiment 3, a risk domain control method based on train formation operation risk causation analysis is provided, realizing an initial exploration of the timely safety concept from identification and deduction to control in the field of rail transit. The risk domain control method based on train formation operation risk causation analysis includes:

[0159] Based on multi-sensor fusion technology, high-speed trains can be monitored in real time for multi-dimensional operating status information, and vehicle-to-vehicle communication can be established between trains operating in virtual double-unit formations.

[0160] A thorough analysis was conducted on the various internal and external risks that may exist in autonomous single-vehicle operation and multi-vehicle collaborative platooning operation.

[0161] Taking the low efficiency of the braking and anti-skid system under abnormal weather conditions and the failure to establish communication links in the queuing workshop as risk sources, and the occurrence of a collision as the accident result, a risk situation propagation model based on risk causal factors is constructed.

[0162] Based on the logical behavior and sequence of each risk node in the risk-causing scenario, the situation propagation model based on risk-causing factors is transformed into a DBN (Database Deployment Network). State monitoring data from the VC-FAO system is continuously input into the DBN model as evidence. Assuming that each node is independent of its parent node, the joint probability of all nodes can be obtained by multiplying the conditional probabilities of each node, specifically expressed as:

[0163]

[0164] Furthermore, it is assumed that each time node must satisfy the Markov assumption, and the current state is only related to the previous state, in order to reduce the complexity of DBN and improve analysis efficiency. Specifically, this is expressed as follows:

[0165]

[0166] Furthermore, assuming that each temporal node is stable within a single time slice, its conditional probability remains unchanged, specifically expressed as follows:

[0167]

[0168] Based on this, a DBN risk situation propagation model is constructed to continuously calculate the probability of each risk node in the network, realizing the DBN visualization update of the collision accident probability under the train formation operation scenario. Combining the risk-accident evolution process, the DBN is used to simulate the occurrence process of risk scenarios, and sensitivity analysis and ranking of each risk node are performed to determine the risk node with the greatest impact on the collision accident probability, specifically represented as follows:

[0169]

[0170] Furthermore, nonlinear vibration theory and Rayleigh equations are used to describe the vibration characteristics and microscopic evolution of the VC-FAO system during the risk situation change process, and a risk state equation is constructed based on this. The sensitivity weight values ​​of the risk nodes that have the greatest impact on the probability of train formation collision accidents are then assigned. As the damping coefficient in the Rayleigh equation, it describes the changing process of the risk situation. Specifically, it is expressed as:

[0171]

[0172] As the spatiotemporal operating state changes cumulatively, the system risk situation deviates from the initial safe state and generally shows an upward trend; and after a certain period of time, the system risk situation exhibits a stable limit cycle; the purpose of using the Rayleigh equation to control the risk domain of the VC-FAO system is to minimize the region formed by its limit cycle or converge to the initial safe state (0,0), and the size of the limit cycle depends on the amplitude of the limit cycle; therefore, a risk control function is introduced based on damping control, specifically expressed as:

[0173]

[0174] Furthermore, the risk domain limit cycle amplitude is solved using the multi-scale method when the Rayleigh equation describes the changes in the system's risk situation. Specifically, it is expressed as:

[0175]

[0176] Based on this, adjust the control variables. Damping coefficients in Rayleigh equations The relationship between them makes the limit cycle amplitude of the risk domain as small as possible, that is, the risk domain converges to the initial safe state (0,0) as much as possible, thereby realizing the control of the risk domain of the queuing from the system level.

[0177] Example 4

[0178] This embodiment 4 provides a non-transitory computer-readable storage medium for storing computer instructions. When these computer instructions are executed by a processor, they implement the risk domain control method based on train formation operation risk causal analysis as described above. The method includes:

[0179] In view of the multifactorial and cumulative nature of risks during train formation operation, a risk situation propagation model is constructed, taking the low efficiency of the braking anti-skid system under abnormal weather conditions and the high communication latency between the train formation workshops as risk sources and the collision of trains in the formation as the accident result.

[0180] By using dynamic Bayesian networks to simulate the occurrence of risk scenarios, sensitivity analysis is performed on each node in the risk situation propagation model and the nodes are ranked according to their sensitivity to identify the risk nodes that have the greatest impact on the probability of collision accidents.

[0181] Based on the applicability of the Rayleigh equation to describe the changes in the risk situation of the VC-FAO system, a risk state equation for the VC-FAO system is constructed. The sensitivity weight value of the node that has the greatest impact on the probability of collision accidents is used as the damping coefficient of the Rayleigh equation to describe the change process of the risk situation.

[0182] Based on the risk state equation and damping control, by introducing control variables, the magnitude of the limit cycle of the risk domain is derived, so that the risk domain constructed by the limit cycle is as small as possible or converges to the initial safe state.

[0183] Example 5

[0184] This embodiment 5 provides a computer device, including a memory and a processor, wherein the processor and the memory communicate with each other, and the memory stores program instructions that can be executed by the processor. The processor calls the program instructions to execute the risk domain control method based on the risk causal analysis of train formation operation as described above, the method including:

[0185] In view of the multifactorial and cumulative nature of risks during train formation operation, a risk situation propagation model is constructed, taking the low efficiency of the braking anti-skid system under abnormal weather conditions and the high communication latency between the train formation workshops as risk sources and the collision of trains in the formation as the accident result.

[0186] By using dynamic Bayesian networks to simulate the occurrence of risk scenarios, sensitivity analysis is performed on each node in the risk situation propagation model and the nodes are ranked according to their sensitivity to identify the risk nodes that have the greatest impact on the probability of collision accidents.

[0187] Based on the applicability of the Rayleigh equation to describe the changes in the risk situation of the VC-FAO system, a risk state equation for the VC-FAO system is constructed. The sensitivity weight value of the node that has the greatest impact on the probability of collision accidents is used as the damping coefficient of the Rayleigh equation to describe the change process of the risk situation.

[0188] Based on the risk state equation and damping control, by introducing control variables, the magnitude of the limit cycle of the risk domain is derived, so that the risk domain constructed by the limit cycle is as small as possible or converges to the initial safe state.

[0189] Example 6

[0190] This embodiment 6 provides an electronic device, including: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes instructions to implement the risk domain control method based on the risk causal analysis of train formation operation as described above. The method includes:

[0191] In view of the multifactorial and cumulative nature of risks during train formation operation, a risk situation propagation model is constructed, taking the low efficiency of the braking anti-skid system under abnormal weather conditions and the high communication latency between the train formation workshops as risk sources and the collision of trains in the formation as the accident result.

[0192] By using dynamic Bayesian networks to simulate the occurrence of risk scenarios, sensitivity analysis is performed on each node in the risk situation propagation model and the nodes are ranked according to their sensitivity to identify the risk nodes that have the greatest impact on the probability of collision accidents.

[0193] Based on the applicability of the Rayleigh equation to describe the changes in the risk situation of the VC-FAO system, a risk state equation for the VC-FAO system is constructed. The sensitivity weight value of the node that has the greatest impact on the probability of collision accidents is used as the damping coefficient of the Rayleigh equation to describe the change process of the risk situation.

[0194] Based on the risk state equation and damping control, by introducing control variables, the magnitude of the limit cycle of the risk domain is derived, so that the risk domain constructed by the limit cycle is as small as possible or converges to the initial safe state.

[0195] In summary, the risk domain control method based on the risk causal analysis of train formation operation described in this embodiment of the invention is effective. Based on multi-sensor fusion technology, real-time perception of multi-dimensional train operation status information is achieved, and real-time interaction of status information between adjacent trains is realized based on vehicle-to-vehicle communication technology. Addressing the multi-factor and cumulative nature of risks during train platooning operation, a risk situation propagation model is constructed, taking low efficiency of the braking and anti-skid system under abnormal weather conditions and failure to establish communication links between platooning units as risk sources, and a collision in the platoon as the accident outcome. Based on the risk-accident evolution process, DBN is used to simulate the occurrence of risk scenarios, and sensitivity analysis and ranking of each risk node are performed to determine the risk node with the greatest impact on the probability of a collision accident. Considering that the risk situation of train platooning operation is constantly changing, nonlinear vibration theory and Rayleigh equations are used to describe the vibration characteristics and micro-evolution process of the VC-FAO system during the risk situation change process, and a risk state equation is constructed based on this. The sensitivity weight value of the node with the greatest impact on the probability of a collision accident is used as the damping coefficient of the Rayleigh equation. Based on the derivation of the limit loop amplitude of the risk domain, the relationship between the control variables and the damping coefficient of the Rayleigh equation is explored to minimize the limit loop amplitude and converge the risk domain to the initial safe state. This invention enables risk domain control for train formation operation from a system-level perspective, reducing the reliance of risk studies on accident data or expert knowledge, and providing assurance for the intelligent, reliable, and safe operation of flexible train formation.

[0196] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0197] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0198] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0199] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment, whereby a series of operational steps are performed to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes ​ The steps of the function specified in one or more boxes.

[0200] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that, based on the technical solutions disclosed in the present invention, various modifications or variations that can be made by those skilled in the art without creative effort should be included within the scope of protection of the present invention.

Claims

1. A risk domain control method based on the causal analysis of risks in train formation operation, characterized in that, include: In view of the multifactorial and cumulative nature of risks during train formation operation, a risk situation propagation model is constructed, taking the low efficiency of the braking anti-skid system under abnormal weather conditions and the high communication latency between the train formation workshops as risk sources and the collision of trains in the formation as the accident result. By using dynamic Bayesian networks to simulate the occurrence of risk scenarios, sensitivity analysis is performed on each node in the risk situation propagation model and the nodes are ranked according to their sensitivity to identify the risk nodes that have the greatest impact on the probability of collision accidents. Based on the applicability of the Rayleigh equation to describe the changes in the risk situation of the VC-FAO system, a risk state equation for the VC-FAO system is constructed. The sensitivity weight value of the node that has the greatest impact on the probability of collision accidents is used as the damping coefficient of the Rayleigh equation to describe the change process of the risk situation. Based on the risk state equation and damping control, by introducing control variables, the magnitude of the limit cycle of the risk domain is derived, so that the risk domain constructed by the limit cycle is as small as possible or converges to the initial safe state.

2. The risk domain control method based on the risk causal analysis of train formation operation according to claim 1, characterized in that, Construct a risk situation propagation model, including: Based on the causes of risks and the consequences of accidents, a risk scenario is assumed: the automatic monitoring system issues formation instructions to two trains operating in moving block mode through the area controller and trackside communication equipment; based on the risk scenario, the evolution process of risk-accident is deduced; the deduced evolution process of risk-accident is transformed into a risk situation propagation model based on the risk-causing factors.

3. The risk domain control method based on the causal analysis of train formation operation risks according to claim 2, characterized in that, Based on the risk scenario, the evolution of the risk-accident process is simulated, including: both trains confirm and start their car-to-car communication equipment, but due to the abnormal occupation of the car-to-car communication equipment of the following train and the abnormal weather conditions, the car-to-car communication equipment of the preceding train experiences high latency, and the car-to-car communication link in the formation fails to be established; at the same time, the environmental monitoring system detects that the preceding train is slipping due to the influence of the rail surface adhesion coefficient, reports to ZC and ATS and sets a temporary speed limit to implement braking; however, due to the failure to establish the car-to-car communication link, the following train cannot obtain the change in the running status of the preceding train in time, still judges that it is running on the predetermined trajectory, and accelerates to catch up according to the formation instructions, which leads to a reduction in the running interval and the occurrence of a collision accident.

4. The risk domain control method based on the risk causal analysis of train formation operation according to claim 1, characterized in that, Identify the risk nodes that have the greatest impact on the probability of a collision, including: Based on logical behavior and sequence, the situation propagation model based on risk-causing factors is transformed into DBN, and the state monitoring data of the VC-FAO system is continuously input into the DBN model as evidence information; Assuming that each node is independent of its parent node, the joint probability of all nodes is obtained by multiplying the conditional probabilities of each node. Assuming that each time node must satisfy the Markov assumption, the current state is only related to the previous state, in order to reduce the complexity of DBN and improve analysis efficiency; Assuming that each temporal node is stable within a single time slice, its conditional probability remains unchanged; Based on the assumptions, a DBN risk situation propagation model is constructed to continuously calculate the probability of each risk node in the network, thereby realizing the dynamic Bayesian network visualization update of the probability of collision accidents in the train formation operation scenario.

5. The risk domain control method based on the risk causal analysis of train formation operation according to claim 4, characterized in that, Since the parameters of each node in the DBN change over time, it is necessary to dynamically add evidence nodes at different times and consider the situational changes in adjacent time slices. In this way, the sensitivity weight of each risk node in the DBN can be calculated, and the risk nodes with the greatest impact on the probability of train formation collision accidents and their sensitivity weight values ​​can be obtained by sorting them.

6. The risk domain control method based on the risk causal analysis of train formation operation according to claim 1, characterized in that, Based on the applicability of the Rayleigh equation to describe the risk situation changes of the VC-FAO system, a risk state equation for the VC-FAO system is constructed. The sensitivity weights of the nodes with the greatest impact on the probability of collision are used as the damping coefficients of the Rayleigh equation to describe the risk situation changes. This includes: using nonlinear vibration theory and the Rayleigh equation to describe the vibration characteristics and microscopic evolution of the VC-FAO system during the risk situation change process, and constructing the risk state equation accordingly; considering the phase plane technique, which allows for intuitive observation of the interrelationships between state variables and their changes over time; and transforming the risk state equation into a first-order differential equation containing only state variables. In the process, the system damping term is negative when it is small, but positive when it is large enough. This means that when using the Rayleigh equation to describe the change in the risk state of the system, there is a moment when the damping term rapidly changes from positive to negative or vice versa, i.e., there is an abrupt change. Using the Lenard plotting method, a zero-slope isochoric line can be obtained in the phase plane, thus proving that the Rayleigh equation, while satisfying the abrupt change, is also applicable to describing the risk state changes of the VC-FAO system, and that its phase trajectory will maintain periodic motion after a period of time. In addition, since each point in the phase trajectory contains the current risk state of the system and the changing trend of the risk state, it can constitute the system risk domain.

7. The risk domain control method based on the risk causal analysis of train formation operation according to claim 1, characterized in that, Based on the risk state equation and damped control, by introducing control variables, the magnitude of the limit cycle amplitude of the risk domain is derived, so that the risk domain constructed by the limit cycle is as small as possible or converges to the initial safe state. This includes: introducing a risk control function into the risk state equation based on damped control, introducing small parameters, and using the multi-scale method to obtain the spatial approximate solution of the risk control function; synthesizing the original time phase using different time scales; obtaining the limit cycle amplitude of the risk domain when the Rayleigh equation describes the changes in the system risk situation; and by adjusting the relationship between the control variables and the damping coefficient of the Rayleigh equation, the risk domain of the queuing trains can be controlled from the system level.

8. A risk domain control system based on causal analysis of train formation operation risks, characterized in that, include: The module is designed to address the multifactorial and cumulative nature of risks during train platooning operations. It takes the low efficiency of the braking and anti-skid system under abnormal weather conditions and the high communication latency between the platooning workshops as risk sources, and the collision of trains in the platooning as the accident result, and constructs a risk situation propagation model. The ranking module is used to simulate the occurrence of risk scenarios using dynamic Bayesian networks, perform sensitivity analysis on each node in the risk situation propagation model and rank them according to their sensitivity to determine the risk nodes that have the greatest impact on the probability of collision accidents. The simulation module is used to simulate the changes in the risk situation of the VC-FAO system, construct the risk state equation of the VC-FAO system, and use the sensitivity weight value of the node that has the greatest impact on the probability of collision accidents as the damping coefficient of the Rayleigh equation to describe the change process of the risk situation. The control module is used to control the risk domain of queuing vehicles from a system-level perspective by combining risk state equations and damping control. By introducing control variables, the magnitude of the limit cycle of the risk domain is derived so that the risk domain constructed by the limit cycle is as small as possible or converges to the initial safe state.

9. A computer device, characterized in that, The system includes a memory and a processor, which communicate with each other. The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the risk domain control method based on the risk causal analysis of train formation operation as described in any one of claims 1-7.

10. An electronic device, characterized in that, include: The electronic device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions that implement the risk domain control method based on the risk causal analysis of train formation operation as described in any one of claims 1-7.

Images