Safety production supervision and scheduling management method and device and medium

Abstract

The present application relates to the field of safety production informatization, and particularly relates to a safety production supervision and scheduling management method, device and medium, the method comprising the following steps: firstly, constructing a full-factor data twin and a cause-and-effect rule library that are real-time synchronized with an actual production system, and decomposing emergency plans into atomic-level instructions; when an accident occurs, triggering all eligible emergency plans at the same time to form a set of instructions to be executed; screening out potential conflict instructions through a two-level conflict detection mechanism, simulating execution of the instructions in the data twin, deducing consequences of the execution and calculating a comprehensive effect value; generating a conflict-free globally optimal scheduling instruction sequence based on the deduced results; and real-time monitoring of the instruction execution state, and dynamically adjusting the scheduling scheme when detecting an execution deviation. The present application can effectively improve the efficiency and reliability of emergency response and reduce accident losses.

Description

Technical Field

[0001] This invention belongs to the field of information technology for safe production, specifically involving methods, devices, and media for safe production supervision, scheduling, and management. Background Technology

[0002] Safety in production is a core management aspect of high-risk industries such as industrial production, chemical manufacturing, and energy extraction. Emergency response and dispatch management for sudden accidents directly impact personnel safety and the extent of property damage. With the development of information technology, safety production monitoring and dispatch systems have evolved from traditional manual operation to digitalization and intelligence. Currently, mainstream safety production monitoring and dispatch systems primarily employ a plan-based emergency response mechanism. Its core technical principle is: pre-prepared independent emergency plans for various possible accident types; when the system detects an accident of a certain type, it automatically triggers the corresponding emergency plan and issues dispatch instructions according to the order specified in the plan. For situations where multiple plans are triggered simultaneously, existing technologies generally adopt a priority-based dispatch strategy, executing instructions from highest to lowest priority. However, existing technologies have the following fundamental technical flaws: priority-based dispatch strategies cannot resolve inherent instruction conflicts. When the instructions of two or more plans conflict causally, even execution according to priority order may lead to serious consequences. The reason for this deficiency lies in the fact that existing technologies lack a causal relationship model between different contingency plans and between contingency plan instructions and the state of the production system, thus failing to understand the impact of instruction execution. There is a lack of ability to extrapolate the consequences of instruction execution based on real-time system status. Existing emergency plans are static processes, unable to predict the chain reactions of instruction execution based on dynamically changing system states. This deficiency also stems from the lack of a data twin synchronized in real-time with the actual production system. Manual conflict detection methods are inefficient and limited. Existing conflict detection relies on pre-labeled rules, failing to comprehensively cover all possible conflict scenarios or detect dynamic conflicts. Furthermore, it cannot achieve globally optimal scheduling of multiple contingency plan objectives. Existing technologies can only sacrifice lower-priority contingency plan objectives to ensure higher-priority objectives, failing to maximize the overall rescue effect. Therefore, a new safety production supervision and scheduling management method is urgently needed, capable of causally extrapolating the consequences of scheduling instruction execution based on real-time system status, automatically detecting and resolving instruction conflicts, and achieving globally optimal scheduling of multiple contingency plan objectives. Summary of the Invention

[0003] The purpose of this invention is to provide a method for safety production supervision, scheduling and management to solve the problems mentioned in the background art.

[0004] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0005] Safety production supervision, scheduling, and management methods include the following steps:

[0006] S1. Construct a data twin of all elements of safe production, establish a rule base for causal relationships between various elements of the production system and scheduling instructions and system status, and calibrate and update the status of the data twin in real time;

[0007] S2. Transform the emergency plan into a structured form, decompose it into atomic-level instructions and label them with attributes, and establish a mapping between instructions and causal relationship rules;

[0008] S3. Collect and preprocess production data in real time, identify accident characteristics, comprehensively analyze accident information, trigger all eligible emergency plans, and form a set of instructions to be executed;

[0009] S4. Perform conflict pre-detection on the instructions to be executed, filter out the set of potentially conflicting instructions, and mark the instructions without conflicts.

[0010] S5. Generate a snapshot of the data twin's state, simulate the execution of potentially conflicting instructions, deduce the execution consequences of single instructions and different combinations of instructions, and calculate the comprehensive effect value.

[0011] S6. Filter feasible instruction sequences, select the one with the highest comprehensive effect value as the optimal instruction sequence, integrate and optimize it with conflict-free instructions to form the final scheduling instruction sequence;

[0012] S7. Issue instructions, monitor execution status in real time, update data twin status, and detect execution deviations;

[0013] S8. When an execution deviation is detected, analyze the cause and regenerate the scheduling plan to achieve dynamic adjustment.

[0014] Furthermore, in step S1, the causal relationship rule base is established based on the operating mechanism of the production system, physical and chemical laws, industry standards and expert experience, and each rule is represented in the form of "condition-result".

[0015] Furthermore, in step S2, the atomic-level instruction refers to an operation with a clear and single objective that can only be completed by one executing entity within a continuous time period; its attributes include instruction ID, associated plan ID, executing entity, execution location, operation object, operation type, operation parameters, preconditions, execution duration, affected object, and contribution to the plan objective.

[0016] Furthermore, in step S4, the conflict pre-detection includes: initial conflict screening based on the operation object, marking instruction pairs with opposite operation types; conflict detection based on preconditions, marking instruction pairs whose preconditions violate each other; and conflict detection based on resource contention, marking instruction pairs vying for the same limited resource.

[0017] Furthermore, in step S5, the comprehensive effect value is a weighted sum of the weight of each contingency plan target and the degree of completion of that contingency plan target, wherein the weight of the contingency plan target is preset according to the accident level and the importance of the contingency plan.

[0018] Furthermore, in step S6, if the combined effect values ​​of multiple feasible instruction sequences are the same, the sequence with the shortest execution time is selected; instruction sequence optimization includes merging instructions that can be executed in parallel, adjusting execution time, and deleting redundant instructions.

[0019] Furthermore, in step S7, the executing entity reports the execution status at a frequency of once per minute; execution deviation refers to the deviation between the actual system status and the predicted status exceeding a preset threshold.

[0020] This application also discloses a safety production supervision and dispatch management device, including: a data twin construction module for constructing a full-element data twin and a causal relationship rule base; a contingency plan processing module for structured conversion of emergency plans and atomized decomposition of instructions; an accident perception module for real-time data collection, accident identification, and triggering of emergency plans; a conflict detection module for pre-detection of instruction conflicts; a causal deduction module for simulating instruction execution and deducing execution consequences; an instruction generation module for generating a globally optimal dispatch instruction sequence; an execution monitoring module for issuing instructions and monitoring execution status; and a dynamic adjustment module for handling execution deviations and regenerating dispatch schemes.

[0021] This application also discloses an electronic device, including:

[0022] At least one processor; and

[0023] A memory communicatively connected to the at least one processor; wherein,

[0024] The memory stores instructions that can be executed by the at least one processor, which are executed by the at least one processor to enable the at least one processor to execute the safety production supervision and scheduling management method of the present invention.

[0025] This application also discloses a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-described safety production supervision and scheduling management method of the present invention.

[0026] Beneficial Effects: This invention, by constructing a full-element data twin and a causal relationship rule base, can accurately predict the chain effects after command execution, identify and resolve essential causal conflicts, and avoid the escalation of accidents caused by command conflicts. This invention achieves dynamic scheduling decisions based on real-time system status. All causal deductions and scheduling decisions are made based on the current actual system status, adapting to the dynamic changes in system status during accident handling and making more reasonable and effective scheduling decisions. This invention comprehensively improves the efficiency and accuracy of command conflict detection. A two-level detection mechanism is adopted: pre-detection quickly filters obvious conflicts, and causal deduction deeply detects potential dynamic conflicts, ensuring both efficiency and comprehensiveness. This invention achieves globally optimal scheduling for multiple contingency plans. A multi-objective optimized scheduling model is established, aiming to maximize the overall rescue effect, maximizing the overall rescue effect while ensuring the achievement of all contingency plan objectives as much as possible. This invention improves the automation level and reliability of emergency response. It achieves fully automated processing from accident detection to command issuance, reducing human intervention and avoiding the subjectivity and arbitrariness of human decision-making. This invention possesses continuous learning and optimization capabilities. The post-event review mechanism enables continuous updates to the causal relationship rule base and emergency response plans, thereby continuously improving system performance. Attached Figure Description

[0027] Figure 1 This is a flowchart of the safety production supervision and scheduling management method of the present invention;

[0028] Figure 2 This is a flowchart of the steps for constructing a data twin of all elements of safe production in an embodiment of the present invention;

[0029] Figure 3 This is a flowchart of the steps for digitizing emergency plans and atomizing instructions in an embodiment of the present invention;

[0030] Figure 4 This is a flowchart of the real-time accident perception and multiple contingency plan triggering steps in an embodiment of the present invention;

[0031] Figure 5 This is a flowchart of the instruction conflict pre-detection step in an embodiment of the present invention;

[0032] Figure 6 This is a flowchart of the steps for causal relationship deduction and conflict-free optimal instruction sequence generation based on data twins in an embodiment of the present invention;

[0033] Figure 7 is a schematic diagram of the attribute information of the set of instructions to be executed in an embodiment of the present invention, which is used to show the core attribute parameters of each atomic-level scheduling instruction participating in conflict pre-detection;

[0034] Figure 8 is a schematic diagram of the physical entity management interface of the safety production full-element data twin management system in an embodiment of the present invention, which is used to realize the digital modeling and information maintenance of various physical entities in the production system;

[0035] Figure 9 is a schematic diagram of the causal rule management interface of the safety production full-element data twin management system in an embodiment of the present invention, which is used for querying, adding, editing and maintaining the causal relationship rule base.

[0036] Figure 10 is a schematic diagram of the system control and real-time monitoring interface of the safety production full-element data twin management system in an embodiment of the present invention, which is used for the status calibration, synchronization control and real-time status monitoring and display of the data twin;

[0037] Figure 11 is a schematic diagram of the emergency plan structured management and atomic instruction decomposition interface in an embodiment of the present invention, which is used for the structured input of emergency plans, the atomic decomposition of the handling process, and the annotation configuration of instruction attributes;

[0038] Figure 12 is a schematic diagram of the real-time monitoring center interface for instruction issuance and execution status in an embodiment of the present invention, which is used for controlling the issuance of scheduling instructions, monitoring the execution progress, and synchronously displaying the status of the production system.

[0039] Figure 13 is a schematic diagram of the real-time status details interface of the data twin in an embodiment of the present invention, which is used to display the detailed status parameters of the core production equipment and mark the execution deviation between the actual status and the predicted status;

[0040] Figure 14 is a schematic diagram of the real-time status and accident triggering interface of the production system in an embodiment of the present invention, which is used to display the real-time operating parameters of the production entity, accident identification alarm information and the list of emergency plans that have been matched and triggered.

[0041] Figure 15 is a schematic diagram of the scheduling instruction generation and conflict inference results interface in an embodiment of the present invention, which is used to display the optimal scheduling instruction sequence, instruction conflict detection results and causal inference impact data of each instruction;

[0042] Figure 16 is a schematic diagram of the execution deviation processing interface in an embodiment of the present invention, used to display the details of the execution deviation event and record the full process log of deviation analysis and dynamic adjustment of the scheduling scheme;

[0043] Figure 17 is a schematic diagram of the digital replay interface of the entire accident process in an embodiment of the present invention, which supports the visual replay of the accident development process and displays the accident parameter change curve, instruction execution record and real-time system status.

[0044] Figure 18 is a schematic diagram of the accident review and causal rule optimization interface in an embodiment of the present invention. It is used to display the verification status of the causal rule base and generate update and optimization suggestions for the rule base based on accident data. Detailed Implementation

[0045] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0046] This invention provides a method for safety production supervision, scheduling, and management, such as... Figure 1 As shown, the steps include:

[0047] S1. Construct a data twin of all elements of safe production, establish a rule base for causal relationships between various elements of the production system and scheduling instructions and system status, and calibrate and update the status of the data twin in real time;

[0048] S2. Transform the emergency plan into a structured form, decompose it into atomic-level instructions and label them with attributes, and establish a mapping between instructions and causal relationship rules;

[0049] S3. Collect and preprocess production data in real time, identify accident characteristics, comprehensively analyze accident information, trigger all eligible emergency plans, and form a set of instructions to be executed;

[0050] S4. Perform conflict pre-detection on the instructions to be executed, filter out the set of potentially conflicting instructions, and mark the instructions without conflicts.

[0051] S5. Generate a snapshot of the data twin's state, simulate the execution of potentially conflicting instructions, deduce the execution consequences of single instructions and different combinations of instructions, and calculate the comprehensive effect value.

[0052] S6. Filter feasible instruction sequences, select the one with the highest comprehensive effect value as the optimal instruction sequence, integrate and optimize it with conflict-free instructions to form the final scheduling instruction sequence;

[0053] S7. Issue instructions, monitor execution status in real time, update data twin status, and detect execution deviations;

[0054] S8. When an execution deviation is detected, analyze the cause and regenerate the scheduling plan to achieve dynamic adjustment.

[0055] Example 1

[0056] This embodiment provides the implementation process of the safety production supervision and dispatch management method (refer to...). Figures 7 to 18 ):

[0057] Step S1: Construction of a data twin of all elements of safe production; This step is performed during the system initialization phase to provide a basic model for subsequent causal inference and scheduling decisions.

[0058] Specifically, such as Figure 2 As shown, this step includes the following sub-steps:

[0059] S110. Production System Physical Entity Modeling. Digital modeling is performed on all physical entities in the production site, including production equipment, pipelines, valves, storage tanks, sensors, actuators, buildings, personnel, vehicles, fire protection facilities, and emergency supplies. Each physical entity is assigned a unique identifier, and its basic attributes such as equipment model, specifications, installation location, design parameters, and operating status are defined.

[0060] S120. System Status Parameter Definition. Define a set of monitorable status parameters for each physical entity, including temperature, pressure, flow rate, liquid level, concentration, rotational speed, current, voltage, switch status, position, speed, etc. Set normal value range, warning value, and danger value for each status parameter.

[0061] S130. Establishment of a causal relationship rule base. Based on the operating mechanism of the production system, physical and chemical laws, industry standards, and expert experience, establish a causal relationship rule base between various elements of the production system and between scheduling instructions and system state. Each rule is represented in the form of "condition-result", clearly defining the time delay and degree of impact of the result.

[0062] S140. Initial State Calibration of Data Twin. By collecting real-time operational data from the production system, the initial state of the data twin is calibrated to ensure that the virtual model state is completely consistent with the actual production system state.

[0063] S150. Establishment of a real-time data twin update mechanism. Establish a real-time data synchronization mechanism between the data twin and the actual production system, setting the data update cycle to 1 second to ensure that the data twin's state is always synchronized with the actual system.

[0064] A full-element data twin refers to a digital virtual model that corresponds one-to-one with the actual production system, capable of reflecting the real-time status and operating parameters of all physical entities within the production system. Optionally, depending on the complexity of the production system, a layered modeling approach can be adopted, first constructing a plant-level overall model, and then refining it into workshop-level and equipment-level sub-models. It is known that real-time synchronization of the data twin is fundamental to ensuring the accuracy of causal inference; if the data update delay exceeds one second, it may lead to a significant deviation between the inference results and the actual situation.

[0065] Step S2: Digitalization of emergency plans and atomic decomposition of instructions; This step is performed during the system initialization phase, transforming traditional text-based emergency plans into structured data that can be processed by computers.

[0066] Specifically, such as Figure 3As shown, this step includes the following sub-steps:

[0067] S210. Structured Conversion of Emergency Response Plans. All emergency response plans of the enterprise are converted into a structured format according to a unified format and stored in the plan database. Each structured plan includes information such as plan number, name, accident type, triggering conditions, response level, plan objectives, responsible department, execution time limit, and handling procedures.

[0068] S220, Atomized Decomposition of Contingency Plan Instructions. The handling process of each contingency plan is decomposed into a series of indivisible atomic-level scheduling instructions. For example, "handling tank leakage" is decomposed into independent instructions such as "closing the feed valve", "opening the emergency shut-off valve", and "setting up a dike".

[0069] S230, Atomic Instruction Attribute Labeling. Each atomic-level instruction is labeled with attributes such as instruction ID, associated plan ID, executing entity, execution location, target object, operation type, operation parameters, preconditions, execution duration, affected objects, and contribution to the plan's target.

[0070] S240, Instruction-Causal Relationship Rule Mapping. A mapping relationship is established between each atomic-level instruction and the relevant rules in the causal relationship rule base, clarifying the causal rules triggered and the resulting system state changes after the instruction is executed.

[0071] Atomic-level instructions refer to operations with a clear and singular objective that can only be completed by a single executor within a continuous time period. Optionally, an instruction template library can be established to standardize the definitions of instructions for common operation types, improving contingency plan conversion efficiency. It is known that atomized instruction decomposition is a prerequisite for instruction conflict detection and causal deduction; if the instruction granularity is too large, potential conflicts cannot be accurately identified.

[0072] Step S3: Real-time accident detection and triggering of multiple contingency plans; This step is continuously executed during system operation and is responsible for detecting the occurrence of accidents and triggering corresponding emergency plans.

[0073] Specifically, such as Figure 4 As shown, this step includes the following sub-steps:

[0074] S310 Real-time data acquisition and preprocessing. Various sensors deployed on the production site collect real-time operational data of the production system according to a preset sampling frequency. The raw data is then cleaned, denoised, outlier-handled, and format-converted.

[0075] S320. Accident Feature Identification and Judgment. Preprocessed real-time data is matched against a preset accident feature model. When the real-time data matches the feature model of a certain type of accident, that type of accident is determined to have occurred.

[0076] S330. Comprehensive Analysis of Accident Information. A comprehensive analysis is conducted on identified accidents to determine their type, level, time of occurrence, location, scope of impact, and severity, and to predict potential secondary accidents.

[0077] S340. Multi-plan matching and triggering. Based on the type, level, and scope of impact of the accident, all emergency plans that meet the triggering conditions are matched from the plan database, and all eligible plans are triggered simultaneously. All atomic-level instructions are collected to form a set of instructions to be executed.

[0078] The accident characteristic model is built upon historical accident data, industry standards, and expert experience, encompassing typical data characteristics and development patterns of various types of accidents. Optionally, multi-sensor data fusion can be introduced to improve the accuracy and reliability of accident identification. It is understood that simultaneously triggering all eligible contingency plans, rather than just the highest-priority plan, is the foundation for achieving globally optimal scheduling in this invention.

[0079] Step S4: Command conflict pre-detection; This step is executed immediately after multiple pre-plans are triggered to quickly filter out command pairs with obvious conflicts.

[0080] Specifically, such as Figure 5 As shown, this step includes the following sub-steps:

[0081] S410. Initial conflict screening based on operands. Traverse the set of instructions to be executed, group the instructions according to operands, and check whether there are instructions with opposite operation types in the instruction group of the same operand. Mark them as first-type conflict instruction pairs.

[0082] S420. Conflict detection based on preconditions. Check whether the precondition of each instruction can be violated by the execution of other instructions. If the execution of one instruction would cause the precondition of another instruction to be unmet, it is marked as a type II conflict instruction pair.

[0083] S430, Conflict Detection Based on Resource Contention. Check if multiple instructions need to use the same limited resource within the same time period. If the resource quantity is insufficient to simultaneously meet the needs of all instructions, mark it as a third-type conflicting instruction pair.

[0084] S440. Generation of Potential Conflicting Instruction Set. All marked conflicting instruction pairs are aggregated to generate a potential conflicting instruction set. Instructions without any conflicts are marked as conflict-free instructions.

[0085] Command conflicts can be categorized into three types: direct operational conflicts, precondition conflicts, and resource contention conflicts. Optionally, a conflict prediction model can be built based on historical conflict data to identify high-risk command combinations in advance. It is known that the pre-detection phase can quickly eliminate conflict-free commands, significantly reducing the computational load of subsequent causal deduction.

[0086] Step S5: Causal relationship deduction based on data twins; This step is the core step of the present invention, which simulates and deduces the execution consequences of potentially conflicting instructions.

[0087] Specifically, such as Figure 6 As shown, this step includes the following sub-steps:

[0088] S510. Data twin state snapshot generation. A snapshot of the data twin's state is taken at the current moment, saving the current state parameter values ​​of all physical entities as the initial state for causal inference.

[0089] S520, Single Instruction Execution Consequence Simulation. For each instruction in the potentially conflicting instruction set, simulate execution based on the snapshot state of the data twin, trigger relevant causal rules, track chain reactions, record the system state and the completion status of contingency plan objectives after the simulation, and calculate the contribution and damage of the instruction to each contingency plan objective.

[0090] S530, Instruction Combination Execution Consequence Deduction. For all possible instruction combinations in the potentially conflicting instruction set, generate all possible execution sequences, simulate the execution of each sequence of instruction combinations in turn, record the system state and the completion status of the contingency plan objectives for each execution sequence, and calculate the comprehensive effect value.

[0091] S540. Analysis of Simulation Results. Organize the simulation results for all single instructions and instruction combinations, generating a simulation result report that includes execution consequences, impact on contingency plan objectives, and a comprehensive effect value. Causal simulation refers to simulating the instruction execution process in a data twin to predict its cascading impact on the system state and all contingency plan objectives. Optionally, parallel computing technology can be used to simultaneously simulate multiple instruction combinations, improving simulation efficiency. It should be noted that the comprehensive effect value is a weighted sum of the weight of each contingency plan objective and the degree of its completion, where the weights of the contingency plan objectives are pre-set based on the incident level and the importance of the contingency plan.

[0092] Step S6: Generation of conflict-free optimal instruction sequence; This step generates the globally optimal scheduling instruction sequence based on the causal deduction results.

[0093] Specifically, such as Figure 6 As shown, this step includes the following sub-steps:

[0094] S610. Selection of Feasible Instruction Sequences. Based on the simulation results, select all conflict-free feasible instruction sequences.

[0095] S620. Optimal Instruction Sequence Selection. Select the sequence with the highest overall performance value from all feasible instruction sequences as the optimal instruction sequence. If the overall performance values ​​are the same, select the sequence with the shortest execution time.

[0096] S630, Instruction Sequence Integration. Conflict-free instructions are integrated with the optimal instruction sequence to form the final complete scheduling instruction sequence, ensuring the sequence's executability.

[0097] S640, Instruction Sequence Optimization. The final instruction sequence is optimized by merging instructions that can be executed in parallel, adjusting instruction execution time to improve resource utilization, and removing redundant instructions.

[0098] Among them, a feasible instruction sequence refers to an instruction sequence in which there is no causal conflict between any two instructions. Optionally, intelligent optimization algorithms such as genetic algorithms and simulated annealing algorithms can be introduced to quickly find an approximate optimal solution when there are a large number of instruction combinations. It is known that globally optimal scheduling can maximize the overall rescue effect while ensuring that all planned objectives are achieved as much as possible.

[0099] Step S7: Instruction Issuance and Real-time Monitoring of Execution Status; This step is responsible for issuing the optimal instruction sequence to the execution entity and monitoring the execution status. The execution status includes four categories: not executed, executing, completed, and execution failed.

[0100] Specifically, this step includes the following sub-steps:

[0101] S710, Instructions are issued in batches. Instructions are issued to each execution entity in batches according to the final scheduling instruction sequence, and instructions that can be executed in parallel are issued simultaneously.

[0102] S720. Real-time collection of execution status. Establish an execution status feedback mechanism, requiring the executing entity to report the execution status once per minute. For instructions that fail to execute, the reason for failure and the current system status must be reported.

[0103] S730, Data Twin Status Synchronization Update. The status of the data twin is updated in real time based on execution status feedback and real-time data collected by sensors.

[0104] S740, Execution Deviation Detection. The actual system state is compared with the system state predicted by causal inference. If the deviation exceeds a preset threshold, it is determined to be an execution deviation.

[0105] Optionally, a visual monitoring interface can be established to display the real-time progress of command execution and changes in system status. It is clear that real-time monitoring can promptly identify problems during execution, providing a basis for dynamically adjusting the scheduling plan.

[0106] Step S8: Execution Deviation Handling and Dynamic Adjustment of Scheduling Scheme; This step is executed when an execution deviation is detected, and the scheduling scheme is dynamically adjusted according to the actual situation. Execution deviations include situations such as instruction execution failure, unexpected changes in system state, and the occurrence of new or secondary accidents. Optionally, a deviation grading mechanism can be established to adopt different handling strategies based on the severity of the deviation.

[0107] Specifically, this step includes the following sub-steps:

[0108] S810. Deviation Cause Analysis. When an execution deviation is detected, immediately analyze the cause of the deviation.

[0109] S820. Regenerate the scheduling scheme. Based on the current actual system status and the cause of the deviation, return to step S3 to re-perform accident analysis, contingency plan triggering, instruction conflict detection, causal deduction, and optimal instruction sequence generation.

[0110] S830. Switching between old and new scheduling schemes. Suspend the execution of instructions that have not yet been issued under the original scheduling scheme, issue the new scheduling scheme to all implementing entities, and notify relevant personnel of the scheme change.

[0111] S840, Adjustment Process Record. This record details the process of adjusting the scheduling plan, including the time of deviation occurrence, the content of the deviation, the cause of the deviation, the process of generating the new plan, and the switchover time. It can be seen that the dynamic adjustment mechanism can adapt to the dynamic changes in the development of an accident, ensuring that the scheduling plan always remains optimal.

[0112] This embodiment achieves accurate prediction of the consequences of dispatching instructions by constructing a full-element data twin and a causal relationship rule base, fundamentally solving the problem of instruction conflicts when multiple plans are triggered simultaneously; it adopts a two-level conflict detection mechanism, which ensures both detection efficiency and detection comprehensiveness; and it generates a globally optimal dispatching instruction sequence based on multi-objective optimization, maximizing the overall rescue effect.

[0113] Example 2

[0114] This embodiment uses a fire and hazardous chemical leak accident that occurred in a storage tank area of ​​a chemical enterprise as an example to illustrate the specific application of the method of the present invention.

[0115] Steps S1-S2

[0116] At 10:00 AM on one of the natural days, during the system initialization phase, the construction of a full-element data twin of the chemical plant's tank area and the digitization and atomization decomposition of emergency plans and instructions were completed. The tank area comprises 10 vertical tanks (T1-T10), with T1-T5 storing gasoline and T6-T10 storing diesel. Each tank is equipped with an inlet valve, outlet valve, emergency shut-off valve, level sensor, temperature sensor, and pressure sensor. A rule base containing thousands of causal relationship rules was established. The structured conversion and atomization decomposition of instructions for the "Fire Emergency Plan" (YJ001) and the "Hazardous Chemical Leakage Emergency Plan" (YJ002) were completed. Specifically, the "Fire Emergency Response Plan" is broken down into atomic-level instructions such as I1 (start fire pump P1), I2 (open fire hydrants around storage tank T1), I3 (close the emergency shut-off valve of storage tank T1), I4 (organize personnel evacuation), I5 (notify the fire department), and I6 (cut off non-fire-fighting power supply); the "Hazardous Chemical Leakage Emergency Response Plan" is broken down into atomic-level instructions such as J1 (close the feed valve of storage tank T1), J2 (open the emergency shut-off valve of storage tank T1), J3 (open the explosion-proof ventilation system), J4 (organize personnel to wear gas masks), J5 (notify the environmental protection department), and J6 (set up a dike). For example, the attribute of instruction I3 is: Instruction The relevant contingency plan Executing entity = Production department, Execution location = Storage tank T1 site, Operation object = Emergency shut-off valve of storage tank T1, Operation type = Closure, Execution duration = 2 minutes, Contribution to plan target The attributes of instruction J2 are: instruction The relevant contingency plan Executing entity = Production department; Execution location = Storage tank T1 site; Operation object = Emergency shut-off valve of storage tank T1; Operation type = Open; Execution duration = 2 minutes; Contribution to plan target. .

[0117] Step S3: The real-time accident perception and multi-plan triggering system detects a sudden rise in the temperature of storage tank T1 to 70°C via temperature sensors around T1. Simultaneously, combustible gas sensors detect that the concentration of combustible gas around T1 has reached 30% of the lower explosive limit, and the video monitoring system detects an open flame in the T1 area. The system comprehensively analyzes the collected data and determines that a fire accident has occurred in storage tank T1, accompanied by a gasoline leak, with an accident level of Level II and an impact area of ​​50 meters around storage tank T1. Based on the accident type and level, the system simultaneously triggers the "Fire Emergency Plan" (YJ001) and the "Hazardous Chemical Leakage Emergency Plan" (YJ002), forming a set of instructions to be executed: {I1,I2,I3,I4,I5,I6,J1,J2,J3,J4,J5,J6}.

[0118] Step S4: Instruction conflict pre-detection;

[0119] S410. Initial conflict screening based on the operation object. It was found that the operation objects of instructions I3 and J2 are both the emergency shut-off valve of storage tank T1, and the operation types are opposite (I3 requires closure, J2 requires opening), and are marked as a first-type conflict instruction pair.

[0120] S420, Precondition-based conflict detection. All instruction preconditions are checked; no instruction preconditions are found to be violated by the execution of other instructions, and there are no Class II conflict instruction pairs.

[0121] S430, Conflict Detection Based on Resource Contention. It was discovered that both instructions I4 and J4 require execution by security personnel, but only one team was present on-site at the time, classifying them as a Type III conflict instruction pair.

[0122] S440. Generation of potential conflicting instruction set. The potential conflicting instruction set is {I3,J2,I4,J4}, and the conflict-free instructions are {I1,I2,I5,I6,J1,J3,J5,J6}.

[0123] Step S5: Causal relationship deduction based on data twins;

[0124] S510. Data twin state snapshot generation. Generate a data twin state snapshot; the key state parameter is: T1 temperature. ℃, T1 liquid level T1 pressure Emergency shut-off valve status = open, fire hydrant network pressure Combustible gas concentration %LEL.

[0125] S520, deduction of the consequences of single instruction execution.

[0126] Simulation command I3 (close emergency shut-off valve): After execution, the feed and discharge of T1 are cut off, but the leakage continues and the concentration of combustible gas continues to rise. The simulation results show that the probability of an explosion occurring at around 10:15 is 85%, the contribution to the fire plan target is 0.3, and the damage to the leakage plan target is 0.7.

[0127] The simulation command J2 (open the emergency shut-off valve): After execution, the gasoline in T1 can be discharged to the standby storage tank through the discharge pipe, and at the same time, fire water is sprayed to cool it down. The simulation results show that the fire was brought under control at about 10:20 and the leak was effectively dealt with. The contribution to the fire plan target is 0.2 and the contribution to the leak plan target is 0.4.

[0128] Instruction I4 (Organize personnel evacuation): After execution, personnel will be evacuated to a safe area. The evacuation process will take 10 minutes. The contribution to the fire emergency plan objective is 0.2, and the contribution to the leakage emergency plan objective is 0.1.

[0129] Instruction J4 (Organize personnel to wear gas masks): After execution, personnel may continue to carry out leak handling work. The contribution to the fire emergency plan objective is 0, and the contribution to the leak emergency plan objective is 0.3.

[0130] S530, deduction of the consequences of instruction combination execution.

[0131] Combination 1: Execute I3, do not execute J2; execute I4, do not execute J4. Overall effect value. .

[0132] Combination 2: Execute J2, do not execute I3; execute J4, do not execute I4. Overall effect value. .

[0133] Combination 3: Execute J2, not I3; execute I4 first, then J4. Overall effect value. .

[0134] S540, Summary of deduction results. Combination 2 has the highest overall effect value, at 0.84.

[0135] Step S6: Generate the conflict-free optimal instruction sequence;

[0136] S610, Feasible instruction sequence filtering. Combination 2 is a feasible instruction sequence, and there is no causal conflict.

[0137] S620, Optimal Instruction Sequence Selection. Combination 2 is selected as the optimal instruction sequence.

[0138] S630, Instruction Sequence Integration. This integrates conflict-free instructions with the optimal instruction sequence to form the final scheduling instruction sequence.

[0139] S640, Instruction Sequence Optimization. The execution order of instructions has been adjusted to allow notifications to the fire department and the environmental protection department to be executed in parallel.

[0140] Step S7: The command issuance and execution status real-time monitoring system issues commands in batches according to the optimized command sequence. Each execution entity executes the command immediately upon receiving it and reports the execution status once per minute. The system updates the data twin status in real time based on the execution status feedback and sensor data.

[0141] Step S8: Deviation Handling and Dynamic Adjustment of Dispatch Scheme. The system detected that the temperature drop rate of T1 was 5°C slower than the predicted result, exceeding the preset threshold of 3°C. Analysis revealed that the deviation was caused by a malfunction in fire pump P2, resulting in insufficient fire water flow. The system immediately re-performed the causal simulation and generated a new dispatch scheme. The new scheme involved activating the backup fire pump P3 and opening more fire hydrants to enhance cooling. The system suspended the execution of any unissued instructions from the original scheme and issued the new instructions to the fire department. The temperature drop rate of T1 returned to the expected level.

[0142] This embodiment successfully resolves the fundamental command conflict between "closing the emergency shut-off valve" and "opening the emergency shut-off valve" in chemical tank area fires and hazardous chemical leaks, thus preventing explosions and minimizing accident losses.

[0143] Example 3

[0144] This embodiment provides an apparatus for implementing the above-mentioned safety production supervision and scheduling management method.

[0145] The device includes:

[0146] The data twin construction module is used to build a data twin of all elements of safe production that is synchronized in real time with the actual production system, and to establish a causal relationship rule base.

[0147] The contingency plan processing module is used to perform structured transformation of emergency plans and atomic decomposition of instructions, annotate the attributes of each atomic-level instruction and establish a mapping with causal relationship rules.

[0148] The accident detection module is used to collect real-time production system operation data, identify accident characteristics, comprehensively analyze accident information, and trigger corresponding emergency plans.

[0149] The conflict detection module is used to pre-detect the instructions to be executed and filter out a set of potentially conflicting instructions.

[0150] The causal deduction module is used to simulate the execution of potentially conflicting instructions in a data twin, deduce their execution consequences, and calculate the overall effect value.

[0151] The instruction generation module is used to generate a conflict-free, globally optimal scheduling instruction sequence based on causal deduction results.

[0152] The execution monitoring module is used to issue scheduling instructions, monitor the execution status of instructions in real time, update the status of the data twin, and detect execution deviations.

[0153] The dynamic adjustment module is used to analyze the cause of the deviation and regenerate the scheduling plan when an execution deviation is detected.

[0154] The causal deduction module further includes:

[0155] The State Snapshot Unit is used to generate a snapshot of the current state of the data twin.

[0156] The single instruction deduction unit is used to deduce the execution consequences of a single instruction.

[0157] The combined deduction unit is used to deduce the execution consequences of different instruction combinations and execution sequences.

[0158] The results processing unit is used to process the simulation results and generate a simulation report.

[0159] Example 4

[0160] This embodiment provides an electronic device for implementing the above-described method for safety production supervision, scheduling, and management.

[0161] The electronic device includes:

[0162] At least one processor;

[0163] Memory that is communicatively connected to at least one processor;

[0164] bus;

[0165] Input / output interfaces;

[0166] Communication unit.

[0167] The memory stores a computer program that can be executed by at least one processor. When the computer program is executed by at least one processor, it causes the at least one processor to execute the safety production supervision and scheduling management method described in Embodiment 1 or Embodiment 2.

[0168] The processor and memory are connected via a bus, the input / output interface is used to connect input and output devices, and the communication unit is used for data communication with external systems.

[0169] This embodiment also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the safety production supervision and scheduling management method described in Embodiment 1 or Embodiment 2.

[0170] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for safety production supervision, scheduling, and management, characterized in that: Includes the following steps: S1. Construct a data twin of all elements of safe production, establish a rule base for causal relationships between various elements of the production system and scheduling instructions and system status, and calibrate and update the status of the data twin in real time; S2. Transform the emergency plan into a structured form, decompose it into atomic-level instructions and label them with attributes, and establish a mapping between instructions and causal relationship rules; S3. Collect and preprocess production data in real time, identify accident characteristics, comprehensively analyze accident information, trigger all eligible emergency plans, and form a set of instructions to be executed; S4. Perform conflict pre-detection on the instructions to be executed, filter out the set of potentially conflicting instructions, and mark the instructions without conflicts. S5. Generate a snapshot of the data twin's state, simulate the execution of potentially conflicting instructions, deduce the execution consequences of single instructions and different combinations of instructions, and calculate the comprehensive effect value. S6. Filter feasible instruction sequences, select the one with the highest comprehensive effect value as the optimal instruction sequence, integrate and optimize it with conflict-free instructions to form the final scheduling instruction sequence; S7. Issue instructions, monitor execution status in real time, update data twin status, and detect execution deviations; S8. When an execution deviation is detected, analyze the cause and regenerate the scheduling plan to achieve dynamic adjustment.

2. The safety production supervision and dispatch management method according to claim 1, characterized in that, In step S1, the causal relationship rule base is established based on the operating mechanism of the production system, physical and chemical laws, industry standards and expert experience, and each rule is represented in the form of "condition-result".

3. The safety production supervision and dispatch management method according to claim 1, characterized in that, In step S2, the atomic-level instruction refers to an operation with a clear and single objective that can only be completed by one execution subject within a continuous time period; its attributes include instruction ID, associated plan ID, execution subject, execution location, operation object, operation type, operation parameters, preconditions, execution duration, affected objects, and contribution to the plan objective.

4. The safety production supervision and dispatch management method according to claim 1, characterized in that, In step S4, the conflict pre-detection includes: initial conflict screening based on the operation object, marking instruction pairs with opposite operation types; conflict detection based on preconditions, marking instruction pairs whose preconditions violate each other; and conflict detection based on resource contention, marking instruction pairs that compete for the same limited resource.

5. The safety production supervision and dispatch management method according to claim 1, characterized in that, In step S5, the comprehensive effect value is a weighted sum of the weight of each contingency plan target and the degree of completion of that target, wherein the weight of the contingency plan target is preset according to the accident level and the importance of the contingency plan.

6. The safety production supervision and dispatch management method according to claim 1, characterized in that, In step S6, if the combined effect values ​​of multiple feasible instruction sequences are the same, the sequence with the shortest execution time is selected; instruction sequence optimization includes merging instructions that can be executed in parallel, adjusting execution time, and deleting redundant instructions.

7. The safety production supervision and dispatch management method according to claim 1, characterized in that, In step S7, the executing entity reports the execution status once per minute; execution deviation refers to the deviation between the actual system status and the predicted status exceeding a preset threshold.

8. An apparatus utilizing the safety production supervision and scheduling management method according to claim 1, characterized in that, include: The data twin building module is used to construct full-element data twins and a causal relationship rule base; The contingency plan processing module is used for the structured transformation of emergency plans and the atomic decomposition of instructions; The accident perception module is used to collect data in real time, identify accidents, and trigger emergency response plans. The conflict detection module is used for pre-detection of instruction conflicts; the causal inference module is used to simulate the execution of instructions and infer the execution consequences. The instruction generation module is used to generate the globally optimal scheduling instruction sequence; The execution monitoring module is used to issue instructions and monitor the execution status; the dynamic adjustment module is used to handle execution deviations and regenerate the scheduling scheme.

9. An electronic device, characterized in that, It includes at least one processor and a memory communicatively connected to the at least one processor, the memory storing a computer program executable by the at least one processor, the computer program being executed by the at least one processor to cause the at least one processor to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method as described in any one of claims 1-7.

Images