Railway station operation and maintenance business collaboration method for multiple systems

Abstract

The application discloses a kind of railway station operation maintenance business cooperation methods for multi-system, it is related to intelligent transportation and building automation technical field.The method includes: analyzing field multi-system real-time monitoring data, mapping obtains the unified operation state view of the logical state and physical location of the equipment of the whole station;In response to emergency events, the joint model is instantiated according to the pre-constructed target plan action sequence and based on the unified view, state matching and mode constraint comparison are carried out, and the action executability judgment result and mode conflict marker information are generated by incremental local simulation;Based on the judgment result, identify the failed action, use the joint model to search for equivalent resources and spatially accessible paths, and reconstruct the optimal alternative action sequence.The application solves the linkage failure problem caused by mode conflict or equipment failure under complex working conditions in traditional plans, realizes dynamic consistency verification and flexible resource reconstruction of emergency response, and guarantees the safety of station operation.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation and building automation technology, and in particular to a collaborative method for the operation and maintenance of railway passenger stations involving multiple systems. Background Technology

[0002] With the expansion of large railway passenger station buildings and the increasing complexity of electromechanical equipment systems, station operation and maintenance have shifted from single-system management to multi-professional collaborative control. The efficient linkage of subsystems such as building automation systems (BAS), automatic fire alarm systems (FAS), automatic fare collection systems (AFC), public information systems (PIS), and power supply in emergency scenarios such as fires, terrorist attacks, or equipment failures is directly related to the safety of tens of thousands of passengers and the efficiency of evacuation, and is the core issue of smart station construction.

[0003] Currently, emergency response at railway passenger stations primarily relies on pre-prepared static contingency plans or simple logical interlocks based on hard-wired connections. Existing integrated management platforms mostly remain at the level of centralized monitoring and alarm display of various subsystems. Interlocking control typically employs point-to-point triggering with fixed rules, such as a fire alarm directly triggering the start of a specific numbered fan. While Building Information Modeling (BIM) technology has been introduced, it is mostly used for static asset management or offline 3D visualization walkthroughs, and has not yet been deeply integrated into the calculation of real-time control logic.

[0004] However, existing technologies have limitations in dealing with complex dynamic scenarios. The core problem lies in the lack of real-time executability verification capabilities for contingency plans and the rigid resource substitution mechanism. Specifically, traditional methods cannot incrementally simulate changes in electrical topology or cumulative capacity effects after an action is executed within milliseconds. This leads to contingency plan instructions easily failing to execute due to violations of hidden mode consistency constraints (such as ring network conflicts or overload) when equipment is under maintenance, degraded, or operating in non-standard modes. Furthermore, the fault tolerance mechanism of existing systems mainly relies on one-to-one hard backup of physical equipment, lacking dynamic reconfiguration capabilities based on space and function. When both primary and backup equipment fail, the system cannot, like human experts, use spatial connectivity calculations (such as opening interlocking valves) to call upon functionally equivalent resources in adjacent areas for remediation, resulting in a break in the emergency response chain. Summary of the Invention

[0005] The purpose of this invention is to provide a collaborative method for the operation and maintenance of railway passenger stations across multiple systems, in order to solve the aforementioned problems in the prior art.

[0006] Technical solution: A collaborative method for railway passenger station operation and maintenance across multiple systems, including:

[0007] Based on the station's multi-source static configuration data, a joint model of multi-system modes and contingency plans was constructed, which aggregated the operational mode constraints, emergency plan action logic, and station spatial evacuation topology.

[0008] Analyze real-time monitoring data from multiple systems on-site to generate a unified operational status view that maps the logical status and physical location of all equipment in the station;

[0009] In response to emergency event triggering information, the target contingency plan action sequence is instantiated based on the multi-system mode and contingency plan joint model, and the target contingency plan action sequence is matched with the unified operation status view for status matching and mode constraint comparison, generating action executability judgment results and mode conflict marking information;

[0010] The system filters out the failed actions indicated in the action executability determination results and mode conflict marking information, retrieves functionally equivalent resources and spatially reachable paths in the multi-system mode and contingency plan joint model, and reconstructs and generates the optimal alternative action sequence that satisfies the current mode constraints.

[0011] Beneficial effects: The present invention addresses the problem of linkage failure caused by mode conflicts or equipment malfunctions in traditional contingency plans under complex operating conditions by achieving dynamic consistency verification and flexible resource reconstruction of emergency response, thus ensuring the safety of station operations. Attached Figure Description

[0012] Figure 1 This is a flowchart illustrating the steps of a collaborative method for railway passenger station operation and maintenance across multiple systems, as described in this application.

[0013] Figure 2 This is a flowchart illustrating the steps of the specific implementation mechanism for pattern conflict detection in the embodiments of this application.

[0014] Figure 3 This is a flowchart illustrating the steps of the incremental local simulation process in the embodiments of this application.

[0015] Figure 4 This is a flowchart illustrating the steps in this application embodiment to convert the underlying physical state of the device into a logical availability conclusion of the upper-layer action. Detailed Implementation

[0016] According to one aspect of this application, an overview of a collaborative approach based on pre-built models and state views is provided.

[0017] like Figure 1 As shown, this embodiment details the overall process of a collaborative method for railway passenger station operation and maintenance across multiple systems. This embodiment constructs a complete technical framework from on-site data perception to emergency response closed-loop, focusing on addressing how to achieve consistent management of operating modes and dynamic executability verification of emergency actions in a heterogeneous multi-system environment.

[0018] A pre-built model is needed as the foundation for collaboration. This embodiment establishes a pre-built multi-system mode and contingency plan joint model by parsing the multi-source static configuration data of the station. The multi-source static configuration data specifically includes station system list data, operation mode configuration data, emergency plan text data, and BIM spatial topology data. Among them, the station system list data records in detail the unique identifiers and physical location coordinates of the equipment and control points under each subsystem such as BAS, FAS, and AFC; the operation mode configuration data defines the target state set of each system under different scenarios such as normal mode, evacuation mode, and fire mode; the emergency plan text data contains descriptions of standard handling procedures under various emergencies; and the BIM spatial topology data provides the station building's zoning, passage connection relationships, and entrance and exit information.

[0019] Building upon this foundation, this embodiment employs formal modeling methods to construct a joint model of multi-system modes and contingency plans. Specifically, this model integrates a multi-system operation mode model, an emergency plan action sequence model, and a spatial evacuation topology model. The multi-system operation mode model defines each operation mode M as a triple containing a scenario identifier, a system state vector, and a set of constraints, and uses a predicate logic rule base to clarify the consistency constraints between modes. For example, in a fire mode targeting a smoke control zone, the constraint set explicitly prohibits the activation of fans supplying air to that zone. The emergency plan action sequence model transforms unstructured contingency plan text into a sequence of action nodes with time dependencies. The spatial evacuation topology model uses a graph theory structure to express the accessibility and connectivity of various areas of the station. This type of pre-constructed joint model provides a unified semantic benchmark and logical rules for subsequent real-time verification and reconstruction.

[0020] Step 101: Analyze the real-time monitoring data of multiple systems on site and map it to obtain a unified operating status view that represents the logical status and physical location of all equipment in the station.

[0021] In this step, the system receives real-time data streams from subsystems such as BAS, FAS, AFC, and PIS. Since each subsystem uses different communication protocols and data formats, these streams need to be cleaned and standardized. A unified operational status view is a normalized representation of the current operational status of all equipment in the station. Specifically, the system maps the collected raw status values, such as equipment operating status, fault alarm codes, maintenance isolation flags, and control mode settings, to status fields under a unified semantic framework. This status is then bound to spatial coordinates by pre-entered physical location information. For example, for a smoke exhaust fan, the record in its unified view includes not only whether it is currently running or stopped, but also its spatial partition ID and whether it is in local control mode. By constructing this view, the semantic barriers between the underlying heterogeneous systems are eliminated, enabling upper-level collaborative logic to make judgments based on a unified data source. In some implementations, this view also includes real-time passenger flow density data and environmental parameter data for each area, providing input for subsequent evacuation assessments.

[0022] Step 102: Respond to the emergency event trigger information, instantiate the target contingency plan action sequence based on the pre-built multi-system mode and contingency plan joint model, match the target contingency plan action sequence with the unified operation status view for status matching and mode constraint comparison, and generate action executability judgment results and mode conflict marking information.

[0023] Upon receiving emergency event trigger information such as a fire alarm, equipment malfunction alarm, or manual triggering, the system retrieves the corresponding contingency plan template from a pre-built model based on the event type and location. Instantiating the target contingency plan action sequence refers to transforming the abstract contingency plan template into a list of operation instructions for specific equipment instances. For example, activating a smoke exhaust fan in a certain area in the template is transformed into activating equipment with ID FAN. 01 Specific instructions.

[0024] The system performs a critical verification process. On one hand, it performs state matching, that is, checks whether the current state of the device involved in the action in the unified operational status view allows it to be operated. For example, if FAN... 01 If the system is currently under maintenance / isolation or without power, the action is deemed unexecutable. On the other hand, a mode constraint comparison is performed to check whether the action would violate the consistency constraints of the current operating mode. For example, if the current system is in a different priority linkage mode that prohibits operation of related air valves, a mode conflict is identified. The generated action executability determination results clearly identify whether each action is executable, conditionally restricted, or unexecutable; the mode conflict flag information records the specific constraint rules and severity that caused the conflict. This step ensures logical security before emergency commands are issued, preventing linkage failures due to equipment malfunctions or mode conflicts.

[0025] Step 103: Based on the action executability determination results and mode conflict marking information, identify failed actions, and use the pre-built multi-system mode and contingency plan joint model retrieval function to retrieve equivalent resources and spatial reachable paths to reconstruct and generate the optimal alternative action sequence that satisfies the current mode constraints.

[0026] For failed actions deemed unexecutable or exhibiting serious conflicts, the system initiates a resource reconfiguration mechanism. Searching for functionally equivalent resources refers to finding other devices or systems within the model that can perform the same or similar functions. For example, when the main exhaust fan fails, the system searches for a backup fan or whether the exhaust system of an adjacent zone can be used to cover the failure by opening specific valves. Searching for spatially reachable paths refers to finding interconnected alternative evacuation routes within the spatial topology model that avoid hazardous areas.

[0027] Based on this, the system constructs a local action sequence containing alternative resources and replaces the failed parts of the original plan with it, forming a new candidate sequence. By verifying the pattern consistency and timing feasibility of the new sequence, the system selects the solution with the lowest risk and highest efficiency as the optimal alternative action sequence. In some implementations, this process also includes converting the optimal alternative action sequence into control commands that can be recognized by each subsystem and issuing them for execution, using a unified operating status view to continuously monitor execution feedback and achieve closed-loop control. This type of dynamic reconfiguration mechanism improves the system's emergency response capability and robustness in complex fault scenarios.

[0028] According to one aspect of this application, a conflict detection mechanism based on action state evolution fragments.

[0029] like Figure 2 As shown, this embodiment elaborates on the specific implementation mechanism of pattern conflict detection. As the core verification engine in the cooperative method, this embodiment introduces an incremental local simulation algorithm, which solves the problems of large computational load and poor real-time performance in traditional full-system simulation, and achieves millisecond-level action conflict prediction.

[0030] Step 201: Traverse the action instances in the target contingency plan action sequence, extract the equipment state influence rules based on the pre-constructed multi-system model and contingency plan joint model, and generate action state evolution fragments that characterize the changes in key state variables before and after action execution.

[0031] In this step, the system analyzes each action instance in the target plan one by one. Equipment state impact rules are predefined in the joint model, describing the physical impact of specific types of actions on the system state. For example, for the action of closing a circuit breaker, the rule defines that the action will cause the circuit breaker's switching state to change from open to closed, and may change the connectivity of related electrical nodes. Based on this rule, the system generates action state evolution fragments. Action state evolution fragments are incremental data structures that do not record the state of the entire system, but only the changes or altered values ​​of local state variables affected by the action. For example, for the action of starting a wind turbine, the evolution fragment may only contain information on the changes in two key variables: the wind turbine's operating state being set to 1 and the corresponding increase in duct flow rate ΔQ. This lightweight data structure lays the foundation for subsequent efficient calculations.

[0032] Step 202: Perform local state evolution simulation based on the unified running state view and action state evolution fragment to obtain the local state prediction results after the action is executed.

[0033] like Figure 3 As shown, this step details the incremental local simulation process. To support this simulation, the system performs a refinement step: extracting equipment operating states, electrical topology connections, and loop load data associated with the current action from the unified operating state view, and constructing a local state snapshot containing equipment state vectors, topology connectivity state vectors, and load distribution vectors. Local State Snapshot S before It is the baseline of the simulation, which only contains the part of the system state that is physically or logically related to the current action.

[0034] Subsequently, the system performs the core incremental overlay operation: incrementally overlaying the action state evolution fragments onto the local state snapshot, updating the local state prediction results that characterize the device state and topology after the action execution. This process can be represented as the local state prediction result S. after Apply means that the rule is applied to S. before For example, in a power distribution circuit scenario, suppose the current load I of a feeder in a local state snapshot is... curr The current is 80 amps, and the action is to start a unit with a rated current of I. motor A high-power water pump with a capacity of 70 amps is used. At this point, the operational state evolution segment includes a load increase of 70 amps. The system calculates the predicted current I of this feeder in the local state prediction result through superposition calculations. pred The current is 150 amperes. For actions that change the valve state, the system updates the connectivity matrix of the nodes in the local topology graph, obtaining a new topology state. By calculating only the locally relevant variables, the system can predict the physical consequences after the action is performed very quickly.

[0035] Step 203: Retrieve the constraint rule set in the pre-built multi-system model and contingency plan joint model, verify whether the local state prediction results violate the preset model consistency constraints, and generate model conflict label information that records the conflict type and severity level.

[0036] After obtaining the prediction results, the system calls the pre-stored constraint rule set for verification. Specifically, the system calls the interlock constraints, capacity constraints, and topology safety constraints in the constraint rule set to perform threshold and logical verification on the topology connectivity state vector and load distribution vector in the local state prediction results.

[0037] Continuing with the example of the distribution circuit above, assume that the maximum safe current carrying capacity I of the feeder is defined in the constraint rule set. max The system predicts a current of 150 amps, which is 120 amps. If the predicted value exceeds the threshold of 120 amps, the system determines that a capacity constraint has been violated. For topology safety constraints, such as prohibiting two different power sources from simultaneously supplying power to the same bus, the system checks the predicted topology for illegal loops or dual-power-source paralleling points. Once a constraint violation is detected, the system generates mode conflict flag information. This information details the type of conflict, such as a capacity overload conflict or an electrical interlocking conflict, as well as the severity level of the conflict. For example, a minor overload might be flagged as Level 1, while a short-circuit risk might be flagged as a Level 3 fatal conflict. This flag information directly guides subsequent steps on whether an alternative solution needs to be found.

[0038] According to one aspect of this application, device availability is multidimensionally aggregated to meet motion requirements.

[0039] like Figure 4 As shown, this embodiment details how to transform the underlying physical state of the device into a logical availability conclusion for the upper-level actions. This embodiment provides a multi-dimensional aggregation calculation method capable of handling various complex action requirement logics such as full satisfaction, redundant satisfaction, and capacity satisfaction.

[0040] Step 401: Analyze the associated device set and action requirement logic of each action instance in the target contingency plan action sequence. The action requirement logic includes full satisfaction logic, minimum satisfaction logic, and capacity-based satisfaction logic.

[0041] In this step, the system clarifies which devices need to be operated for each action and what success criteria must be met. The associated device set refers to a list of all physical objects involved in the execution of the action. The action requirement logic defines the rules for determining whether an action is executable. Specifically, the "full satisfaction" logic requires all devices in the set to be available, commonly seen in the disconnection of critical safety loops; the "at least one satisfaction" logic requires at least one device in the set to be available, commonly seen in dual-machine systems that serve as backups; and the "capacity satisfaction" logic requires the total capacity of available devices in the set to meet a preset business demand, commonly seen in scenarios requiring smoke extraction or lighting brightness.

[0042] Step 402: Extract the real-time operation field of the associated device set from the unified operation status view, and map it to a discrete device availability level according to the preset standardization rules.

[0043] To standardize the representation of different types of equipment status, the system performs a standardized status mapping. The system reads the equipment's original status fields, such as operating status, fault codes, and control modes, and converts them into a unified equipment availability level according to preset rules. For example, level 3 represents fully available, meaning the equipment is fault-free, power is normal, and it is in automatic control mode; level 2 represents limited availability, such as the equipment being in local control mode but allowing manual intervention, or the equipment operating at degraded performance; level 1 represents only monitoring functionality and no controllability; and level 0 represents completely unavailable, such as equipment failure or maintenance isolation. Through discretized grading, the system can calculate the health status of individual equipment.

[0044] Step 403: Use the action requirement logic to aggregate the device availability level of each device in the associated device set, calculate the remaining capacity and redundancy required to meet the action execution, and generate an action executability judgment result that includes an executability label, capacity gap value and redundancy risk level.

[0045] This is the core calculation step in this embodiment. The system aggregates the levels of individual devices based on the parsed logical type.

[0046] For fully satisfied logic, the system checks if there are any devices in the set whose level is lower than a specific threshold, such as level 2. If so, the action is determined to be unexecutable.

[0047] For at least one logically satisfied device, the system statistical level meets the requirement. If the number is ≥1, the system is deemed executable, and the number of available devices is calculated as a redundancy index.

[0048] Specifically, N avail =Count({d j |L(d j )≥L min}); where N availThe number of available redundant devices, where Count represents the counting function, and d j For the j-th device in the set of associated devices, L(d) j L represents the availability level of the device. min This is the minimum availability level threshold required to meet the action execution requirements. This formula is used for redundancy calculations under at least one satisfying logic condition.

[0049] For logic based on capacity satisfaction, the system performs numerical accumulation calculations. For example, the total air volume required for a certain smoke extraction operation. req The capacity is 50,000 cubic meters per hour. The associated equipment set includes three fans: F1, F2, and F3. After mapping, F1 is classified as Class 3 with a rated airflow of 20,000 cubic meters per second; F2 is Class 2, operating at reduced capacity due to a power issue, with a current available airflow of 15,000 cubic meters per second; F3 is Class 0 and unavailable due to a fault. The system calculates the total available capacity C. total =20000+15000=35000; Subsequently, the system calculates the capacity gap C. gap C gap =Volume req -C total Among them, Volume req C represents the total air volume or total business demand required for the operation. total C represents the total available capacity of the associated device set, which is the sum of the available capacities of each device. gap =50000-35000=15000. Due to the existence of a gap, the system marks this action as partially satisfied or requiring substitution, and records the gap value. The generated action executability determination result not only includes a Boolean conclusion on whether it is executable, but also includes specific gap data and risk level, providing a calculation basis for subsequent accurate search for alternative resources.

[0050] In another alternative implementation, Judgment=(C total ≥α*C req Partial Satisfied Not Satisfied Where Judgment is the result of the capacity satisfaction logic, and C total The total available capacity of the device set, where α is a preset lower limit capacity ratio coefficient, for example, 0.5, C req For the target capacity required for the action, Partial Satisfied Indicates a partially satisfied state, Not Satisfied This indicates a non-compliant state. This formula is used to determine whether to accept a degraded operation or a partially compliant alternative.

[0051] According to one aspect of this application, alternative resource reconfiguration is based on functional units and spatial topology.

[0052] This embodiment details the process of identifying available alternative resources from both functional logic and spatial physics dimensions after a failed action is detected. The core technical problem addressed by this embodiment is how to automatically discover, when the original equipment specified in the contingency plan becomes unusable due to fault or conflict, hidden resources that are physically different but functionally equivalent, or that can cover the original target area through specific control methods, thereby maximizing emergency response capabilities.

[0053] Step 501: Based on the action executability determination result, select actions marked as unexecutable or having high-risk conflicts from the target plan action sequence to form a set of actions that need to be replaced.

[0054] In this step, the system reads the judgment result generated by the previous embodiment. This result contains the execution status label of each action. The system extracts the actions that need to be replaced according to a preset screening strategy. Specifically, actions marked as unexecutable are directly included in the set; actions marked as conditionally executable but accompanied by high-level risks, such as electrical operations with a risk level of 2 or higher, are also included in the set according to the safety strategy. The system packages this action with its index position in the original sequence, the original target area ID, and the required function type label to form a standardized set of actions that need to be replaced.

[0055] Step 502: Traverse the set of actions to be replaced, and use the pre-built multi-system model and the pre-stored functional equivalence relationship definition and spatial evacuation topology model in the contingency plan joint model to retrieve candidate equipment that can functionally cover the original target area and candidate evacuation channels that are spatially connected, and construct a set of alternative resource candidates.

[0056] This step is the overall control point for resource retrieval. To achieve accurate retrieval, the system performs further actions, specifically including:

[0057] Step 601: Map each action in the set of actions to be replaced to a corresponding functional unit, and retrieve candidate functional units that have an equivalent or similar relationship with the functional unit in the pre-built multi-system model and contingency plan joint model.

[0058] Here, a functional unit refers to a capability service node abstracted from a physical device after removing its specific model number. For example, for the failed action of starting exhaust fan A, the system maps it to the exhaust function unit in area A. Subsequently, the system queries the equivalence relation graph in the joint model. This graph predefines the substitute relationships between different functional units. For example, the graph defines an exhaust function unit in area B that has a similar relationship to the exhaust function unit in area A, provided that area A and area B are physically adjacent and connected. Through this mapping, the system finds potential candidate units, rather than being limited to backup units of the same model.

[0059] Step 602: Extract the coverage area of ​​the candidate functional unit, perform spatial overlay analysis on it and the target area of ​​the original action in the spatial evacuation topology model, identify the directly accessible resources of the target area contained in the coverage area and the convertible accessible resources that can form a connection path by changing the state of the access control or air valve, and generate a set of spatially accessible candidate alternative devices.

[0060] This step introduces convertible reachability analysis technology. The system obtains the physical service range of candidate functional units, such as the duct coverage area of ​​fan B. If this coverage area directly includes the target area of ​​the original action, it is marked as a directly reachable resource. More importantly, if the two do not overlap, the system uses a spatial evacuation topology model to perform path search, determining whether there are nodes whose on / off state can be changed through control actions, such as normally closed fire dampers or fire doors. If so, the system simulates changing the state of such nodes and then determines whether the service range of the candidate functional unit can be extended to the target area. For example, the system calculation finds that as long as the connecting damper V between area A and area B is opened... AB The suction effect of wind turbine B can cover area A. At this point, wind turbine B is identified as a convertible and reachable resource and added to the candidate set.

[0061] Step 603: The failed evacuation path nodes in the original plan are obstacle points. A local path search algorithm is executed on the spatial evacuation topology model to generate a set of candidate alternative evacuation paths that avoid the obstacle points and meet the evacuation direction requirements.

[0062] For evacuation guidance actions involving spatial movement, the system marks nodes unavailable due to fire or equipment failure as obstacles on the topology map, setting their passage weight to infinity. Subsequently, the system runs a path search algorithm, such as the A* algorithm or a variant of Dijkstra's algorithm, to search for the optimal path from the current passenger flow gathering point to the safety exit. During the search process, the system strictly adheres to evacuation direction requirements, prohibiting the generation of paths that involve running in the opposite direction or traversing areas with higher hazard levels.

[0063] Step 701: Obtain the real-time operating load and performance limitation status of each device in the set of spatially accessible candidate alternative devices from the unified operating status view.

[0064] To ensure the retrieved resources are indeed available at the current moment, the system further performs a real-time verification step. The system accesses real-time data and reads the current operating status of the equipment. For example, it reads the current operating frequency, current value, and inverter performance limit flags of fan B. If fan B is currently operating at full load, or if the operator has locked the maximum frequency limit, this information will be recorded in detail.

[0065] Step 702: Calculate the available redundancy capacity of each candidate device based on the device design parameters, and compare it with the capacity gap of the original actions in the set of actions to be replaced.

[0066] The system is based on the equipment's designed rated capacity C rated and current real-time load C load Calculate the available redundancy capacity C avail The specific calculation formula is C. avail =C rated *k perf -C load Among them, C avail C represents the available redundancy capacity of the equipment. rated k is the design rated capacity of the equipment. perf C is a performance factor used to indicate whether a device has performance limitations. load This represents the current real-time load of the equipment. For example, if fan B has a rated airflow of 60,000, currently outputting 20,000, and has no performance limitations, then its available redundancy capacity is 40,000. The system will then add this value to the original operating capacity gap C. gap For example, 15000, for comparison.

[0067] Step 703: Eliminate devices with insufficient available redundant capacity or those in a faulty state, and select device combinations that can make up for the capacity gap through single or multiple device combinations to form a candidate set of alternative resources that has been verified in real time.

[0068] Based on the comparison results, the system executes filtering logic. If the available redundancy capacity of a single device is greater than or equal to the gap, that device is selected as an independent candidate. If a single device is insufficient, the system attempts to combine multiple devices. For example, if the combined redundancy capacity of devices B and C exceeds the gap, it is selected as a combined solution. In the resulting set, each element is a set of effective resources that are functionally equivalent, spatially reachable, and currently have sufficient capacity to execute the task, providing a solid material foundation for the subsequent generation of specific control sequences.

[0069] According to one aspect of this application, alternative sequence construction and temporal consistency reconstruction.

[0070] This embodiment details how to transform selected static alternative resources into a dynamically executable sequence of control instructions. It focuses on resolving timing conflicts and deadlocks that may arise when alternative solutions are embedded into the original plan, ensuring the reconstructed plan is logically sound along the timeline.

[0071] Step 801: For each action in the set of actions to be replaced, select alternative equipment combinations or alternative evacuation routes from the candidate set of alternative resources that have been verified in real time, and instantiate a local alternative subsequence containing pre-preparation actions, core alternative actions and post-recovery actions based on the action template in the pre-built multi-system mode and contingency plan joint model.

[0072] In this step, the system no longer simply replaces the device ID, but reconstructs the action chain. The system calls a pre-stored action template in the federated model, which defines the standard process for operating a certain type of resource. For example, for the scheme of replacing fan A with fan B in the aforementioned embodiment, the system instantiates a local substitution sub-sequence containing three stages. In the pre-preparation action stage, a command is generated to open the interlocking air valve V. AB And confirm the arrival signal; in the core alternative action phase, generate a command to start fan B and adjust it to the specified frequency; in the post-recovery action phase, generate a command to shut down fan B and reset air valve V after the emergency ends. AB This type of structured subsequence enables complex cross-system borrowing operations to be executed correctly, avoiding the situation where a replacement device is directly started without establishing a physical path.

[0073] Step 802: Insert the local substitution subsequence into the position of the original substitution action in the target plan action sequence to form several sets of candidate substitution action sequences to be verified.

[0074] The system embeds the generated local sub-sequences as a whole module into the timeline of the original plan. To ensure smooth transitions, the system inherits all the pre-triggered conditions of the original failed action and links the subsequent dependent actions of the original action to the end node of the sub-sequence. The system generates multiple sets of candidate alternative action sequences, each corresponding to a different alternative resource combination scheme.

[0075] Step 901: Based on the action dependency rules and mutual exclusion constraints in the pre-constructed multi-system model and contingency plan joint model, abstract all actions in the candidate alternative action sequence into nodes, and abstract the sequential dependency and interlocking relationship between actions into directed edges to construct an action dependency graph.

[0076] To verify the logical feasibility of the new sequence, the system employs graph theory for modeling. The system treats all atomic actions in the sequence, both existing and newly inserted, as nodes in the graph. Edges are constructed based on the rules in the model. Dependency rules are transformed into directed edges; for example, if action X must be completed before action Y, an edge is created pointing from X to Y. Mutual exclusion constraints are transformed into bidirectional constraint edges or prohibition of coexistence markers; for example, if actions M and N are interlocked, this is represented in the graph as the two actions not being on parallel branches. Through this step, the linear list of actions is transformed into an action dependency graph capable of expressing complex logical relationships.

[0077] Step 902: Perform a topological sorting algorithm on the action dependency graph to detect whether there are dependency loops, remove sequences with loops, and determine the successfully sorted sequences as temporally consistent alternative action sequence candidates.

[0078] This is a crucial algorithmic step in timing verification. The system runs a topological sorting algorithm, such as the Kahn algorithm, on the action dependency graph. This algorithm attempts to arrange all nodes in the graph into a linear sequence such that for every directed edge from node U to node V, U is listed before V. If, during the sorting process, it is found that not all nodes can be output, i.e., a cycle exists in the graph, it usually indicates a logical deadlock. For example, a newly inserted alternative action requires the power to be turned off first, while a subsequent action retained in the original sequence requires the power to remain on to execute, thus creating a logical contradiction. The system automatically identifies and removes invalid sequences containing cycles. Only sequences that successfully complete the topological sorting and whose generated linear order conforms to all physical constraints are marked as temporally consistent alternative action sequence candidates and enter the evaluation phase.

[0079] According to one aspect of this application, scenario adaptive evaluation and closed-loop execution.

[0080] This embodiment further illustrates the closed-loop execution process of the solution. It provides a decision-making mechanism based on multi-dimensional indicators and scenario weights, enabling the system to intelligently balance safety, efficiency, and consistency in different emergency scenarios, and select the truly optimal solution.

[0081] Step 1001: Based on the event type in the emergency event triggering information, load the corresponding indicator weight vector. The indicator weight vector defines the weight ratio of safety, evacuation efficiency and pattern consistency in the current scenario.

[0082] The system identifies the current business scenario context. In the pre-built model, different event types are associated with different decision-making strategies. For example, for a fire scenario, the system loads an indicator weight vector W. fire In the middle, the security weight w safe Set to 0.7, evacuation efficiency weight w eff Set to 0.2, mode consistency weight w mode The value is set to 0.1, reflecting the principle of prioritizing life safety. For non-emergency equipment failure scenarios, the system loads vector W... fault In this scenario, the security weight may be adjusted to 0.4, while more emphasis is placed on the pattern consistency weight of 0.4 and the efficiency weight of 0.2, to reduce interference with normal operations.

[0083] Step 1002: Combine the unified operation status view to conduct a multi-dimensional evaluation of each temporally consistent alternative action sequence candidate, calculate the safety score, evacuation efficiency score and pattern consistency score, and use the index weight vector to calculate the comprehensive score.

[0084] The system calculates and scores each candidate sequence. The safety score primarily assesses the probability of equipment overload exceeding limits and the stability of the electrical topology during sequence execution; the evacuation efficiency score is calculated using normalized calculations based on path length and estimated evacuation time; and the mode consistency score evaluates the degree of deviation between the alternative and the original operating mode. The system obtains a comprehensive score, expressed by the formula: Score = w safe *S safe +w eff *S eff +w mode *S mode Wherein, Score is the overall score of the alternative action sequence, w safe For security weights, S safe For safety rating, w eff S is the weight for evacuation efficiency. eff To score evacuation efficiency, w mode S is the pattern consistency weight. mode Scoring is given for pattern consistency.

[0085] In another alternative implementation, Score total =w safe *S safe +w eff *S eff +w cons *S cons +w res *S res Among them, Score total For the overall score, w safe For security weights, S safe For safety rating, w eff S is the weight for evacuation efficiency. eff To score evacuation efficiency, w cons S is the pattern consistency weight. cons To score pattern consistency, w res For resource consumption and redundancy weights, S res Score resource usage and redundancy.

[0086] Step 1003: Select the sequence with the highest comprehensive score as the optimal alternative action sequence.

[0087] The system sorts all candidate sequences based on their comprehensive scores and selects the one with the highest score as the execution plan. In some optional implementations, if there are multiple plans with extremely close scores, the system can prompt a human operator to provide assistance for confirmation, or break the tie by prioritizing the plan with the fewest actions according to preset secondary sorting rules.

[0088] As a complete closed loop of the method, after determining the optimal alternative action sequence, the system translates the logical actions in the sequence into specific communication protocol instructions, which are then sent to the field subsystem controllers such as BAS and FAS via interfaces. The system continuously monitors the feedback signals of each device using a unified operational status view. If the execution result is detected to be inconsistent with the expected instruction, such as a wind turbine failing to start as scheduled, the system will trigger the aforementioned conflict detection and resource reconfiguration process, generating new corrective instructions, thereby achieving dynamic closed-loop control throughout the entire process. Even if a secondary failure occurs during execution, this mechanism can still maintain business continuity and security.

[0089] According to one aspect of this application, emergency event analysis and closed-loop model correction.

[0090] This embodiment details the event analysis process in the initial stage of emergency response and the model self-correction mechanism in the later stage of execution, ensuring the integrity of the method throughout its entire lifecycle from trigger input to feedback update.

[0091] Step 102: Respond to the emergency event trigger information, instantiate the target contingency plan action sequence based on the pre-built multi-system mode and contingency plan joint model, match the target contingency plan action sequence with the unified operation status view for status matching and mode constraint comparison, and generate action executability judgment results and mode conflict marking information.

[0092] As a further refinement of this step, this embodiment describes in detail the specific logic for responding to emergency event triggering information. The system reads the emergency event triggering information, which includes the event type, event location, event time, and event severity level. For example, it receives information from the fire alarm controller regarding an alarm from a smoke detector in the west area of ​​the station hall, with a severity level of Level 1.

[0093] Based on the above information, the system executes the contingency plan location. Specifically, the system uses the event type and location as index keys to search within a pre-built multi-system model and contingency plan joint model. The model stores a contingency plan index table, and the system filters out all candidate contingency plans applicable to the current location and type. Subsequently, the system considers the current station's operational mode configuration, such as whether it is currently in peak holiday mode, and selects the highest priority target contingency plan that best suits the current station structure from the candidate plans. This process ensures that the invoked contingency plan is most relevant to the current actual situation, rather than a generic static text.

[0094] Based on this, the system further performs a comprehensive judgment on executability. The system reads the device availability judgment result and the mode conflict marking information. The system applies preset comprehensive judgment rules to merge the two types of information. For example, if the device availability of an action is fully available and the mode conflict marking is conflict-free, then the final state of the action is marked as fully executable; if the device availability is limited, or there is a minor mode conflict of level 1, then it is marked as conditionally executable; if there is a device unavailable or a fatal conflict of level 3, then it is marked as unexecutable. This comprehensive judgment result is the direct basis for triggering the subsequent replacement mechanism.

[0095] As a further extension, this embodiment also describes in detail the closed-loop tracking and model update process after the action is executed.

[0096] After the optimal alternative action sequence is issued and executed, the system continuously monitors the execution feedback of each device based on a unified operational status view. The system records the actual execution status and execution time of each instruction and compares it with the expected action logic to form execution deviation record data. For example, if the system records that the actual time taken for a fireproof rolling shutter door to fully close from the issuance of the instruction is 45 seconds, while the model estimates 30 seconds, then a time deviation of 15 seconds is recorded.

[0097] After the linkage process is completed, the system uses the aforementioned execution deviation recording data to revise and update the pre-built multi-system model and contingency plan joint model. Specifically, the system can update equipment availability statistics, such as marking slow-responding equipment as having degraded performance; the system can also revise local constraints, such as adding new interlocking rules to the model based on interlocking issues discovered during actual execution. Through a closed-loop feedback mechanism, the multi-system model and contingency plan joint model can continuously evolve as the operation progresses, making subsequent emergency responses more accurate and efficient.

[0098] According to one aspect of this application, a gate and guidance linkage scenario based on passenger flow coordination.

[0099] This embodiment details the specific application of the present invention in passenger transport organization operations. It demonstrates how the method of the present invention can be used to achieve cross-system collaborative replacement of AFC, PIS, and BAS when a ticket gate malfunctions, causing passenger congestion.

[0100] In the scenario set in this embodiment, the station is in peak holiday mode. The emergency event trigger information is: a set of turnstiles at gate A on the station hall level experiences a sudden network failure, resulting in an 80% reduction in the gate's throughput capacity. The system instantiates the target contingency plan based on the pre-built model, and the original plan requires maintaining normal throughput at gate A.

[0101] Step 1001: Generate action executability determination results. The system, combined with the unified operational status view analysis, found that 8 out of the 10 turnstiles associated with gate A are offline, with an availability level of 0. Based on the capacity fulfillment logic, the current available throughput capacity of 20 people per minute is far below the planned demand of 100 people per minute, resulting in a capacity gap C. gap The rate is 80 people per minute. Therefore, the original action was deemed unexecutable.

[0102] Step 1002: Construct a candidate set of alternative resources. The system maps the ticket checking function of gate A to a functional unit and retrieves functionally equivalent resources in the model. The model defines the adjacent gate B as a candidate functional unit. The system performs spatial reachability analysis and finds that there is a physical barrier between gate A and gate B, but in the spatial evacuation topology model, there is a normally closed emergency evacuation door D between them. AB The system recognizes that if gate D is... AB If the status changes from closed to open, the coverage area of ​​gate B can be extended to the waiting area of ​​gate A. Therefore, the system will open gate D. AB+ The system identifies some turnstiles at Gate B as accessible alternatives. Gate C is also located further away, but by controlling the PIS screen to display diversion information, some subsequent passengers can be guided directly to Gate C.

[0103] Step 1003: Reconstruct and generate the optimal alternative action sequence. The system generates an alternative sequence that includes cross-system linkage. The specific steps of this sequence include: First step (PIS system): Displaying guidance information on the PIS screen in the entrance hall, asking passengers to proceed to gates B and C for waiting; Second step (BAS system): Remotely opening evacuation door D connecting areas A and B. AB The system will adjust the lighting brightness in the area to provide guidance; the third step (AFC system) will temporarily switch the five exit gates at gate B to entry mode (tidal control) to make up for the capacity shortage at gate A.

[0104] Step 1004 involves timing consistency verification and evaluation. The system constructs an action dependency graph to ensure that the opening of evacuation doors is completed before the guidance of passenger flow. In the evaluation phase, the system loads a weight vector for peak passenger flow scenarios, focusing on assessing evacuation efficiency scores. Calculations show that the combined scheme of PIS guidance + BAS door opening + AFC tidal switching has the highest overall score and is selected as the optimal alternative action sequence for execution. This embodiment fully demonstrates the versatility and flexibility of the method of this invention in non-fire-related operational coordination scenarios.

[0105] According to one aspect of this application, the mathematical model and formal definition of the core algorithm.

[0106] This embodiment details the key formulas, data structure definitions, and logical expressions used for calculation during system implementation, enabling those skilled in the art to reproduce the core algorithm of this invention based on this type of mathematical model.

[0107] For calculating device availability levels, this embodiment employs a multi-dimensional feature-weighted scoring model. During state standardization, the system constructs an availability feature vector (Avail) for each device. feature This vector contains the running state score F run Power status score F power Control mode score F ctrl Maintenance status score F maint And performance limit score F perf These scores are all normalized to the range of 0 to 1. For example, when the operating state is normal, F run Set to 1, and to 0 in case of a fault; when the control mode is automatic, F ctrl The value is 1 for local mode and 0.5 for in-situ mode. The system calculates the overall availability score of the device using a weighted summation formula. The specific formula is: Score = w r *F run +w p *F power +w c *F ctrl +w m *F maint +w l *F perf Among them, the score is the overall availability rating of the device, w r For the running state weights, F run For the running status score, w p For power state weights, F power Score the power supply status, w c To control the mode weights, F ctrl To score the control mode, w m To maintain the state weights, F maint To maintain the state score, w l For performance-constrained weights, F perf This is the performance limitation score. Among them, w r w p The values ​​are preset weighting coefficients for each dimension, and their sum is 1. The calculated Score is then mapped to a discrete availability level. For example, a Score ≥ 0.9 maps to Level 3 (fully available), 0.6-0.9 maps to Level 2 (limited availability), and < 0.6 maps to Level 0 (unavailable). This calculation method allows status assessment to no longer rely on simple thresholds, but rather to comprehensively reflect the health of the device.

[0108] Regarding the determination of action requirement logic, this embodiment provides a redundancy margin calculation formula based on capacity satisfaction logic. Let action Act... i The associated set of devices is D, and the target capacity required for the action is C. req For each device d in the set j C avail (d j )=C rated (d j )*Score(d j )*k limit (d j ); where C avail (d j ) for device d j Available capacity, C rated (d j ) for device d j Rated capacity, Score(d) j ) for device d j Usability score, k limit (d j ) for device d j The performance limitation factor. The total available capacity C of the system's computation set. total This is equal to the sum of the available capacity of all devices in the set. Based on this, the system calculates the redundancy margin. The formula is: Margin = C total -C req Where Margin represents the redundancy margin in action execution, and C total C represents the total available capacity of the device set. req Let Margin be the target capacity required for the action. If Margin ≥ 0, the action is considered executable; if Margin < 0, its absolute value is the capacity gap C that needs to be filled by alternative resources. gap The system also calculates the risk level R. level It is defined as the proportion of total available capacity provided by devices with an availability level below 3 in the set. This risk metric calculates the current scheme's reliance on unhealthy devices.

[0109] For verifying pattern consistency constraints, this embodiment uses predicate logic expressions for formal modeling. Constraint rules are defined as the structure IF Condition THEN Prohibition. Here, IF Condition defines the scenario or prerequisite for the rule to take effect, and THEN Prohibition defines the actions, states, or behaviors that must be prevented, restricted, or prohibited when the condition is true. Condition is a Boolean expression composed of local state variables. The local state vector S is defined as a quadruple: sdev ,s topo ,s load ,s env . s dev This represents a set of device states, for example, fan A: running, valve B: open; topo This represents the topology connectivity matrix, used to describe the connectivity of electrical or wind tunnel nodes; s load This represents the load value of the critical node; s env This indicates an environmental condition. For example, a constraint rule regarding the prevention of electrical short circuits is expressed as: IF(Breaker) A.state =CLOSED) and (Breaker) B.state =CLOSED) and (Topo.Connect(Node A Node B )=TRUE)and(Source) A.Phase ≠Source B.Phase The THEN conflict type is equal to asynchronous closing, and the severity is equal to 3. During incremental simulation, the system substitutes the state changes caused by the action into the above Condition expression. If the result is true, the corresponding conflict flag is triggered. This formalized expression allows the constraint verification logic to be automatically parsed and executed by a computer program.

[0110] To address the issue of lacking real-time executability verification capabilities and being prone to violating hidden pattern constraints, this solution employs an incremental local simulation mechanism based on action state evolution fragments. By constructing local state snapshots and superimposing incremental changes generated by actions, the system can accurately predict the physical impact of actions on electrical topology connectivity and loop loads before the actions are executed. This allows for the identification of dynamic conflicts that traditional static rules cannot detect within milliseconds, such as asynchronous closing and capacity exceeding limits, ensuring the logical security of command issuance.

[0111] To address the issues of rigid resource substitution mechanisms and a lack of spatial and functional reconfiguration capabilities, this solution employs a resource substitution reconfiguration method based on functional units and spatial topology. The system breaks down the one-to-one binding limitation of physical equipment, automatically calculating and identifying convertible and reachable implicit resources by analyzing the spatial evacuation topology model. For example, by changing valve states, adjacent area fans can cover the target area. This method enables the system to generate effective alternative action sequences through cross-regional and cross-system functional reconfiguration even in extreme conditions where both primary and backup equipment fail, thus improving the robustness and continuity of emergency response.

Claims

1. A railway station operation and maintenance service coordination method for multiple systems, characterized in that, Comprise: Analyzing field multi-system real-time monitoring data, mapping to obtain a unified operation state view representing the logical state and physical location of the entire station equipment; In response to emergency event trigger information, instantiating target preplan action sequence based on pre-built multi-system mode and preplan joint model, performing state matching and mode constraint comparison between target preplan action sequence and unified operation state view, generating action executability judgment result and mode conflict marking information; Identifying failed actions based on action executability judgment result and mode conflict marking information, using pre-built multi-system mode and preplan joint model to search for equivalent resources and spatially accessible paths, and reconstructing an optimal alternative action sequence that meets the current mode constraints; Performing state matching and mode constraint comparison between target preplan action sequence and unified operation state view, generating mode conflict marking information, comprising: Traversing action instances in target preplan action sequence, extracting equipment state influence rules based on pre-built multi-system mode and preplan joint model, and generating action state evolution segments representing changes in key state variables before and after action execution; Performing local state evolution simulation based on unified operation state view and action state evolution segments, obtaining local state prediction results after action execution; Searching constraint rule set in pre-built multi-system mode and preplan joint model, checking whether local state prediction results violate preset mode consistency constraints, and generating mode conflict marking information recording conflict type and severity level; Performing local state evolution simulation, comprising: Extracting device operation state, electrical topology connection relationship and loop load data associated with the current action from the unified operation state view, and constructing a local state snapshot containing device state vector, topology connectivity state vector and load distribution vector; Incrementally adding action state evolution segments to the local state snapshot to update the local state prediction results representing the device state and topology structure after action execution; Calling interlock constraints, capacity constraints and topology safety constraints in the constraint rule set to perform threshold and logical checking on the topology connectivity state vector and load distribution vector in the local state prediction results; Performing state matching and mode constraint comparison between target preplan action sequence and unified operation state view, generating action executability judgment result, comprising: Analyzing the associated device set and action requirement logic of each action instance in the target preplan action sequence, the action requirement logic including full satisfaction logic, at least satisfaction logic and capacity-based satisfaction logic; the action requirement logic defines the rules for determining action executability; Extracting real-time operation fields of the associated device set from the unified operation state view, and mapping them to discretized device availability levels according to preset standardized rules; Using action requirement logic to aggregate device availability levels of each device in the associated device set, calculating the remaining capacity and redundancy required to satisfy action execution, and generating action executability judgment result containing executability label, capacity gap value and redundancy risk level.

2. The method of claim 1, wherein, Using pre-built multi-system mode and preplan joint model to search for equivalent resources and spatially accessible paths, comprising: Filtering out the actions marked as unexecutable or with high risk conflicts from the target preplan action sequence based on the action executability determination results, to form a set of actions to be replaced; Traversing the set of actions to be replaced, using the pre-built multi-system mode and preplan joint model and the pre-stored functional equivalence relationship definition and space evacuation topology model to respectively search for candidate devices that can functionally cover the original target area and candidate evacuation channels that are spatially connected, to build a candidate set of replacement resources.

3. The method of claim 2, wherein, Building the candidate set of replacement resources includes: Mapping each action in the set of actions to be replaced to a corresponding functional unit, and searching for a candidate functional unit with an equivalent relationship or a similar relationship with the functional unit in the pre-built multi-system mode and preplan joint model; Extracting the coverage area of the candidate functional unit, and performing spatial overlay analysis on the coverage area and the target area of the original action in the space evacuation topology model to identify directly reachable resources whose coverage areas contain the target area and convertible reachable resources that can form a connected path by changing the access control or damper state, to generate a set of candidate replacement devices that are spatially reachable; The failed evacuation path nodes in the original preplan are obstacle points, and a local path search algorithm is performed on the space evacuation topology model to generate a set of candidate replacement evacuation paths that avoid the obstacle points and meet the evacuation direction requirements.

4. The method of claim 3, wherein, Building the candidate set of replacement resources further includes: Obtaining the real-time running load and performance limitation state of each device in the set of candidate replacement devices that are spatially reachable from the unified running state view; Calculating the available redundant capacity of each candidate device in combination with the device design parameters, and comparing it with the capacity gap of the original action in the set of actions to be replaced; Eliminating devices with insufficient available redundant capacity or in a fault state, and selecting device combinations that can make up the capacity gap through single-machine or multi-machine combination, to form a real-time verified candidate set of replacement resources.

5. The method of claim 4, wherein, Reconstructing an optimal replacement action sequence that meets the current mode constraints includes generating candidate replacement action sequences: For each action in the set of actions to be replaced, selecting a replacement device combination or a replacement evacuation path from the real-time verified candidate set of replacement resources, and instantiating a local replacement sub-sequence containing a pre-action, a core replacement action, and a post-recovery action according to the action template in the pre-built multi-system mode and preplan joint model; Inserting the local replacement sub-sequence into the position of the original action to be replaced in the target preplan action sequence to form a number of candidate replacement action sequences to be verified.

6. The method of claim 5, wherein, Reconstructing an optimal replacement action sequence that meets the current mode constraints also includes timing reconstruction of the candidate replacement action sequences: According to the action dependency rules and mutual exclusion constraints in the pre-built multi-system mode and preplan joint model, abstracting all actions in the candidate replacement action sequences as nodes, and abstracting the sequential dependency and mutual exclusion relationship between actions as directed edges, to build an action dependency graph; Performing a topological sorting algorithm on the action dependency graph to detect whether there is a dependency loop, eliminating sequences with loops, and determining the sequences that pass the sorting as candidate replacement action sequences that are time-consistent.

7. The method of claim 6, wherein, Reconstructing an optimal replacement action sequence that meets the current mode constraints includes evaluation and optimization: According to the event type in the emergency event trigger information, a corresponding index weight vector is loaded, and the index weight vector defines the weight proportion of safety, evacuation efficiency and mode consistency in the current scene; In combination with the unified running state view, each time sequence consistent candidate alternative action sequence is multi-dimensionally evaluated, safety score, evacuation efficiency score and mode consistency score are calculated, and a comprehensive score is calculated by using the index weight vector for weighted calculation; The sequence with the highest comprehensive score is selected as the optimal alternative action sequence.

Images