Intelligent safety monitoring and evaluating method for long-distance transfer of precast beam

Abstract

The invention discloses an intelligent safety monitoring and evaluation method for long-distance transport of precast beams, and particularly relates to the technical field of transport safety of precast beam engineering, comprising the steps of S1, risk prevention and control equipment installation, S2, simulation modeling, S3, risk evaluation and S4, management and control early warning measure execution. According to the method, zero-cost and destructive mass trial and error are carried out in a virtual environment, the effectiveness of multi-dimensional decision logic is verified, accurate hierarchical management and control from slight prompt to forced intervention and remote linkage is realized through execution links according to risk levels, the interference of invalid alarm on a driver is effectively avoided, and the safety of the driver is improved. And through the vehicle-road-cloud coordination and full-process data recording function, the three core pain points of easy overturning due to high gravity center, long vehicle body, large blind area and easy fatigue in long-distance transportation in the precast beam transportation process are thoroughly solved, and the active safety, accident prevention capability and later liability definition scientificity in the transportation process are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of precast beam engineering transportation safety, and in particular relates to an intelligent safety monitoring and evaluation method for long-distance transportation of precast beams. BACKGROUND

[0002] The technical field of precast beam engineering transportation safety includes a logistics transportation risk control system for large concrete precast components from the precast yard to the construction site during the process of road and bridge construction. The core content of this field involves monitoring the dynamic stability of super-long and super-heavy beam transportation vehicles, spatial position sensing in complex road environments, and driving behavior management of long-time operators. The overall technical field systematically covers real-time posture capture of high-center-of-gravity cargo vehicles using physical sensing devices, identification of obstacles and road alignment on the transportation path through vehicle-mounted sensing terminals, and development of standardized operation procedures and emergency response strategies for special transportation operations. The main purpose is to ensure the physical stability and compliance of transportation vehicles in different working conditions through mechanical analysis and environmental monitoring methods.

[0003] The existing technology mainly relies on the manual experience and visual observation of the driver, lacks multi-source real-time sensing capability for the overturning posture caused by the high center of gravity of the beam transportation vehicle, abnormal tire pressure, and the huge blind area around the vehicle body, and is difficult to accurately identify and actively intervene in the driver's fatigue or distraction behavior in long-distance transportation. At the same time, due to the inability to conduct destructive real vehicle tests, there is a lack of quantitative risk assessment system based on digital twin simulation, and it is impossible to perform precise graded early warning or remote collaborative control according to the dynamically calculated comprehensive risk level. It can only passively respond to safety accidents and lacks active prevention and full-process digital recording means. SUMMARY

[0004] The main purpose of the present application is to provide an intelligent safety monitoring and evaluation method for long-distance transportation of precast beams, which can effectively solve the problems mentioned in the background art.

[0005] To achieve the above-mentioned purpose, the technical solution adopted by the present application is as follows: An intelligent safety monitoring and evaluation method for long-distance transportation of precast beams, comprising the following steps: S1: risk control equipment installation, installing a sensor group on the beam transportation vehicle for real-time collection of posture data representing the stability of the beam transportation vehicle, environmental data representing external environmental risks, and multi-source heterogeneous data representing the state of the driver, and transmitting through a 5G vehicle-mounted communication module; and installing a detachable reflective marker; S2: simulation modeling, inputting the multi-source heterogeneous data into a traffic safety simulation software, establishing a precast beam transportation benchmark scene model without integrated installation of risk control equipment and a precast beam transportation evaluation scene model with integrated installation of risk control equipment; S3: risk assessment, pre-set safety evaluation index of precast beam transfer, under the same initial conditions, respectively running the baseline scenario model and the evaluation scenario model, comparative analysis of simulation results, according to the pre-set safety evaluation index to evaluate and output the integrated risk level k of overturning risk, collision risk and driver state risk; S4: execute control warning measures, according to the risk level, execute corresponding risk control measures, and record the alarm information through the recording unit.

[0006] Preferably, the sensor group in S1 comprises: an attitude perception module comprising a plurality of inclinometers installed on the longitudinal center axis of the vehicle near the mass center position, a tire pressure detector; an environment perception module comprising four high-definition cameras, one installed on the front of the vehicle, one installed on the rear of the vehicle, and two respectively installed on the left and right sides of the vehicle body; a collision avoidance radar installed at the rear of the vehicle body; a driver state perception module comprising a fatigue warning instrument; The detachable reflective marker installation method in S1 includes but is not limited to fixing the detachable device with a rope on the top of the precast beam on both sides of the edge, pasting the reflective marker on a long stainless steel strip, and then fixing the stainless steel strip with the reflective marker on both sides of the detachable device. The reflective marker adopts a yellow and white striped high-contrast pattern.

[0007] Preferably, the simulation modeling in S2 specifically comprises: establishing a baseline scenario model, in a pre-set traffic safety simulation software, according to the target transportation route, road structure characteristics, beam body and beam truck structure size parameters, traffic flow, weather conditions, transportation time, a baseline transportation scenario model is constructed without integrating the risk prevention and control device logic; establishing an evaluation scenario model, in the simulation software, copying the baseline scenario model, and embedding the decision logic of the risk prevention and control device to form an evaluation transportation scenario model.

[0008] Preferably, the risk assessment in S3 specifically comprises: running comparative simulation, under the same initial conditions of running speed and acceleration, respectively running the baseline scenario model and the evaluation scenario model, comparing the running posture of the beam truck in straight road section, slope, curve and road intersection under the two scenarios, and collecting the pre-set safety evaluation index data, quantitatively calculating and outputting a comprehensive risk level k.

[0009] Preferably, the decision logic of the risk prevention and control device embedded in the evaluation scenario model at least comprises: A1: the decision logic rules of the inclinometer and the tire pressure detector are as follows: Lateral stability control: When the simulated tilt angle data received by the simulation model exceeds the first-level threshold θ1 (3°), the decision logic will output the instruction of "cautious steering" to the driver model in the simulation; when it exceeds the second-level threshold θ2 (5°), the logic will forcibly limit the simulated steering wheel angle rate and trigger the simulated vehicle to automatically perform a deceleration operation. Tire pressure-related risks: When the simulated tire pressure data exceeds the safe threshold range of 8.0~9.0 bar, if the tire pressure is below 8.0 bar, the lateral stiffness parameter of the vehicle tire in the simulation is reduced to more realistically reproduce the decrease in handling caused by insufficient tire pressure in the model. If the tire pressure is above 9.0 bar, the adhesion coefficient between the tire and the road surface in the simulation model is reduced to simulate the physical characteristics of reduced tire contact area and decreased grip caused by excessive tire pressure. A2: The decision-making logic rules for high-definition cameras and collision avoidance radar are as follows: Warning phase: When a surrounding obstacle is detected and the calculated TTC (Time to Collision) is less than the preset threshold T1 (5.0 seconds), the decision logic triggers a warning on the dashboard of the simulated vehicle; Assisted braking phase: When TTC is less than the more stringent threshold T2 (3.0 seconds), the logic will override the original throttle input of the driver model and apply a graded partial braking force (0.3g) to the simulated vehicle. Emergency avoidance phase: When TTC is less than the limit threshold T3 (1.5 seconds) and there is a safe space to the side, the logic will instruct the driver model to perform an automatic emergency steering operation; Data fusion: The detection results of cameras and radar are fused. When both identify high risk and the data is consistent, the confidence level for triggering a decision is the highest. When only a single sensor identifies the risk, a more conservative early warning strategy is adopted. A3: The decision logic rules of the driver condition monitoring system are as follows: Tiered alerts: When continuous yawning or a slight distraction of the gaze away from the road for more than 2 seconds is detected, a voice alert is triggered in the simulation; Forced intervention: When the system detects that the eyes are closed for more than 1.5 seconds, the decision logic will force the simulated vehicle to turn on the hazard warning lights and automatically execute a series of operations of "deceleration - pull over - stop - turn off the engine", and record the event as a high-risk event. A4: The decision-making logic rules for detachable reflective signs are as follows: Following distance logic: In the evaluation model, the AI ​​behavior rules for vehicles in the following traffic flow are set: When the distance D between the following vehicle and the beam transport vehicle is less than 50 meters, the AI ​​of the following vehicle will perform an active lane change to avoid the vehicle; if it cannot change lanes, it will be forced to decelerate to maintain a safe distance of more than 50 meters.

[0010] Preferably, the safety evaluation indicators in S3 include, but are not limited to, accident rate, number of serious traffic conflicts, cumulative exposure time to high-risk conditions, and frequency of dangerous driving operations.

[0011] Preferably, the comprehensive risk level k output in S4 is determined in the following way: Generation of overturning risk level k1: The cumulative times Tb and Te of the beam transport vehicle's lateral tilt angle exceeding the safety threshold α (5°) during 1000 runs of the baseline and evaluation scenario models are statistically analyzed using the formula: Calculate the overturning risk reduction rate R1, and map the numerical range of R1 to the preset overturning risk level k1; Generation of collision risk level k2: The number of severe traffic conflicts Cb and Ce occurring during multiple runs of the baseline scenario model and the evaluation scenario model are statistically analyzed. A severe traffic conflict is defined as a TTC (Time to Collision) less than a threshold β (1.5 seconds), calculated using the formula: Calculate the collision risk reduction rate R2, and map the numerical range of R2 to the preset collision risk level k2; Generation of driver state risk level k3: In multiple runs of the evaluation scenario model, the average frequency F of events triggering warnings or mandatory interventions due to simulated driver distraction or fatigue is statistically analyzed. e According to F e The numerical range is directly mapped to the preset driver status risk level k3; The overall risk level is obtained by weighting and fusing the rollover risk level, collision risk level, and driver status risk level. As shown in the following formula: In the formula, —Overturning risk level, —Collision risk level, —Driver's condition risk level; —Percentage of weight given to the risk of capsizing; —Percentage of collision risk; —Percentage weighting of driver condition risk.

[0012] Preferably, the intensity of the S4 early warning measures is positively correlated with the risk level, specifically including: Level 1 warning: When the overall risk level k is lower than the first preset threshold a, the dashboard graphic warning is triggered to provide a notification; Level 2 warning: When the comprehensive risk level k is greater than or equal to the first preset threshold a and lower than the second preset threshold b, the vehicle-mounted audible and visual alarm is triggered to issue an intermittent alarm. Level 3 warning: When the comprehensive risk level k is greater than or equal to the second preset threshold b, the vehicle-mounted audible and visual alarm is triggered to issue a continuous emergency alarm, and the alarm information and vehicle location data are uploaded to the remote monitoring platform.

[0013] Preferably, the control and early warning measures in S4 include using vehicle-mounted audible and visual alarms, dashboard graphic warnings, and remote monitoring and control platforms.

[0014] Preferably, the data recording unit in S4 is used to store the multi-source heterogeneous data, alarm records and operation logs for post-event analysis and responsibility determination.

[0015] Compared with the prior art, the present invention has the following beneficial effects: This invention constructs a comprehensive perception network integrating attitude, environment, driver status monitoring, and physical warnings through S1, utilizing 5G to ensure data real-time performance and eliminate physical blind spots. Subsequently, through digital twin simulation technology (S2 and S3), zero-cost, destructive, and massive trial-and-error testing is conducted in a virtual environment to verify the effectiveness of multi-dimensional decision-making logic, including lateral stability control, automatic emergency avoidance, and fatigue intervention. The fuzzy safety concept is transformed into a quantifiable comprehensive risk level k encompassing rollover, collision, and human factors. Finally, in conjunction with the execution phase (S4), precise hierarchical control is achieved based on the risk level, ranging from "minor warnings" to "mandatory intervention" and "remote linkage." This effectively avoids interference from invalid alarms to the driver and, through "vehicle-road-cloud" collaboration and full-process data recording, completely solves the three core pain points in precast beam transportation: high center of gravity leading to rollover, long vehicle body with large blind spots, and fatigue during long-distance transportation. This significantly improves the proactive safety of the transportation process, accident prevention capabilities, and the scientific nature of post-accident liability determination. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall process of the present invention. Detailed Implementation

[0017] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0018] like Figure 1 As shown, an intelligent safety monitoring and assessment method for long-distance transport of precast beams includes the following steps: S1: Risk control equipment installation: Sensor groups are installed on the beam transport vehicle to collect attitude data that characterizes the stability of the beam transport vehicle, environmental data that characterizes the external environmental risks, and multi-source heterogeneous data that characterizes the driver's state in real time, and transmit them through a 5G vehicle communication module; and detachable reflective signs are installed to warn other vehicles. The sensor group in S1 includes: The attitude sensing module includes at least one inclinometer, which is installed on the longitudinal center axis of the vehicle near the center of gravity, for real-time monitoring of the lateral tilt angle of the beam transport vehicle chassis, with a range of -90° to 90° and a measurement accuracy of 0.1°; and a tire pressure gauge, for real-time monitoring of tire pressure, with a monitoring range of 0 to 14.0 bar. The environmental perception module includes four high-definition cameras: one installed at the front of the vehicle, one at the rear, and the other two installed on the left and right sides of the vehicle, respectively. These cameras are used to acquire environmental image data within a range of 10 to 15 meters around the beam transport vehicle to eliminate blind spots. The camera image resolution is no less than 1920×1080 pixels. A collision avoidance radar is installed at the rear of the beam transport vehicle, with a detection range of 5 to 10 meters. The driver status perception module includes at least one driver-oriented fatigue warning device, which is used to identify and issue voice warnings for dangerous driving behaviors such as closing eyes, yawning, smoking, making phone calls, looking down, and looking around during the transportation of precast beams. The technical performance indicators of the fatigue early warning device must at least meet the following requirements: The accuracy rate for identifying the above-mentioned dangerous driving behaviors is ≥98%; System response time < 500 milliseconds; Working illuminance range: 5 Lux to 100,000 Lux; Image acquisition frame rate ≥ 30fps; It combines active monitoring by electronic devices with passive warnings by high-brightness reflective strips, effectively preventing rear-end collisions or scrapes with beam transport vehicles, especially at night or in poor visibility conditions.

[0019] The installation method for the detachable reflective sign in S1 is as follows: first, fix the detachable device to the two sides of the top of the precast beam with ropes, then attach the reflective sign to the long stainless steel strip, and then fix the stainless steel strip with the reflective sticker to both sides of the detachable device. The reflective sign adopts a yellow and white striped high-contrast pattern, and the retroreflection coefficient should not be less than 500cd / (lux·m²).

[0020] It solves the three major pain points of beam transport vehicles: "large vehicle size and large blind spots", "high center of gravity and easy to overturn", and "driver fatigue during long-distance transportation", ensuring that massive video and sensor data can be transmitted in milliseconds, providing communication support for subsequent real-time early warning.

[0021] S2: Simulation modeling, inputting multi-source heterogeneous data into traffic safety simulation software to establish a precast beam transfer benchmark scenario model without integrated risk control equipment and a precast beam transfer evaluation scenario model with integrated risk control equipment. Simulation modeling in S2 specifically includes: Establish a baseline scenario model. In the preset traffic safety simulation software, construct a baseline transportation scenario model without integrated risk control equipment logic based on the target transportation route, road structure characteristics, structural dimension parameters of the beam and beam transport vehicle, traffic flow, weather conditions, and transportation time.

[0022] An assessment scenario model is established. The baseline scenario model is replicated in simulation software, and the decision-making logic of the risk control equipment is embedded to form an assessment transportation scenario model.

[0023] Transporting precast beams is extremely dangerous. In reality, it is impossible to conduct destructive experiments, such as deliberately causing the vehicle to overturn to test the alarm. Simulation modeling provides a safe and low-cost testing platform. Before actual use, by comparing two models, we can intuitively see how many potential accidents this system can avoid and verify the effectiveness of the prevention and control logic.

[0024] S3: Risk assessment. Pre-set safety evaluation indicators for the transportation of precast beams. Under the same initial conditions, run the benchmark scenario model and the evaluation scenario model respectively, compare and analyze the simulation results, evaluate according to the preset safety evaluation indicators, and output a comprehensive risk level k that integrates overturning risk, collision risk and driver status risk. The risk assessment in S3 specifically includes: Run comparative simulations. Under the same initial conditions of running speed and acceleration, run the baseline scenario model and the evaluation scenario model no less than 1000 times respectively. Compare the running posture of the beam transport vehicle on straight road sections, slopes, curves and road intersections under the two scenarios, and collect preset safety evaluation index data. Quantitatively calculate and output a comprehensive risk level k. Transforming the vague concept of "safety" into a specific numerical risk level k makes it easier for management to intuitively grasp the status of transportation risks.

[0025] The decision logic for evaluating risk control devices embedded in the scenario model includes at least the following: A1: The decision logic rules for inclinometers and tire pressure gauges are as follows: Lateral stability control: When the simulated tilt angle data received by the simulation model exceeds the first-level threshold θ13°, the decision logic will output a "cautious steering" instruction to the driver model in the simulation; when it exceeds the second-level threshold θ25°, the logic will forcibly limit the simulated steering wheel angle rate and trigger the simulated vehicle to automatically perform a deceleration operation. Tire pressure-related risks: When the simulated tire pressure data exceeds the safe threshold range of 8.0~9.0 bar, if the tire pressure is below 8.0 bar, the lateral stiffness parameter of the vehicle tire in the simulation is reduced to more realistically reproduce the decrease in handling caused by insufficient tire pressure in the model. If the tire pressure is above 9.0 bar, the adhesion coefficient between the tire and the road surface in the simulation model is reduced to simulate the physical characteristics of reduced tire contact area and decreased grip caused by excessive tire pressure.

[0026] A2: The decision-making logic rules for high-definition cameras and collision avoidance radar are as follows: Warning phase: When a surrounding obstacle is detected and the calculated TTC collision time is less than the preset threshold T15.0 seconds, the decision logic triggers a warning on the dashboard of the simulated vehicle; Assisted braking phase: When TTC is less than the more stringent threshold T23.0 seconds, the logic will override the original throttle input of the driver model and apply a graded partial braking force of 0.3g to the simulated vehicle; Emergency avoidance phase: When TTC is less than the limit threshold T31.5 seconds and there is a safe space in the lateral direction, the logic will instruct the driver model to perform an automatic emergency steering operation; Data fusion: The detection results of cameras and radar are fused. When both identify high risk and the data is consistent, the confidence level for triggering a decision is the highest. When only a single sensor identifies the risk, a more conservative early warning strategy is adopted.

[0027] A3: The decision logic rules of the driver condition monitoring system are as follows: Tiered alerts: When mild distraction behaviors such as continuous yawning or looking away from the road for more than 2 seconds are detected, a voice alert is triggered in the simulation.

[0028] Forced intervention: When the system detects that the eyes are closed for more than 1.5 seconds, the decision logic will force the simulated vehicle to turn on the hazard warning lights and automatically execute a series of operations of "deceleration - pull over - stop - turn off the engine", and record the event as a high-risk event.

[0029] A4: The decision-making logic rules for detachable reflective signs are as follows: Following distance logic: In the evaluation model, the AI ​​behavior rules for vehicles in the following traffic flow are set: When the distance D between the following vehicle and the beam transport vehicle is less than 50 meters, the AI ​​of the following vehicle will perform an active lane change to avoid the vehicle; if it cannot change lanes, it will be forced to decelerate to maintain a safe distance of more than 50 meters.

[0030] The logic of A1-A4 covers the entire process from minor risk to extreme danger, avoiding a "one-size-fits-all" approach to alarms. It can automatically save lives in critical moments, such as by forcing the brakes, while reducing unnecessary interference with the driver.

[0031] Safety evaluation indicators in S3 include at least: accident rate, number of serious traffic conflicts, cumulative exposure time to high-risk conditions, and frequency of dangerous driving operations.

[0032] It comprehensively considers three dimensions: vehicle rollover, environmental collision, and human fatigue, thus avoiding the one-sidedness of evaluation based on a single indicator.

[0033] S4: Implement control and early warning measures, carry out corresponding risk control measures according to the risk level, and record alarm information through the recording unit.

[0034] The overall risk level k output in S4 is determined in the following way: Generation of overturning risk level k1: The cumulative time Tb and Te during 1000 runs of the baseline and evaluation scenario models, where the lateral tilt angle of the beam transport vehicle exceeded the safety threshold α5°, were calculated using the formula: Calculate the overturning risk reduction rate R1, and map the numerical range of R1 to the preset overturning risk level k1.

[0035] The collision risk level k2 is generated by statistically analyzing the number of severe traffic conflicts (Cb and Ce) occurring in multiple runs of the baseline and evaluation scenario models. A severe traffic conflict is defined as a TTC collision time less than a threshold β1.5 seconds, calculated using the formula: Calculate the collision risk reduction rate R2, and map the numerical range of R2 to the preset collision risk level k2.

[0036] Generation of driver state risk level k3: In multiple runs of the evaluation scenario model, the average frequency F of events triggered by simulated driver distraction or fatigue behavior, leading to warnings or mandatory interventions, is statistically analyzed. e According to F e The numerical range is directly mapped to the preset driver status risk level k3.

[0037] The overall risk level is obtained by weighting and fusing the rollover risk level, collision risk level, and driver condition risk level. As shown in the following formula: In the formula, —Overturning risk level, —Collision risk level, —Driver's condition risk level; —Percentage of weight given to the risk of capsizing; —Percentage of collision risk; —Percentage weighting of driver condition risk.

[0038] The intensity of S4 warning measures is positively correlated with the risk level, specifically including: Level 1 Warning: When the overall risk level k is lower than the first preset threshold a, a graphical warning will be triggered on the dashboard to provide a notification; Level 2 warning: When the comprehensive risk level k is greater than or equal to the first preset threshold a and lower than the second preset threshold b, the vehicle-mounted audible and visual alarm is triggered to issue an intermittent alarm. Level 3 warning: When the comprehensive risk level k is greater than or equal to the second preset threshold b, the vehicle-mounted audible and visual alarm is triggered to issue a continuous emergency alarm, and the alarm information and vehicle location data are uploaded to the remote monitoring platform.

[0039] By using tiered warning systems, a piercing alarm is only issued when there is a genuine danger, thus preventing drivers from shutting down the system or becoming annoyed by frequent false alarms.

[0040] The control and early warning measures in S4 include using vehicle-mounted audible and visual alarms, dashboard graphic warnings, and remote monitoring and control platforms.

[0041] The three-level early warning upload platform makes transportation not just a matter for the driver alone, but also allows back-end management personnel to intervene and command in real time, realizing collaborative management of "vehicle-road-cloud".

[0042] The data recording unit in S4 is used to store multi-source heterogeneous data, alarm records, and operation logs for post-event analysis and responsibility determination.

[0043] Similar to an aircraft's "black box," in the event of an accident or dispute, complete data records provide irrefutable legal evidence for accident cause analysis and liability determination.

[0044] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A smart safety monitoring and assessment method for long-distance transport of precast beams, characterized in that, Includes the following steps: S1: Risk control equipment installation: Install sensor groups on the beam transport vehicle to collect attitude data characterizing the stability of the beam transport vehicle, environmental data characterizing the external environmental risks, and multi-source heterogeneous data characterizing the driver's state in real time, and transmit them through a 5G vehicle communication module; and install detachable reflective signs. S2: Simulation modeling, inputting the multi-source heterogeneous data into traffic safety simulation software to establish a precast beam transfer benchmark scenario model without integrated risk control equipment and a precast beam transfer evaluation scenario model with integrated risk control equipment. S3: Risk assessment. Pre-set safety evaluation indicators for the transportation of precast beams. Under the same initial conditions, run the benchmark scenario model and the evaluation scenario model respectively, compare and analyze the simulation results, evaluate according to the preset safety evaluation indicators, and output a comprehensive risk level k that integrates overturning risk, collision risk and driver status risk. S4: Implement control and early warning measures, execute corresponding risk control measures according to the risk level, and record alarm information through the recording unit.

2. The intelligent safety monitoring and evaluation method for long-distance transport of precast beams according to claim 1, characterized in that: The sensor group in S1 includes: Attitude perception module, including several inclinometers, is installed on the longitudinal centerline of the vehicle near the center of gravity; tire pressure monitoring system; The environmental perception module includes four high-definition cameras: one installed at the front of the vehicle, one at the rear, and two on the left and right sides of the vehicle, respectively; and one collision avoidance radar installed at the rear of the beam transport vehicle. Driver status perception module, including fatigue warning device; The installation method of the detachable reflective sign in S1 includes, but is not limited to, fixing the detachable device to the two sides of the top of the precast beam with ropes, attaching the reflective sign to the long stainless steel strip, and then fixing the stainless steel strip with the reflective sticker to both sides of the detachable device. The reflective sign adopts a high-contrast pattern of yellow and white stripes.

3. The intelligent safety monitoring and evaluation method for long-distance transport of precast beams according to claim 1, characterized in that: The simulation modeling in S2 specifically includes: Establish a baseline scenario model. In the preset traffic safety simulation software, based on the target transportation route, road structure characteristics, structural dimension parameters of the beam and beam transport vehicle, traffic flow, weather conditions, and transportation time, construct a baseline transportation scenario model that does not integrate the logic of the aforementioned risk prevention and control equipment. An evaluation scenario model is established. The baseline scenario model is copied into the simulation software, and the decision-making logic of the risk control equipment is embedded to form an evaluation transportation scenario model.

4. The intelligent safety monitoring and evaluation method for long-distance transport of precast beams according to claim 1, characterized in that: The risk assessment in S3 specifically includes: Run comparative simulations. Under the same initial conditions of running speed and acceleration, run the baseline scenario model and the evaluation scenario model respectively. Compare the running posture of the beam transport vehicle on straight road sections, slopes, curves and road intersections in the two scenarios. Collect preset safety evaluation index data, quantify and calculate and output a comprehensive risk level k.

5. The intelligent safety monitoring and evaluation method for long-distance transport of precast beams according to claim 4, characterized in that, The decision-making logic of the risk control equipment embedded in the assessment scenario model includes at least the following: A1: The decision logic rules for inclinometers and tire pressure gauges are as follows: Lateral stability control: When the simulated tilt angle data received by the simulation model exceeds the first-level threshold θ1 (3°), the decision logic will output the "careful steering" instruction to the driver model in the simulation; when it exceeds the second-level threshold θ2 (5°), the logic will forcibly limit the simulated steering wheel angle rate and trigger the simulated vehicle to automatically perform a deceleration operation. Tire pressure-related risks: When the simulated tire pressure data exceeds the safe threshold range of 8.0~9.0 bar, if the tire pressure is below 8.0 bar, the lateral stiffness parameter of the vehicle tire in the simulation is reduced to more realistically reproduce the decrease in handling caused by insufficient tire pressure in the model. If the tire pressure is above 9.0 bar, the adhesion coefficient between the tire and the road surface in the simulation model is reduced to simulate the physical characteristics of reduced tire contact area and decreased grip caused by excessive tire pressure. A2: The decision-making logic rules for high-definition cameras and collision avoidance radar are as follows: Warning phase: When a surrounding obstacle is detected and the calculated TTC (Time to Collision) is less than the preset threshold T1 (5.0 seconds), the decision logic triggers a warning on the dashboard of the simulated vehicle; Assisted braking phase: When TTC is less than the more stringent threshold T2 (3.0 seconds), the logic will override the original throttle input of the driver model and apply a graded partial braking force (0.3g) to the simulated vehicle. Emergency avoidance phase: When TTC is less than the limit threshold T3 (1.5 seconds) and there is a safe space to the side, the logic will instruct the driver model to perform an automatic emergency steering operation; Data fusion: The detection results of cameras and radar are fused. When both identify high risk and the data is consistent, the confidence level for triggering a decision is the highest. When only a single sensor identifies the risk, a more conservative early warning strategy is adopted. A3: The decision logic rules of the driver condition monitoring system are as follows: Tiered alerts: When continuous yawning or a slight distraction of the gaze away from the road for more than 2 seconds is detected, a voice alert is triggered in the simulation; Forced intervention: When the system detects that the eyes are closed for more than 1.5 seconds, the decision logic will force the simulated vehicle to turn on the hazard warning lights and automatically execute a series of operations of "deceleration - pull over - stop - turn off the engine", and record the event as a high-risk event. A4: The decision-making logic rules for detachable reflective signs are as follows: Following distance logic: In the evaluation model, the AI ​​behavior rules for vehicles in the following traffic flow are set: When the distance D between the following vehicle and the beam transport vehicle is less than 50 meters, the AI ​​of the following vehicle will perform an active lane change to avoid the vehicle; if it cannot change lanes, it will be forced to decelerate to maintain a safe distance of more than 50 meters.

6. The intelligent safety monitoring and evaluation method for long-distance transport of precast beams according to claim 4, characterized in that: The safety evaluation indicators in S3 include, but are not limited to, accident rate, number of serious traffic conflicts, cumulative exposure time to high-risk conditions, and frequency of dangerous driving operations.

7. The intelligent safety monitoring and evaluation method for long-distance transport of precast beams according to claim 1, characterized in that: The overall risk level k output in S4 is determined in the following way: Generation of overturning risk level k1: The cumulative times Tb and Te of the beam transport vehicle's lateral tilt angle exceeding the safety threshold α (5°) during 1000 runs of the baseline and evaluation scenario models are statistically analyzed using the formula: Calculate the overturning risk reduction rate R1, and map the numerical range of R1 to the preset overturning risk level k1; Generation of collision risk level k2: The number of severe traffic conflicts Cb and Ce occurring during multiple runs of the baseline scenario model and the evaluation scenario model are statistically analyzed. A severe traffic conflict is defined as a TTC (Time to Collision) less than a threshold β (1.5 seconds), calculated using the formula: Calculate the collision risk reduction rate R2, and map the numerical range of R2 to the preset collision risk level k2; Generation of driver state risk level k3: In multiple runs of the evaluation scenario model, the average frequency F of events triggering warnings or mandatory interventions due to simulated driver distraction or fatigue is statistically analyzed. e According to F e The numerical range is directly mapped to the preset driver status risk level k3; The overall risk level is obtained by weighting and fusing the rollover risk level, collision risk level, and driver status risk level. As shown in the following formula: In the formula, —Overturning risk level, —Collision risk level, —Driver's condition risk level; —Percentage of weight given to the risk of capsizing; —Percentage of collision risk; —Percentage weighting of driver condition risk.

8. The intelligent safety monitoring and evaluation method for long-distance transport of precast beams according to claim 7, characterized in that: The intensity of the S4 early warning measures is positively correlated with the risk level, specifically including: Level 1 warning: When the overall risk level k is lower than the first preset threshold a, the dashboard graphic warning is triggered to provide a notification; Level 2 warning: When the comprehensive risk level k is greater than or equal to the first preset threshold a and lower than the second preset threshold b, the vehicle-mounted audible and visual alarm is triggered to issue an intermittent alarm. Level 3 warning: When the comprehensive risk level k is greater than or equal to the second preset threshold b, the vehicle-mounted audible and visual alarm is triggered to issue a continuous emergency alarm, and the alarm information and vehicle location data are uploaded to the remote monitoring platform.

9. The intelligent safety monitoring and evaluation method for long-distance transport of precast beams according to claim 8, characterized in that: The control and early warning measures in S4 include using vehicle-mounted audible and visual alarms, dashboard graphic warnings, and remote monitoring and control platforms.

10. The intelligent safety monitoring and evaluation method for long-distance transport of precast beams according to claim 9, characterized in that: The data recording unit in S4 is used to store the multi-source heterogeneous data, alarm records and operation logs for post-event analysis and responsibility determination.

Images