Real-time management and control system of power plant vehicles based on Beidou positioning technology

Abstract

The application discloses a real-time management and control system of power plant vehicles based on Beidou positioning technology, and relates to the field of real-time management and control of power plant vehicles, and comprises a data module, a constraint module, an authorization module, an event module, an evolution module, a traceability module and an output module.The data module is used for collecting vehicle management and control data and real-time process state data and preprocessing; the constraint module is used for generating a working condition adaptive safety constraint space; the authorization module is used for generating a working condition task authorization state; the event module is used for generating a credible vehicle operation event sequence; the evolution module is used for generating a vehicle safety risk event and a risk level; the traceability module is used for generating a responsibility traceability evidence chain; and the output module is used for generating a closed-loop management and control strategy and updating a real-time management and control result.The application adopts working condition decoupling and dynamic constraint, realizes closed-loop management and control of power plant vehicles, and has the advantages of real-time safety and responsibility traceability.

Description

Technical Field

[0001] This invention relates to the field of real-time management and control of power plant vehicles, and more particularly to a real-time management and control system for power plant vehicles based on BeiDou positioning technology. Background Technology

[0002] Currently, in the field of real-time vehicle management in power plants, existing technologies typically rely on Beidou positioning terminals, electronic fences, and manual dispatch records to monitor vehicle locations and issue alarms when vehicles enter fixed restricted areas or deviate from set routes. However, these methods are mostly based on static area rules and are difficult to dynamically adjust safety areas, road traffic relationships, and task authorization status in conjunction with the real-time process status of the power plant, which can easily lead to false alarms, missed alarms, or delayed management.

[0003] At the same time, existing technologies for judging abnormal vehicle behavior are mostly limited to determining whether a vehicle has crossed a boundary or deviated from its route. They lack decoupling analysis of risk sources such as spatial intrusion, task deviation, operational condition coupling, and delayed response. They also have difficulty in linking alarm information with the basis for tracing responsibility, which makes it impossible to quickly form a closed-loop management strategy and traceable handling results after a vehicle safety risk occurs. Summary of the Invention

[0004] One objective of this invention is to propose a real-time control system for power plant vehicles based on BeiDou positioning technology. This invention employs operating condition decoupling and dynamic constraints to achieve closed-loop control of power plant vehicles, possessing the advantages of real-time safety and traceability of responsibility.

[0005] A real-time control system for power plant vehicles based on BeiDou positioning technology according to an embodiment of the present invention includes: The data module is used to collect vehicle control data and real-time process status data, and generate standardized control data after preprocessing. The constraint module is used to generate a basic safety constraint space and a dynamic process coupling region based on standardized control data, and to update the basic safety constraint space using the dynamic process coupling region to generate an operating condition adaptive safety constraint space. The authorization module is used to match vehicle tasks with operating conditions based on standardized control data and adaptive safety constraint space, and generate operating condition task authorization status. The event module is used to input standardized control data, adaptive safety constraint space of operating conditions and authorization status of operating conditions into the event layer of the process model of operating condition decoupling point, perform road reachability verification and trajectory reliability screening, and generate a reliable vehicle operation event sequence. The evolution module is used to input the trusted vehicle operation event sequence into the decoupling layer, decouple and generate spatial intrusion factor, task deviation factor, operating condition coupling factor and handling delay factor, and modify the event intensity based on the operating condition coupling factor to generate vehicle safety risk events and risk levels. The traceability module is used to generate vehicle safety alarm information based on vehicle safety risk events and risk levels, and input vehicle safety alarm information, spatial intrusion factors, task deviation factors, operating condition coupling factors and handling delay factors into the traceability layer to generate a chain of evidence for responsibility traceability. The output module is used to generate closed-loop control strategies based on vehicle safety alarm information and the chain of evidence for liability tracing, and to update real-time control results.

[0006] Optionally, the vehicle management data includes vehicle Beidou positioning data, vehicle task data, vehicle identity data, driver data, factory road data, factory safety area data, and vehicle safety alarm handling data. The real-time process status data includes process risk status, work permit status, and equipment operating status. The preprocessing includes data alignment, coordinate unification, trajectory cleaning, business association, process status merging, and time sequence encapsulation.

[0007] Optionally, the constraint module includes: A basic security constraint space is constructed based on standardized management and control data; Identify process risk areas and process impact periods based on real-time process status data, and generate dynamic process coupling regions; Based on real-time process status data, a working condition influence coefficient is generated, and based on the working condition influence coefficient, boundary expansion, road restriction strength correction, and safety level correction are performed on the dynamic process coupling area to generate working condition constraint quantities. By updating the regional boundary relationships, road traffic relationships, and basic safety rules in the basic safety constraint space using the operating condition constraints, an adaptive safety constraint space for operating conditions is generated.

[0008] Optionally, the authorization module includes: Read vehicle task data from standardized management and control data, and extract the task area, task route and task time period corresponding to the vehicle task; Based on the adaptive safety constraint space of the working condition, spatial matching calculations are performed on the task area, task route and task time period to obtain the task space matching result; Based on the task space matching results, the executable status of the vehicle task is determined, and the working condition task authorization status is generated.

[0009] Optionally, the event module includes: Based on the work condition task authorization status and standardized control data, generate candidate vehicle operating trajectories; Standardized control data, adaptive safety constraint space for operating conditions, and authorized status of operating conditions are input into the event layer of the process model at the operating condition decoupling point; The decoupling point process model includes an event layer, a decoupling layer, and a tracing layer; In the event layer, road reachability verification is performed on the candidate vehicle trajectories, and road reachability verification results are generated. Based on the road reachability verification results, a reliable trajectory screening is performed on the candidate vehicle trajectories to remove location points that do not meet the requirements of road reachability, temporal continuity, and authorization validity, thereby generating a set of reliable trajectory points. Based on a set of trusted trajectory points, a trusted sequence of vehicle operation events is generated.

[0010] Optionally, the evolution module includes: The trusted vehicle operation event sequence is input into the decoupling layer, and the trusted vehicle operation events are encoded to generate vehicle operation event feature vectors. Risk source decoupling projection and factor scaling are performed on the feature vector of vehicle operation events to generate spatial intrusion factor, task deviation factor, operating condition coupling factor and handling delay factor; Based on the spatial intrusion factor, mission deviation factor, operating condition coupling factor, and handling delay factor, the basic intensity of vehicle risk is generated, and the basic intensity of vehicle risk is coupled and modulated based on the operating condition coupling factor to generate the modified event intensity. Based on the revised event intensity, as well as spatial intrusion factors, mission deviation factors, operating condition coupling factors, and handling delay factors, vehicle safety risk events and risk levels are generated.

[0011] Optionally, the traceability module includes: Generate vehicle safety warning information based on vehicle safety risk events and risk levels; Input vehicle safety alarm information, spatial intrusion factors, task deviation factors, operating condition coupling factors, and handling delay factors into the traceability layer; In the tracing layer, evidence nodes for spatial intrusion, task deviation, operational condition coupling, and delayed handling are generated respectively. A chain-like association is executed on evidence nodes of spatial intrusion, mission deviation, operational condition coupling, and delayed handling to generate a chain of evidence for tracing responsibility.

[0012] Optionally, the output module includes: The type of control action is determined based on vehicle safety alarm information and the chain of evidence for liability tracing, and a closed-loop control strategy is generated. Link the closed-loop management strategy with vehicle safety alarm information to generate the strategy distribution content; The system collects execution feedback corresponding to the content of the strategy, updates the status of vehicle safety alarm information and the chain of evidence for liability tracing based on the execution feedback, and generates updated real-time control results.

[0013] The beneficial effects of this invention are: This invention proposes a real-time control system for power plant vehicles based on BeiDou positioning technology. By generating a basic safety constraint space and a dynamic process coupling region, and updating the basic safety constraint space using the dynamic process coupling region, the system enables vehicle control rules to dynamically adjust according to real-time operating conditions such as ammonia leaks, hot work, unit maintenance, and equipment isolation. Furthermore, this invention employs a process model based on operating condition decoupling points to decouple four types of risk sources—space intrusion, task deviation, operating condition coupling, and delayed response—from a reliable vehicle operation event sequence. It also corrects event intensity based on operating condition coupling factors, allowing risk assessment to move beyond simple boundary crossings or route deviations. Instead, it combines real-time process status to determine the specific source and severity of vehicle risks, thereby reducing false alarms, missed alarms, and control delays caused by static fencing rules.

[0014] This invention also inputs vehicle safety alarm information along with spatial intrusion factors, task deviation factors, operating condition coupling factors, and response delay factors into the traceability layer to generate a chain of evidence for responsibility tracing. This enables alarm results to establish a correspondence with the source of risk, time of occurrence, location of occurrence, and situation of response delay. The output module further generates a closed-loop control strategy based on the vehicle safety alarm information and the chain of evidence for responsibility tracing, and updates the real-time control results based on execution feedback. This achieves closed-loop management from risk identification and alarm tracing to strategy issuance and result updating, improving the real-time performance, interpretability, and responsibility tracing capabilities of power plant vehicle safety control. Attached Figure Description

[0015] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is an overall flowchart of a real-time control system for power plant vehicles based on BeiDou positioning technology proposed in this invention. Figure 2 This invention presents a flowchart of the adaptive safety constraint space construction and vehicle task authorization process for a real-time control system for power plant vehicles based on BeiDou positioning technology. Figure 3 This is a flowchart illustrating the process of generating reliable vehicle operation events, constructing risk evolution, and establishing a chain of evidence for tracing responsibility in a real-time control system for power plant vehicles based on BeiDou positioning technology, as proposed in this invention. Detailed Implementation

[0016] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0017] refer to Figures 1-3A real-time control system for power plant vehicles based on BeiDou positioning technology includes: The data module is used to collect vehicle control data and real-time process status data, and generate standardized control data after preprocessing. The constraint module is used to generate a basic safety constraint space and a dynamic process coupling region based on standardized control data, and to update the basic safety constraint space using the dynamic process coupling region to generate an operating condition adaptive safety constraint space. The authorization module is used to match vehicle tasks with operating conditions based on standardized control data and adaptive safety constraint space, and generate operating condition task authorization status. The event module is used to input standardized control data, adaptive safety constraint space of operating conditions and authorization status of operating conditions into the event layer of the process model of operating condition decoupling point, perform road reachability verification and trajectory reliability screening, and generate a reliable vehicle operation event sequence. The evolution module is used to input the trusted vehicle operation event sequence into the decoupling layer, decouple and generate spatial intrusion factor, task deviation factor, operating condition coupling factor and handling delay factor, and modify the event intensity based on the operating condition coupling factor to generate vehicle safety risk events and risk levels. The traceability module is used to generate vehicle safety alarm information based on vehicle safety risk events and risk levels, and input vehicle safety alarm information, spatial intrusion factors, task deviation factors, operating condition coupling factors and handling delay factors into the traceability layer to generate a chain of evidence for responsibility traceability. The output module is used to generate closed-loop control strategies based on vehicle safety alarm information and the chain of evidence for liability tracing, and to update real-time control results.

[0018] In this embodiment, vehicle management data includes vehicle Beidou positioning data, vehicle task data, vehicle identity data, driver data, factory road data, factory safety area data, and vehicle safety alarm handling data. Real-time process status data includes process risk status, work permit status, and equipment operating status. Preprocessing includes data alignment, coordinate unification, trajectory cleaning, business association, process status merging, and time sequence encapsulation.

[0019] In this embodiment, the constraint module includes: A basic security constraint space is constructed based on standardized management and control data; The generation of the basic safety constraint space specifically includes: constructing a road traffic map from the factory area road data, overlaying the factory area safety zone data onto the road traffic map, and associating the road nodes and road edges in the road traffic map with the corresponding area boundaries to generate a basic safety constraint space that includes road traffic relationships, area boundary relationships, and basic safety rules; Identify process risk areas and process impact periods based on real-time process status data, and generate dynamic process coupling regions; The generation of dynamic process coupling regions specifically includes: reading the process risk status, work permit status and equipment operation status from real-time process status data, and extracting the status type, status location, status start time, status end time and status source from them. When the status type represents ammonia area leakage, hot work, unit maintenance or other process risk status, the plant area corresponding to the status location is taken as the core area, and the dynamic process coupling region is generated according to the spatial adjacency relationship between the core area and adjacent roads and adjacent work areas. Based on real-time process status data, a working condition influence coefficient is generated, and based on the working condition influence coefficient, boundary expansion, road restriction strength correction, and safety level correction are performed on the dynamic process coupling area to generate working condition constraint quantities. The generation of the operating condition impact coefficient specifically includes: reading the state type and the process risk classification table in the power plant safety management regulations; matching the state type with the process risk classification table to obtain the state basic risk level; reading the regional basic safety level of the area corresponding to the state location; fusing the state basic risk level and the regional basic safety level to generate the state risk quantity. The regional basic safety level is determined based on the fixed risk attributes of the corresponding area in the plant's safety area data. The level fusion adopts a high-level priority approach combined with an incremental approach to adjacent levels. The state source is read, and the source credibility level is determined based on the state source. The source credibility level is determined according to the power plant's distributed control system and work permit system. The order of the unified and manual confirmation records is determined from high to low. The specific level value is set according to the false alarm rate, false alarm rate and confirmation delay statistics of each status source in the power plant's historical process status records. The status start time and current processing time are read, the status duration is calculated, and the corresponding status reference period is determined according to the status type. The status reference period is the standard time length required to complete the response and handling for a certain type of real-time process status. The status duration is divided by the status reference period to obtain the duration ratio. The status duration is generated by a piecewise monotonically increasing mapping function. The status risk quantity, status credibility quantity and status duration quantity are normalized and then weighted and fused to generate the operating condition influence coefficient. The generation of working condition constraint quantities specifically includes: reading the core area boundary coordinate parameters and road connectivity of the dynamic process coupling area, reading the basic safety level in the basic safety constraint space, obtaining the traffic weight parameter based on the weighted sum of road length and road level, calculating the distance offset of the core area boundary coordinates based on the working condition influence coefficient to obtain the boundary expansion amount, multiplying the working condition influence coefficient with the traffic weight parameter, normalizing it to obtain the road restriction strength, weighting the working condition influence coefficient with the basic safety level and mapping it to the level to obtain the corrected safety level, and encapsulating the boundary expansion amount, road restriction strength and corrected safety level in a unified manner to form the working condition constraint quantities used to update the basic safety constraint space; By updating the regional boundary relationships, road traffic relationships, and basic safety rules in the basic safety constraint space using the load condition constraint quantities, an adaptive safety constraint space for the load condition is generated. The generation of the adaptive safety constraint space for operating conditions specifically includes: parameterizing the regional boundary relationships, road traffic relationships, and basic safety rules in the basic safety constraint space into boundary coordinate parameters, road connectivity weight parameters, and basic safety level parameters, respectively; using the boundary extension quantity in the operating condition constraints to perform distance superposition calculation on the boundary coordinate parameters to generate updated regional boundary coordinates; using the road restriction intensity in the operating condition constraints to perform multiplicative correction calculation on the road connectivity weight parameters to generate updated road connectivity weights; using the corrected safety level in the operating condition constraints to perform weighted fusion calculation on the basic safety level parameters to generate updated safety constraint intensity; and aligning the three types of update results in terms of spatial mapping relationships to generate a unified adaptive safety constraint space for operating conditions.

[0020] In this embodiment, the authorization module includes: Read vehicle task data from standardized management and control data, and extract the task area, task route and task time period corresponding to the vehicle task; Based on the adaptive safety constraint space of the working condition, spatial matching calculations are performed on the task area, task route and task time period to obtain the task space matching result; The task space matching results include the number of regions matched, the number of routes matched, and the number of times matched. The generation of the region matching quantity specifically includes: mapping the task region to the region boundary relationship in the adaptive safety constraint space of the working condition; reading the area of ​​the task region, the overlapping area between the task region and the allowed entry area, the overlapping area between the task region and the restricted entry area, and the overlapping area between the task region and the prohibited entry area; dividing the three types of overlapping areas by the area of ​​the task region respectively to obtain the overlapping ratio of the allowed area, the overlapping ratio of the restricted area, and the overlapping ratio of the prohibited area; and weighting the three types of overlapping ratios to generate the region matching quantity, which is used to characterize the degree of spatial adaptation between the vehicle task region and the region boundary relationship in the adaptive safety constraint space of the working condition. The generation of route matching quantity specifically includes: mapping the task route to the road traffic relationship in the adaptive safety constraint space of the working condition; reading the total length of the task route, the length of the passable road segment, the length of the restricted road segment, and the length of the prohibited road segment in the task route; dividing the length of the three types of road segments by the total length of the task route respectively to obtain the proportion of passable routes, the proportion of restricted routes, and the proportion of prohibited routes; and weighting the three types of route proportions to generate the route matching quantity, which is used to characterize the degree of traffic adaptation between the vehicle's task route and the road traffic relationship in the adaptive safety constraint space of the working condition. The generation of the time matching quantity specifically includes: mapping the task time period to the effective time period corresponding to the basic safety rule in the adaptive safety constraint space of the working condition; reading the total duration of the task time period, the overlap duration between the task time period and the allowed execution time period, the overlap duration between the task time period and the restricted execution time period, and the overlap duration between the task time period and the prohibited execution time period; dividing the three types of overlap duration by the total duration of the task time period respectively to obtain the overlap ratio of allowed execution time period, the overlap ratio of restricted execution time period, and the overlap ratio of prohibited execution time period; and weighting the three types of overlap ratios to generate the time matching quantity, which is used to characterize the degree of time adaptation between the vehicle task time period and the effective time period corresponding to the basic safety rule in the adaptive safety constraint space of the working condition; The executable status of vehicle tasks is determined based on the task space matching results, and the working condition task authorization status is generated. The generation of the work condition task authorization status specifically includes: jointly determining the area matching quantity, route matching quantity, and time matching quantity to generate a field-based work condition task authorization status, specifically including an authorization status field, a restriction source field, and an authorization valid time period field. The authorization valid time period field is generated based on the overlap between the task time period and the corresponding valid time period of the basic safety rules in the work condition adaptive safety constraint space. When there is a proportion of prohibited areas overlapping in the area matching quantity, a proportion of prohibited routes overlapping in the route matching quantity, or a proportion of prohibited execution time periods overlapping in the time matching quantity, the authorization status field is written as "prohibited execution," and the corresponding prohibited area, prohibited road, and prohibited time period are specified. When there is no overlap of prohibited areas, prohibited routes, or prohibited execution times, and there is overlap of restricted areas, restricted routes, or restricted execution times in the area matching data, or in the time matching data, the authorization status field is written to restricted execution, and the corresponding restricted area, restricted road, and restricted time period are written to the restriction source field. When there is no overlap of prohibited areas, prohibited routes, prohibited execution times, restricted areas, restricted routes, or restricted execution times, the authorization status field is written to allowed execution.

[0021] In this embodiment, the event module includes: Based on the work condition task authorization status and standardized control data, generate candidate vehicle operating trajectories; The generation of candidate vehicle trajectories specifically includes: sorting the BeiDou positioning data of vehicles with the same vehicle number from first to last according to the positioning time to form a vehicle positioning point sequence; associating each positioning point in the vehicle positioning point sequence with vehicle identity data, vehicle task data and vehicle running time data to generate candidate vehicle trajectories. Standardized control data, adaptive safety constraint space for operating conditions, and authorized status of operating conditions are input into the event layer of the process model at the operating condition decoupling point; The decoupling point process model includes an event layer, a decoupling layer, and a tracing layer; The training process of the decoupling point process model specifically includes: collecting historical vehicle control data, historical real-time process status data, historical vehicle safety alarm handling data, and manual review records from the power plant; constructing training samples according to vehicle number and time sequence; the historical vehicle safety alarm handling data includes alarm time, alarm location, alarm type, handling action, and handling completion status; labeling vehicle safety risk events based on manual review records and historical alarm handling results; labeling risk levels according to the risk classification requirements in the power plant safety management regulations; generating spatial intrusion labels based on the overlap between the vehicle's historical operating location and restricted access areas, prohibited access areas, restricted access roads, and prohibited access roads; and generating task deviation labels based on the vehicle's actual operating area, The deviations of the actual operating road and actual operating time from the task area, task route, and task time period are generated, respectively. Specifically, they are calculated based on the area overlap ratio, route matching ratio, and time overlap ratio. The working condition coupling label is generated based on the spatial proximity, road association, and time overlap between the vehicle's historical operating events and the working condition-affected area, working condition-affected road, and working condition-affected time period. The handling delay label is calculated based on the authorization status change time, the time when the vehicle leaves the area or road corresponding to the source of the restriction, and the historical handling completion time. It is also generated by proportionalizing the response benchmark time under the corresponding task type, area level, or road level. The above labels are calculated from historical data and then corrected for abnormal samples by combining manual review records. During model training, training samples are input into the event layer to generate a reliable vehicle operation event sequence. This reliable vehicle operation event sequence is then input into the decoupling layer to generate spatial intrusion factors, task deviation factors, operating condition coupling factors, and handling delay factors. Based on the operating condition coupling factors, the event intensity is corrected, and the vehicle safety risk events and risk levels are output. Vehicle safety alarm information and the four types of risk factors are then input into the tracing layer to output a chain of evidence for liability tracing. The loss function consists of event identification loss, risk level loss, factor decoupling loss, intensity correction loss, and tracing consistency loss, with the event identification loss constraining the identification of vehicle safety risk events. The results are consistent with the risk event labels. The risk level loss is used to constrain the risk level to be consistent with the labeled level. The factor decoupling loss is used to constrain the four types of risk factors to fit the spatial intrusion label, task deviation label, working condition coupling label, and handling delay label, respectively. The intensity correction loss is used to constrain the intensity of the corrected event to be consistent with the historical risk severity. The Adam optimizer is used to iteratively update the model parameters during training. When the accuracy of vehicle safety risk event identification and the accuracy of risk level identification both meet the requirements of power plant vehicle safety management, the model is considered to have converged. The parameters of the event layer, decoupling layer, and tracing layer are saved for subsequent real-time inference. The generation of factor decoupling loss specifically includes: calculating the difference between the spatial intrusion factor and the spatial intrusion label output by the decoupling layer to generate spatial intrusion loss; calculating the difference between the task deviation factor and the task deviation label to generate task deviation loss; calculating the difference between the working condition coupling factor and the working condition coupling label to generate working condition coupling loss; calculating the difference between the handling lag factor and the handling lag label to generate handling lag loss; and summing the spatial intrusion loss, task deviation loss, working condition coupling loss and handling lag loss after normalization to generate factor decoupling loss. This invention makes the following specific improvements to the original spatiotemporal point process model: First, the original method of directly inputting vehicle positioning event sequences into the model is improved to inputting reliable vehicle operation events. Specifically, before vehicle events enter the model, road reachability verification, temporal continuity verification, and authorization validity verification are performed on the vehicle's BeiDou positioning data, considering road traffic relationships, regional boundary relationships, and work condition task authorization status. BeiDou drift points, abnormal jump points, and invalid authorization points are eliminated to generate a reliable vehicle operation event sequence, thereby improving the reliability of the model's input data. Second, the original method of calculating event intensity based solely on vehicle spatiotemporal events is improved to incorporating real-time process status-driven intensity modeling. Specifically, real-time process status data is introduced into the vehicle risk modeling process, based on dynamic process coupling regions and adaptive safety constraints. In terms of space, the correlation between vehicle operation events and the affected areas, roads, and time periods is calculated to generate operating condition coupling factors. These factors are then used to correct the basic intensity of vehicle risks, enabling the risk intensity to dynamically adjust with real-time changes in operating conditions such as ammonia leaks, hot work, and unit maintenance. Thirdly, the original model's output of single event intensity is improved to a four-source risk factor decoupling and responsibility tracing output. Specifically, vehicle operation events are decoupled from four dimensions: spatial constraints, task authorization, operating condition correlation, and response, generating spatial intrusion factors, task deviation factors, operating condition coupling factors, and response delay factors. Based on these four types of factors, vehicle safety risk events, risk levels, and responsibility tracing evidence chains are generated, enabling the model not only to identify risks but also to explain their sources and support subsequent closed-loop management. In the event layer, road reachability verification is performed on the candidate vehicle trajectories, and road reachability verification results are generated. The road reachability verification operation specifically includes: mapping adjacent positioning points in the candidate vehicle trajectory to road traffic relationships in the adaptive safety constraint space; reading the road nodes, road edges, and road traffic status corresponding to adjacent positioning points; generating a reachability verification pass result when there is a road connection between adjacent positioning points and the vehicle displacement and road length within the positioning point interval meet the vehicle speed constraint; and generating a reachability verification exception result when there is no road connection between adjacent positioning points or the vehicle displacement exceeds the vehicle speed constraint within the positioning point interval. Based on the road reachability verification results, a reliable trajectory screening is performed on the candidate vehicle trajectories to remove location points that do not meet the requirements of road reachability, temporal continuity, and authorization validity, thereby generating a set of reliable trajectory points. The generation of the trusted trajectory point set specifically includes: judging the credibility of the positioning points in the candidate vehicle running trajectory based on the road reachability verification results; retaining the positioning points that pass the road reachability verification as trusted trajectory points; marking the positioning points that fail the road reachability verification and cannot form a continuous road passage relationship with the preceding and following positioning points as abnormal jump points and removing them; marking the positioning points that are outside the authorized valid time period as untrusted points during the authorized time period and removing them; marking the positioning points whose road or area conflicts with the restriction source field in the work condition task authorization status as authorized relationship abnormal points and removing them; and combining the retained trusted trajectory points according to the vehicle number, positioning time, road, area and authorization status fields to form a trusted trajectory point set. Generate a reliable vehicle operation event sequence based on a set of reliable trajectory points; The generation of a trusted vehicle operation event sequence specifically includes: using a single trusted trajectory point in the trusted trajectory point set as an event unit, writing the vehicle number, task number, running time, road, region, authorization status field, and restriction source field to generate a trusted vehicle operation event; grouping the trusted vehicle operation events according to the vehicle number and arranging them in order of running time to form a trusted vehicle operation event sequence.

[0022] In this embodiment, the evolution module includes: The trusted vehicle operation event sequence is input into the decoupling layer, and the trusted vehicle operation events are encoded to generate vehicle operation event feature vectors. The generation of vehicle operation event feature vectors specifically includes: encoding the vehicle location field, operation time field, location area field, location road field, authorization status field, and restriction source field in a trusted vehicle operation event into event features, converting them into location coordinate components, time interval components, duration components, area constraint components, road passage components, authorization status components, and restriction source components, and concatenating them sequentially to obtain the vehicle operation event feature vector; Risk source decoupling projection and factor scaling are performed on the feature vector of vehicle operation events to generate spatial intrusion factor, task deviation factor, operating condition coupling factor and handling delay factor; The risk source decoupling projection is a four-dimensional risk source decoupling projection, which specifically includes spatial constraint projection, task authorization projection, operating condition association projection, and handling response projection, generating spatial intrusion representation, task deviation representation, operating condition coupling representation, and handling lag representation. The generation of spatial intrusion representation specifically includes: calculating the shortest distance between the vehicle's position coordinates and the area boundary coordinates to generate the boundary distance; calculating the lateral offset distance between the vehicle's position coordinates and the road centerline coordinates to generate the road offset; reading the area entry constraint level and the road traffic constraint level corresponding to the area and the road traffic constraint level, and performing level numerical processing on the area entry constraint level and the road traffic constraint level to generate the spatial constraint level, wherein the area entry constraint level is determined by the area risk attribute in the plant safety area data and the process risk type in the real-time process status data, and the road traffic constraint level is generated by the road attribute in the plant road data and the real-time process status data; reading the duration of the vehicle continuously being in the same area or the same road, and proportionalizing the duration with the allowable stay time of the corresponding area or road to generate the spatial duration; and performing a weighted sum after normalizing the boundary distance, road offset, spatial constraint level, and spatial duration to generate the spatial intrusion representation. The generation of the task deviation representation specifically includes: calculating the overlap ratio between the vehicle's current location and the task area in the vehicle task data; calculating the route matching ratio between the vehicle's current road and the task route in the vehicle task data; calculating the overlap ratio between the running time field and the authorized valid time period field; subtracting the task area overlap ratio, route matching ratio, and authorized time period overlap ratio from 1 to generate the area deviation, route deviation, and time deviation; reading the authorization status field, processing the level numerical values ​​to generate different standardized constraint values; and normalizing the area deviation, route deviation, time deviation, and standardized constraint values ​​respectively, then performing a weighted sum to generate the task deviation representation. The generation of the working condition coupling representation specifically includes: calculating the shortest distance between the vehicle location and the boundary of the working condition influence area, and scaling this shortest distance to the boundary expansion of the working condition influence area to generate the working condition spatial proximity quantity; calculating the road connectivity distance between the road where the vehicle is located and the road affected by the working condition, and scaling this road connectivity distance to the corresponding reachable distance in the factory road network to generate the working condition road association quantity; calculating the overlap duration between the running time field and the working condition influence period, and dividing the overlap duration by the duration of the vehicle running event to generate the working condition time overlap quantity; reading the safety level corresponding to the working condition influence area or the working condition influence road, and generating the working condition level quantity after the level is numerically processed; multiplying the working condition spatial proximity quantity, the working condition road association quantity, the working condition time overlap quantity, and the working condition level quantity after normalization to generate the basic working condition coupling value; and weighting the basic working condition coupling value to obtain the working condition coupling representation. The generation of the handling lag representation specifically includes: reading the status change time of the authorization status field, the restriction source field, and the departure time of the vehicle leaving the area, road, or time period corresponding to the restriction source; subtracting the status change time from the departure time to generate the response interval; reading the response base time under the corresponding task type, area level, or road level, and dividing the response interval by the response base time to generate the response ratio; reading the continuous dwell time of the vehicle in the area, road, or time period corresponding to the restriction source, and dividing the continuous dwell time by the response base time to generate the lag duration ratio; reading the area constraint level, road traffic constraint level, and time period constraint level corresponding to the restriction source field; after numerical processing of the levels, obtaining the area standardized constraint value, road standardized constraint value, and time period standardized constraint value; calculating the square root of the sum of the squares of the three types of standardized constraint values ​​to generate the source constraint quantity; and then normalizing the response ratio, lag duration ratio, and source constraint quantity respectively and performing a weighted sum to generate the handling lag representation. The factor scaling process specifically includes: performing a unified scaling mapping on the space intrusion representation, task deviation representation, operating condition coupling representation, and handling lag representation respectively, so that the four types of representations are mapped to the same numerical range, generating space intrusion factor, task deviation factor, operating condition coupling factor, and handling lag factor; Based on the spatial intrusion factor, mission deviation factor, operating condition coupling factor, and handling delay factor, the basic intensity of vehicle risk is generated, and the basic intensity of vehicle risk is coupled and modulated based on the operating condition coupling factor to generate the modified event intensity. The generation of the corrected event intensity specifically includes: constructing a risk factor vector from the spatial intrusion factor, task deviation factor, operating condition coupling factor, and handling delay factor; calculating the basic vehicle risk intensity by summing the squares of the risk factor vector; converting the operating condition coupling factor into a coupling modulation coefficient; multiplying the coupling modulation coefficient and the basic vehicle risk intensity and adding the bias to generate the corrected event intensity. The generation of the coupling modulation coefficient specifically includes: normalizing the operating condition coupling factor to generate a normalized operating condition coupling value; reading the operating condition level quantity corresponding to the operating condition coupling factor; and determining the modulation upper limit value based on the operating condition level quantity. The modulation upper limit value is determined according to the risk amplification requirements corresponding to different operating condition level quantities in the power plant safety management regulations. The value 1 is used as the basic modulation value when there is no operating condition coupling effect. The normalized operating condition coupling value is multiplied by the difference between the modulation upper limit value and the basic modulation value, and the calculation result is added to the basic modulation value to generate the coupling modulation coefficient. When the normalized operating condition coupling value is 0, the coupling modulation coefficient is equal to the basic modulation value. Based on the revised event intensity, as well as spatial intrusion factor, mission deviation factor, operating condition coupling factor, and handling delay factor, vehicle safety risk events and risk levels are generated. The generation of vehicle safety risk events specifically includes: normalizing the spatial intrusion factor, task deviation factor, operating condition coupling factor, and handling delay factor; squaring the normalized results; dividing each squared value by the sum of the four squared values ​​to generate the contribution ratios of spatial intrusion, task deviation, operating condition coupling, and handling delay; multiplying the corrected event intensity by the contribution ratios of spatial intrusion, task deviation, operating condition coupling, and handling delay to generate the risk components of spatial intrusion, task deviation, operating condition coupling, and handling delay; determining the dominant risk source based on the risk component with the largest value; and generating vehicle safety risk event types based on the dominant risk source, including spatial intrusion, task deviation, operating condition coupling, and handling delay types. The generation of risk levels specifically includes: performing interval mapping on the modified event intensity to generate a basic risk level. The level boundaries used for interval mapping are determined based on the risk classification requirements corresponding to vehicle entry into restricted areas, entry into prohibited areas, deviation from authorized routes, exceeding authorized time periods, and delayed handling in the power plant safety management regulations. The dominant risk source is determined based on the risk component with the largest numerical value, and the risk component percentage corresponding to the dominant risk source is read. The basic risk level is then adjusted against the dominant risk source to generate a risk level. Specifically, the basic risk level is quantified to generate a risk level. The base level value is calculated by multiplying the spatial intrusion contribution ratio by the spatial constraint level value to generate a spatial correction value. The task deviation contribution ratio is multiplied by the task deviation factor corresponding to the authorized status field to generate a task correction value. The working condition coupling contribution ratio is multiplied by the working condition level value to generate a working condition correction value. The handling delay contribution ratio is multiplied by the response ratio to generate a delay correction value. The spatial correction value, task correction value, working condition correction value, and delay correction value are normalized and summed to generate a level correction value. The base level value is added to the level correction value to generate a corrected level value, which is mapped to the risk level.

[0023] In this embodiment, the traceability module includes: Generate vehicle safety warning information based on vehicle safety risk events and risk levels; Vehicle safety alarm information includes the following fields: alarm number, vehicle number, event time, event location, risk event type, and risk level. Input vehicle safety alarm information, spatial intrusion factors, task deviation factors, operating condition coupling factors, and handling delay factors into the traceability layer; In the tracing layer, evidence nodes for spatial intrusion, task deviation, operational condition coupling, and delayed handling are generated respectively. The generation of evidence nodes specifically includes: normalizing four types of factors, multiplying the normalized results by the corrected level values ​​to obtain the evidence strength of spatial intrusion, task deviation, operating condition coupling, and handling delay; associating the four types of evidence strength with the corresponding factors and the corresponding vehicle safety alarm information to generate evidence nodes of spatial intrusion, task deviation, operating condition coupling, and handling delay. A chain-like association is performed on evidence nodes of spatial intrusion, mission deviation, operational condition coupling, and delayed handling to generate a chain of evidence for tracing responsibility. The generation of the evidence chain for liability tracing specifically includes: using vehicle safety alarm information as the main node of the evidence chain, and using spatial intrusion evidence nodes, task deviation evidence nodes, operating condition coupling evidence nodes, and delayed handling evidence nodes as branch nodes of the evidence chain; determining the time order of the evidence chain according to the event time field; determining the primary and secondary order of the evidence chain according to the strength of spatial intrusion evidence, task deviation evidence, operating condition coupling evidence, and delayed handling evidence; marking the evidence node with the strongest evidence as the dominant evidence node; and linking the main node, dominant evidence node, and other evidence nodes in a chain to generate the evidence chain for liability tracing.

[0024] In this embodiment, the output module includes: The type of control action is determined based on vehicle safety alarm information and the chain of evidence for liability tracing, and a closed-loop control strategy is generated. The closed-loop management strategy uses risk level as the basis for management intensity, risk event type as the basis for management action, and traceability content ranked first in the chain sorting results as the main management basis, including vehicle restriction strategy, route adjustment strategy, area exit strategy, and handling review strategy. Link the closed-loop management strategy with vehicle safety alarm information to generate the strategy distribution content; The generation of policy distribution content specifically includes: associating the closed-loop control policy with alarm number, vehicle number, event location, and event time, and writing the policy type, execution object, execution location, execution time range, and feedback requirements to generate policy distribution content. The execution object is the vehicle corresponding to the vehicle number, the execution location is determined according to the event location, and the execution time range is determined according to the event time and the corresponding handling time limit for the risk level. The handling time limit is determined by the handling requirements for different risk levels in the power plant vehicle safety management procedure. The execution feedback corresponding to the content of the collection strategy is collected, and the status of vehicle safety alarm information and liability tracing evidence chain is updated based on the execution feedback to generate updated real-time control results. The collection of execution feedback specifically includes: collection of policy reception status, vehicle location change, policy execution status, and disposal completion status. Among them, the policy reception status is used to indicate whether the policy content has been received, the vehicle location change is used to indicate whether the vehicle has left the corresponding risk location or entered the adjusted route, the policy execution status is used to indicate whether vehicle restriction, route adjustment, area exit, or disposal review has been executed, and the disposal completion status is used to indicate whether the vehicle's safety alarm information has completed closed-loop disposal. The real-time control results update specifically includes: matching the execution feedback with the execution object, execution location, execution time range, and feedback requirements in the policy issuance content; when the execution feedback indicates that the vehicle has completed the closed-loop control policy and the handling completion status is "completed", the vehicle safety alarm information is updated to the "handled" status, and the responsibility tracing evidence chain is updated to the "closed-loop" status; when the execution feedback indicates that the policy has not been received, the vehicle has not left the risk location, or the handling completion status is not "completed", the vehicle safety alarm information is updated to the "continuous control" status, and the unclosed-loop mark of the responsibility tracing evidence chain is retained. The updated real-time control results include the vehicle's current alarm status, closed-loop control policy, execution feedback, and responsibility tracing evidence chain status.

[0025] Example 1: To verify the feasibility of this system in implementation, it was applied to a real-time vehicle management scenario within a power plant. The power plant includes a main building, ammonia area, coal yard, oil depot, ash and slag transfer area, and maintenance work area. The plant's roads have fixed traffic rules, but these rules also change temporarily due to hot work, unit maintenance, equipment isolation, and the monitoring status of the ammonia area. Before implementation, vehicle management relied mainly on positioning terminals, static electronic fences, and manual verification by dispatchers. When operating conditions changed but the fence boundaries did not change synchronously, vehicles often continued to approach restricted areas along their original routes, the system only generated ordinary boundary violation warnings, or dispatching was delayed.

[0026] In this embodiment, continuous operation records were selected as verification samples, with a total of 36 in-plant vehicles connected. The vehicle positioning terminals uploaded vehicle BeiDou positioning data at a rate of once per second. During this period, the system collected approximately 964,000 vehicle positioning points, read 118 vehicle task records, 168 real-time process status records, and 57 vehicle safety alarm and handling records. The data module aligns, unifies coordinates, and cleans trajectories of vehicle BeiDou positioning data, vehicle task data, driver data, plant road data, plant safety area data, and real-time process status data to form standardized management and control data. On-site verification revealed that the maximum instantaneous positioning deviation in the steel structure obstruction area on the east side of the main plant reached 9.3m. Traditional systems in this area misjudged six short-term drifts as boundary violations. This system eliminated the corresponding abnormal jump points through road reachability verification and trajectory reliability screening.

[0027] The ammonia concentration at the monitoring point on the east side of the ammonia zone rose from 0.6 mg / m³ to 7.8 mg / m³ and remained there for 6 minutes and 40 seconds. Simultaneously, the work permit system indicated temporary passage of maintenance personnel on an adjacent road. Based on this, the constraint module generated a dynamic process coupling area, extending the eastern boundary of the ammonia zone 65m towards the adjacent road and updating the eastern branch road from normal passage to restricted passage. Maintenance vehicle C-17, whose original task route passed through this branch road, had its task authorization status updated from allowed to restricted execution by the authorization module, which also recorded the restricted road and restricted time period. The vehicle continued driving approximately 22.4m from the dynamic boundary. The event module retained five consecutive reliable location points to form a vehicle operation event. The evolution module generated a spatial intrusion factor, a task deviation factor, a condition coupling factor, and a handling delay factor, and adjusted the event intensity based on the condition coupling factor, generating a medium-risk alarm. When this alarm occurred, the vehicle was still approximately 18.5m from the fixed ammonia zone boundary, indicating that the system could identify condition-related risks before the static fence was triggered.

[0028] The output module generates a route adjustment strategy based on vehicle safety alarm information and the chain of evidence for liability tracing, suggesting that C-17 detour via the north side passage of the main plant. The vehicle terminal subsequently receives the strategy, the driver decelerates and turns, leaving the area affected by the dynamically restricted road. The system updates the alarm status to "handled" and the chain of evidence for liability tracing to "closed loop." Simultaneously, an outsourced construction vehicle within the road affected by unit maintenance is identified as being under restricted execution status. The system generates a handling review strategy, and on-duty personnel subsequently complete the permission verification and release the vehicle, avoiding work interruptions caused by a direct prohibition of passage.

[0029] To facilitate the evaluation of the implementation effect of this system, the daytime continuous operation records of the power plant were selected for comparative verification using the same caliber. Both the traditional method and this system used the same vehicle range, the same data collection period, the same operating time, and the same evaluation samples.

[0030] Table 1 Comparison of Real-time Vehicle Management and Control in the Factory Area (using the same criteria)

[0031] As shown in Table 1, the average response time of the traditional method is 436 seconds. This is mainly because static fences cannot actively read real-time process status, requiring dispatchers to verify work orders, monitoring footage, and on-site feedback after an alarm. This system directly incorporates process risk status, work permit status, and equipment operating status into the constraint space update, reducing the response time to 28 seconds. It can update restricted roads and time periods before vehicles enter fixed boundaries. The number of false alarms for location anomalies decreased from 17 to 4 because the event layer first performs road reachability verification and trajectory reliability screening, eliminating jump points caused by obstructions from the main plant's steel structure.

[0032] In terms of risk identification, the system achieves a 94.8% accuracy rate in identifying road conditions affecting traffic, with a false negative rate reduced to 3.2%, significantly outperforming traditional methods. This performance improvement stems from the system's ability to not only determine whether a vehicle has crossed a static fence, but also incorporate vehicle tasks, dynamic process coupling areas, and operational task authorization status into its judgment. Furthermore, the system generates four types of factors at the decoupling layer: spatial intrusion, task deviation, operational condition coupling, and delayed response. This allows it to distinguish between ordinary route deviations and temporary risks caused by operational conditions. The average closed-loop response time has decreased from 12.6 minutes to 6.8 minutes, indicating that the traceability layer and output module can correlate alarms, risk sources, restriction fields, and execution feedback, enabling dispatchers to more quickly determine response actions and complete status write-back.

[0033] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A real-time control system for power plant vehicles based on BeiDou positioning technology, characterized in that, include: The data module is used to collect vehicle control data and real-time process status data, and generate standardized control data after preprocessing. The constraint module is used to generate a basic safety constraint space and a dynamic process coupling region based on standardized control data, and to update the basic safety constraint space using the dynamic process coupling region to generate an operating condition adaptive safety constraint space. The authorization module is used to match vehicle tasks with operating conditions based on standardized control data and adaptive safety constraint space, and generate operating condition task authorization status. The event module is used to input standardized control data, adaptive safety constraint space of operating conditions and authorization status of operating conditions into the event layer of the process model of operating condition decoupling point, perform road reachability verification and trajectory reliability screening, and generate a reliable vehicle operation event sequence. The evolution module is used to input the trusted vehicle operation event sequence into the decoupling layer, decouple and generate spatial intrusion factor, task deviation factor, operating condition coupling factor and handling delay factor, and modify the event intensity based on the operating condition coupling factor to generate vehicle safety risk events and risk levels. The traceability module is used to generate vehicle safety alarm information based on vehicle safety risk events and risk levels, and input vehicle safety alarm information, spatial intrusion factors, task deviation factors, operating condition coupling factors and handling delay factors into the traceability layer to generate a chain of evidence for responsibility traceability. The output module is used to generate closed-loop control strategies based on vehicle safety alarm information and the chain of evidence for liability tracing, and to update real-time control results.

2. The real-time control system for power plant vehicles based on BeiDou positioning technology according to claim 1, characterized in that, The vehicle management data includes vehicle Beidou positioning data, vehicle task data, vehicle identity data, driver data, factory road data, factory safety area data, and vehicle safety alarm handling data. The real-time process status data includes process risk status, work permit status, and equipment operating status. The preprocessing includes data alignment, coordinate unification, trajectory cleaning, business association, process status merging, and time sequence encapsulation.

3. The real-time control system for power plant vehicles based on BeiDou positioning technology according to claim 1, characterized in that, The constraint module includes: A basic security constraint space is constructed based on standardized management and control data; Identify process risk areas and process impact periods based on real-time process status data, and generate dynamic process coupling regions; Based on real-time process status data, a working condition influence coefficient is generated, and based on the working condition influence coefficient, boundary expansion, road restriction strength correction, and safety level correction are performed on the dynamic process coupling area to generate working condition constraint quantities. By updating the regional boundary relationships, road traffic relationships, and basic safety rules in the basic safety constraint space using the operating condition constraints, an adaptive safety constraint space for operating conditions is generated.

4. The real-time control system for power plant vehicles based on BeiDou positioning technology according to claim 1, characterized in that, The authorization module includes: Read vehicle task data from standardized management and control data, and extract the task area, task route and task time period corresponding to the vehicle task; Based on the adaptive safety constraint space of the working condition, spatial matching calculations are performed on the task area, task route and task time period to obtain the task space matching result; Based on the task space matching results, the executable status of the vehicle task is determined, and the working condition task authorization status is generated.

5. A real-time control system for power plant vehicles based on BeiDou positioning technology according to claim 1, characterized in that, The event module includes: Based on the work condition task authorization status and standardized control data, generate candidate vehicle operating trajectories; Standardized control data, adaptive safety constraint space for operating conditions, and authorized status of operating conditions are input into the event layer of the process model at the operating condition decoupling point; The decoupling point process model includes an event layer, a decoupling layer, and a tracing layer; In the event layer, road reachability verification is performed on the candidate vehicle trajectories, and road reachability verification results are generated. Based on the road reachability verification results, a reliable trajectory screening is performed on the candidate vehicle trajectories to remove location points that do not meet the requirements of road reachability, temporal continuity, and authorization validity, thereby generating a set of reliable trajectory points. Based on a set of trusted trajectory points, a trusted sequence of vehicle operation events is generated.

6. A real-time control system for power plant vehicles based on BeiDou positioning technology according to claim 1, characterized in that, The evolution module includes: The trusted vehicle operation event sequence is input into the decoupling layer, and the trusted vehicle operation events are encoded to generate vehicle operation event feature vectors. Risk source decoupling projection and factor scaling are performed on the feature vector of vehicle operation events to generate spatial intrusion factor, task deviation factor, operating condition coupling factor and handling delay factor; Based on the spatial intrusion factor, mission deviation factor, operating condition coupling factor, and handling delay factor, the basic intensity of vehicle risk is generated, and the basic intensity of vehicle risk is coupled and modulated based on the operating condition coupling factor to generate the modified event intensity. Based on the revised event intensity, as well as spatial intrusion factors, mission deviation factors, operating condition coupling factors, and handling delay factors, vehicle safety risk events and risk levels are generated.

7. A real-time control system for power plant vehicles based on BeiDou positioning technology according to claim 1, characterized in that, The traceability module includes: Generate vehicle safety warning information based on vehicle safety risk events and risk levels; Input vehicle safety alarm information, spatial intrusion factors, task deviation factors, operating condition coupling factors, and handling delay factors into the traceability layer; In the tracing layer, evidence nodes for spatial intrusion, task deviation, operational condition coupling, and delayed handling are generated respectively. A chain-like association is executed on evidence nodes of spatial intrusion, mission deviation, operational condition coupling, and delayed handling to generate a chain of evidence for tracing responsibility.

8. A real-time control system for power plant vehicles based on BeiDou positioning technology according to claim 1, characterized in that, The output module includes: The type of control action is determined based on vehicle safety alarm information and the chain of evidence for liability tracing, and a closed-loop control strategy is generated. Link the closed-loop management strategy with vehicle safety alarm information to generate the strategy distribution content; The system collects execution feedback corresponding to the content of the strategy, updates the status of vehicle safety alarm information and the chain of evidence for liability tracing based on the execution feedback, and generates updated real-time control results.

Images