Electronic fence-based hierarchical alarm threshold dynamic adjustment control method and system

Abstract

This invention relates to the field of electronic fence and hierarchical alarm control technology, specifically to a method and system for dynamic adjustment control of hierarchical alarm thresholds based on electronic fences. The method includes: constructing a dynamic risk distribution field; calculating the normalized risk index of the location of the intelligent agent; quantifying and generating the comprehensive protection capability factor of the intelligent agent; quantifying and generating the task criticality factor of the intelligent agent; calculating and generating the individualized risk tolerance threshold of the intelligent agent; comparing the normalized risk index with the individualized risk tolerance threshold, and combining it with a preset safety margin coefficient to determine the current alarm state; calculating the difference between the normalized risk index and the individualized risk tolerance threshold to obtain the risk tolerance deviation; based on the risk tolerance deviation, generating an individualized risk tolerance threshold for the next cycle through closed-loop correction; and executing a hierarchical correction strategy according to the current alarm state. This invention solves the technical contradiction of security systems hindering critical rescue efforts and realizes the cumulative punishment and management of risk exposure.

Description

Technical Field

[0001] This invention relates to the field of electronic fence and hierarchical alarm control technology, specifically to a method and system for dynamic adjustment control of hierarchical alarm thresholds based on electronic fences. Background Technology

[0002] In high-risk work environments such as petrochemical plants, traditional electronic fences rely primarily on rigid area divisions and uniform access standards. This static and indiscriminate risk management logic fails to differentiate the protective capabilities of different agents and does not consider the extreme importance of tasks such as emergency rescue. This creates a technical contradiction where safety strategies hinder the execution of critical tasks, and its binary alarm mechanism easily leads to unnecessary downtime, making it difficult to achieve refined, tiered responses. Therefore, establishing an individualized risk threshold control method that can dynamically assess environmental risks, agent states, and task requirements to achieve synergistic optimization of safety and operational efficiency is an urgent technical problem to be solved. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for dynamically adjusting and controlling graded alarm thresholds based on electronic fences. This method can dynamically calculate individualized risk tolerance thresholds based on the protective capabilities of the intelligent agent and the criticality of the task, solving the technical contradiction that the rigid standardization of traditional electronic fences hinders the execution of critical rescue and other tasks. Furthermore, this method can realize graded alarm judgment and refined graded correction strategies, avoiding unnecessary downtime caused by binary alarms, and achieving synergistic optimization of safety and operational efficiency. Specifically, the technical solution of this invention includes:

[0004] A hierarchical alarm threshold dynamic adjustment control method based on electronic fences includes:

[0005] S1, based on the collected multi-dimensional environmental data, constructs a dynamic risk distribution field;

[0006] S2, based on the dynamic risk distribution field, calculates the normalized risk index of the agent's location;

[0007] S3, based on the collected multi-source state data of the intelligent agent, quantifies and generates the comprehensive protection capability factor of the intelligent agent;

[0008] S4, based on the collected task scheduling instructions, quantifies and generates the key factors of the intelligent agent's tasks;

[0009] S5 calculates the individualized risk tolerance threshold of the generated agent based on the preset basic risk threshold, combined with the comprehensive protection capability factor and the mission criticality factor.

[0010] S6. Compare the normalized risk index with the individualized risk tolerance threshold, and combine them with the preset safety margin coefficient to determine the current alarm status.

[0011] S7, calculate the difference between the normalized risk index and the individualized risk tolerance threshold to obtain the risk tolerance deviation;

[0012] S8, based on the risk tolerance deviation, uses closed-loop correction to generate an individualized risk tolerance threshold for the next cycle;

[0013] S9 executes a graded correction strategy based on the current alarm status.

[0014] Preferably, S2 includes:

[0015] Obtain the real-time or predicted gas concentration at the location of the intelligent agent;

[0016] Obtain the preset danger threshold corresponding to the real-time or predicted gas concentration;

[0017] The normalized risk index is obtained by normalizing the real-time or predicted gas concentration with the preset danger threshold.

[0018] Preferably, S3 includes:

[0019] The equipment protection index and the agent's own state index are obtained by analyzing the multi-source state data of the agent;

[0020] Based on the equipment protection index and the agent's own state index, the comprehensive protection capability factor is determined through weighted calculation.

[0021] S4 includes:

[0022] From the task scheduling instructions, the task priority and task consequence index are obtained;

[0023] Based on task priority and task consequence index, the key factors of the task are determined through weighted calculation.

[0024] Preferably, S5 includes:

[0025] Based on the comprehensive protection capability factor and the preset first weight coefficient, a protection capability product term is generated;

[0026] Based on the task criticality factor and the preset second weight coefficient, a task criticality product term is generated;

[0027] Multiply the product of the basic risk threshold and the protection capability to obtain the intermediate product;

[0028] The intermediate product is multiplied by the task-critical product term to determine the individualized risk tolerance threshold.

[0029] Preferably, S6 includes:

[0030] If the normalized risk index is not greater than the individualized risk tolerance threshold, the current alarm status is determined to be a safe status.

[0031] If the normalized risk index is greater than the individualized risk tolerance threshold, but not greater than the product of the individualized risk tolerance threshold and the safety margin coefficient, then the current alarm status is determined to be a level 1 alarm status.

[0032] If the normalized risk index is greater than the product of the individualized risk tolerance threshold and the safety margin coefficient, the current alarm status is determined to be a level 2 alarm status.

[0033] Preferably, S8 includes:

[0034] In response to a risk tolerance deviation greater than zero, a threshold tightening factor is calculated and generated based on the risk tolerance deviation and a preset correction coefficient.

[0035] The individualized risk tolerance threshold and the threshold tightening factor are calculated to determine the individualized risk tolerance threshold for the next cycle.

[0036] If the risk tolerance deviation is not greater than zero, the individualized risk tolerance threshold will be determined as the individualized risk tolerance threshold for the next period.

[0037] Preferably, S9 includes:

[0038] If the current alarm status is a safe state, the agent is allowed to perform the predetermined task.

[0039] If the current alarm status is a Level 1 alarm status, then based on the dynamic risk distribution field, the lowest risk path is planned and pushed.

[0040] If the current alarm status is a level 2 alarm status, the individualized risk tolerance threshold for the next cycle will be forcibly adjusted to the preset minimum value, and the fastest evacuation route will be planned and pushed.

[0041] A hierarchical alarm threshold dynamic adjustment control system based on electronic fence, comprising:

[0042] The risk field construction unit is used to construct a dynamic risk distribution field based on the collected multi-dimensional environmental data;

[0043] The risk index calculation unit is used to calculate the normalized risk index of the agent's location based on the dynamic risk distribution field.

[0044] The protection capability quantification unit is used to quantify and generate the comprehensive protection capability factor of the agent based on the collected multi-source state data of the agent.

[0045] The task quantization unit is used to quantify the key factors of the task in generating the intelligent agent based on the collected task scheduling instructions.

[0046] The individual threshold calculation unit is used to calculate the individualized risk tolerance threshold of the generated agent based on a preset basic risk threshold, combined with a comprehensive protection capability factor and a task criticality factor.

[0047] The alarm status determination unit is used to compare the normalized risk index with the individualized risk tolerance threshold, and combine it with the preset safety margin coefficient to determine and generate the current alarm status.

[0048] The deviation calculation unit is used to calculate the difference between the normalized risk index and the individualized risk tolerance threshold to obtain the risk tolerance deviation.

[0049] The threshold correction unit is used to generate an individualized risk tolerance threshold for the next cycle by closed-loop correction based on the risk tolerance deviation.

[0050] The strategy execution unit is used to execute hierarchical correction strategies based on the current alarm status.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. This invention breaks through rigid regional restrictions and dynamically generates individualized risk tolerance thresholds by quantifying the protective capabilities of intelligent agents and the criticality of tasks, thereby achieving differentiated management;

[0053] 2. By quantifying the criticality of tasks, this invention allows intelligent agents performing important tasks such as emergency rescue to obtain a higher risk tolerance, thus resolving the technical contradiction that security systems hinder critical rescue efforts.

[0054] 3. The alarm status of this invention is divided into safe, level one and level two, realizing hierarchical response. When a level one alarm occurs, the system only replans the lowest risk path without interrupting the task, thus improving the continuity of operation.

[0055] 4. This invention has a closed-loop correction capability. When the risk exposed by the agent exceeds its individual tolerance, the system will automatically tighten the risk threshold for the next cycle, thereby realizing the cumulative punishment and management of risk exposure. Attached Figure Description

[0056] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0057] Figure 1 This is a flowchart of the method of the present invention;

[0058] Figure 2 This is a structural diagram of the system of the present invention. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0060] Example 1:

[0061] Please see Figure 1 A hierarchical alarm threshold dynamic adjustment control method based on electronic fences includes:

[0062] S1, based on the collected multi-dimensional environmental data, constructs a dynamic risk distribution field;

[0063] S2, based on the dynamic risk distribution field, calculates the normalized risk index of the agent's location;

[0064] S3, based on the collected multi-source state data of the intelligent agent, quantifies and generates the comprehensive protection capability factor of the intelligent agent;

[0065] S4, based on the collected task scheduling instructions, quantifies and generates the key factors of the intelligent agent's tasks;

[0066] S5 calculates the individualized risk tolerance threshold of the generated agent based on the preset basic risk threshold, combined with the comprehensive protection capability factor and the mission criticality factor.

[0067] S6. Compare the normalized risk index with the individualized risk tolerance threshold, and combine them with the preset safety margin coefficient to determine the current alarm status.

[0068] S7, calculate the difference between the normalized risk index and the individualized risk tolerance threshold to obtain the risk tolerance deviation;

[0069] S8, based on the risk tolerance deviation, uses closed-loop correction to generate an individualized risk tolerance threshold for the next cycle;

[0070] S9 executes a graded correction strategy based on the current alarm status.

[0071] This embodiment provides a hierarchical alarm threshold dynamic adjustment control method based on electronic fences. This method is a complete and self-consistent technical closed loop, including the following steps:

[0072] S1, based on the collected multi-dimensional environmental data, constructs a dynamic risk distribution field;

[0073] This step involves building a real-time, three-dimensional hazard diffusion model as the data foundation for all subsequent risk assessments;

[0074] Specifically, the system in this embodiment collects multi-dimensional environmental data in a hazardous materials leak scenario at a petrochemical plant. In the specific technical context of this invention, "multi-dimensional environmental data" refers to a heterogeneous set of data used to describe the spatiotemporal evolution characteristics of a hazardous source, serving as input to a diffusion model. In this embodiment, it originates from and includes:

[0075] Real-time hazardous gas concentration obtained through a gas sensor array ;

[0076] Real-time meteorological data obtained from anemometers Such as wind speed and wind direction;

[0077] Strength of potential leakage sources obtained through the distributed control system (DCS) and the status of the factory ventilation system Such as the start and stop of the fan;

[0078] Based on these real-time multimodal input data, combined with computational fluid dynamics (CFD) simulations or Gaussian plume models, the system constructs a dynamic, probabilistic risk distribution field. ;Should The field can accurately simulate the spatiotemporal diffusion characteristics of hazardous areas under the influence of factors such as wind speed and ventilation, and can predict the spatial diffusion characteristics of any point. Concentration in the near future;

[0079] Specifically, when using the Gaussian plume model, the system can base its analysis on the leakage source strength collected by S1. Unit: mass / second, real-time meteorological data Specifically, average wind speed Units: meters per second and preset atmospheric diffusion coefficient. Its value can be determined by looking up a table based on the Pasqual stability level, and can be calculated for any point downwind using the following continuous point source diffusion formula. exist Concentration at time :

[0080] ;

[0081] in The effective height of the leakage source, in meters. Using this formula, the system can construct a dynamic risk distribution field in real time. It responds to S2's call and outputs the location of any agent. Predicted concentration ;

[0082] S2, based on the dynamic risk distribution field, calculates the normalized risk index of the agent's location;

[0083] This step involves converting the complex concentration field constructed in S1 into a unified, dimensionless risk level, providing a standardized input for subsequent access control and solving the problem of inconsistent standards caused by the different toxicity or explosion thresholds of different hazards.

[0084] Specifically, the system is built from S1 Extracting intelligent agents in the field Current location Real-time or predicted gas concentration The system introduces a preset danger threshold. ;

[0085] Preset danger threshold This refers to a concentration baseline value set according to industry standards or the Material Safety Data Sheet (MSDS), whose dimensions are... Same; its function is as the denominator for risk normalization; This is a key adjustable parameter, the source of which is selected according to the emergency response level; for example, during normal operation. Occupational exposure limit (OEL) can be adopted; during emergency response. The immediate life-threatening concentration (IDLH) or lower explosive limit (LEL) can be used. This step calculates the normalized risk index using the following formula. :

[0086] ;

[0087] in, The normalized risk index is a dimensionless output of this step. For real-time or predictive gas concentration, the source is constructed from S1. field; These are preset danger thresholds, derived from industry standards;

[0088] Using this formula, when When this occurs, it indicates that the risk at that point has exceeded the preset legal threshold; The value is the core input for subsequent steps S6 and S7;

[0089] S3, based on the collected multi-source state data of the intelligent agent, quantifies and generates the comprehensive protection capability factor of the intelligent agent;

[0090] This step involves quantitatively evaluating the intelligent agent. Current actual level of protection and health status;

[0091] Specifically, the system collects multi-source state data of the intelligent agent; multi-source state data of the intelligent agent refers to a set of parameters describing the agent's own protection and state; for personnel, the source can be the intelligent badge they wear or a visual recognition system to obtain the PPE level of the protective equipment they wear. For robots, their explosion-proof rating can be obtained from their equipment management system. Self-state indices such as remaining battery power and sensor availability. ;

[0092] The system then quantifies this data to generate a comprehensive protection capability factor. ; It is a dimensionless factor, used as the input to S5;

[0093] S4, based on the collected task scheduling instructions, quantifies and generates the key factors of the intelligent agent's tasks;

[0094] This step involves quantitatively evaluating the intelligent agent. The importance and urgency of the current task, in order to balance the risks involved in carrying out the task;

[0095] Specifically, the system collects task scheduling instructions; task scheduling instructions refer to a set of task parameters obtained from emergency response or production scheduling systems; their sources include those assigned to intelligent agents. Task priority For example, Level 1 - daily inspection vs. Level 10 - shutting off leaking valves, and the task consequence index. That is, the task is effective in reducing or Contribution assessment;

[0096] The system then quantifies these instructions to generate task criticality factors. ; It is a dimensionless factor, used as the input to S5;

[0097] S5 calculates the individualized risk tolerance threshold of the generated agent based on the preset basic risk threshold, combined with the comprehensive protection capability factor and the mission criticality factor.

[0098] This step breaks away from the rigid logic of traditional electronic fences treating all individuals with the same level of danger. Based on the Individualized Risk Tolerance (TIRT) principle proposed in this invention, it assigns a risk level to each agent based on task requirements. Dynamically generate personalized risk passes;

[0099] Specifically, the system sets a basic risk threshold. Basic risk threshold This refers to a basic risk exposure cap shared by all agents, for example... So that it is consistent with S2 Correspondingly, its source is the safety procedure setting, which is dimensionless; the system uses the following formula, combined with S3. and S4 Calculate the individualized risk tolerance threshold :

[0100] ;

[0101] in, The individualized risk tolerance threshold is a dimensionless output of this step, representing the agent's risk tolerance. exist A cycle of individualized risk access; The basic risk threshold is a dimensionless preset parameter. The comprehensive protection capability factor is dimensionless, and its input source is the calculation result of S3; As a critical factor for the task, it is dimensionless, and its input comes from the calculation results of S4; The weighting coefficient is a dimensionless, adjustable parameter; to further clarify, its source can be determined through regression analysis based on safety drill data; for example, a calibration dataset can be collected, which includes data from different drill scenarios. Below, the actual protection level adopted by the intelligent agent. Execution level of tasks and the maximum risk exposure index allowed by safety procedures in this scenario. ;by As the dependent variable, with and Using the least squares method as the independent variable, The formula is fitted to determine the calibration. and The value;

[0102] This formula has protective capabilities. The higher the level, or the more challenging the task. The more important it is, the more important it is. The higher the level, the more it allows it to enter higher-risk areas to complete critical missions, resolving the technical contradiction that hinders rescue efforts due to safety concerns;

[0103] S6. Compare the normalized risk index with the individualized risk tolerance threshold, and combine them with the preset safety margin coefficient to determine the current alarm status.

[0104] This step involves real-time determination of the intelligent agent. The immediate danger level;

[0105] Specifically, the system introduces a safety margin factor. Safety margin coefficient This refers to a preset multiplier used to distinguish between warnings and emergencies, for example... This originates from the definition of the tolerable short-term exceedance range in industry safety standards;

[0106] The system compares intelligent agents in real time. Risk index of location This originates from S2 and its own individualization threshold. Originating from S5, the alarm status is determined according to the following logic: when When, it is determined to be a safe state; when When, it is determined to be a Level 1 alarm state; when When this occurs, it is determined to be a level two alarm status;

[0107] S7, calculate the difference between the normalized risk index and the individualized risk tolerance threshold to obtain the risk tolerance deviation;

[0108] This step involves introducing the concept of ErrorSignal from cybernetics to quantify the degree to which the agent exceeds its limits, thus providing input for the closed-loop correction of S8.

[0109] Specifically, the system calculates the risk tolerance deviation using the following formula. :

[0110] ;

[0111] in: The risk tolerance deviation is a dimensionless output of this step. The input source is S2; The input source is S5;

[0112] this As the error signal for closed-loop correction; when When this occurs, it indicates that the agent is exposed to a risk that exceeds its individual tolerance, triggering the negative feedback correction of S8.

[0113] S8, based on the risk tolerance deviation, uses closed-loop correction to generate an individualized risk tolerance threshold for the next cycle;

[0114] This step involves constructing a negative feedback system as a penalty mechanism for the agent's risk exposure; when the agent is exposed to an environment exceeding the limits, i.e. The system must dynamically tighten its security threshold to prevent the accumulation of risks;

[0115] Specifically, the system is based on S7. Calculate the individualized risk tolerance threshold for the next cycle. ; It will be stored and will replace the one in S5 in the next computing cycle. The value is used, and the calculation formula is as follows:

[0116] ;

[0117] in: The corrected, individualized threshold for the next cycle is the dimensionless output of this step. This is the original threshold for the current period, and its input source is S5; The risk tolerance bias is input from S7; The correction factor is a dimensionless, adjustable sensitivity coefficient, which can be set to, for example, 0.1.

[0118] The technical motivation behind this formula lies in: using This ensures that action is taken only when the risk exceeds the limit. Only when it is positive will the threshold tightening be triggered, that is The term is less than 1; while within the safe zone, that is... When it is negative, The item is 0. The threshold remains unchanged, preventing erroneous threshold increases;

[0119] The longer or deeper the exposure, The larger, The faster the drop, the easier it is for the agent to trigger an alarm in the next cycle, because its S5 threshold input is reduced;

[0120] S9, based on the current alarm status, execute the graded correction strategy;

[0121] This step involves immediately executing the corresponding control action based on the alarm status determined in S6; this step and S8, which calculates the threshold for the next cycle, are executed in parallel.

[0122] In response to a safe state, the agent is allowed to perform its predetermined tasks.

[0123] In response to a Level 1 alarm, the system performs a Level 1 correction; issues a path deviation or risk proximity warning; and activates the lowest-risk path planning algorithm, based on S1. The system calculates a new path to the task point with the lowest risk exposure for the agent and pushes the path to its terminal.

[0124] In response to a Level 2 alarm, the system executes Level 2 correction; issues an immediate Level 2 evacuation alarm; plans the fastest evacuation route to the nearest safe exit; and simultaneously, the system forcibly... That is, the calculation result of S8 is set to 0 or a very small value, so that its access permission will be invalidated instantly in the next cycle;

[0125] The method described in this embodiment upgrades the traditional, rigid area-locking electronic fence into a personalized, precise risk navigation system through a complete closed loop of S1-S9. It no longer simply allows / prohibits entry, but dynamically calculates the individualized risk tolerance S5 based on the agent's protection capabilities S3 and task importance S4; through S5... The formula achieves safety and high efficiency. With efficiency, high The collaborative optimization of S7 and S8 resolved the technical contradictions that hindered critical rescue efforts. Furthermore, through deviation calculation in S7 and negative feedback correction in S8, dynamic tightening of thresholds was achieved, effectively managing the cumulative risk exposure of agents. At the same time, through state grading in S6 and grading strategy in S9, a progressive and precise response was achieved from alarm to path replanning, Level 1, and then to forced evacuation, Level 2, avoiding unnecessary downtime and ensuring safety in extreme situations.

[0126] Example 2:

[0127] S2 includes:

[0128] Obtain the real-time or predicted gas concentration at the location of the intelligent agent;

[0129] Obtain the preset danger threshold corresponding to the real-time or predicted gas concentration;

[0130] The normalized risk index is obtained by normalizing the real-time or predicted gas concentration with the preset danger threshold.

[0131] The further detailed implementation of S2 in the hierarchical alarm threshold dynamic adjustment control method based on electronic fences as described in Example 1 includes:

[0132] The system obtains real-time or predicted gas concentrations at the location of the intelligent agent. As described in S1, this data originates from the dynamic risk distribution field constructed in S1. ;

[0133] The system also obtains preset danger thresholds corresponding to real-time or predicted gas concentrations. As mentioned above, These are benchmark values ​​set according to chemical safety regulations, such as IDLH values.

[0134] The system normalizes the real-time or predicted gas concentration with a preset danger threshold. In this embodiment, this normalization is achieved by performing a division operation, i.e. Thus, the dimensionless normalized risk index is obtained. ;

[0135] This embodiment explicitly defines normalization processing as... A threshold with clear physical meaning and regulatory basis is used for division to ensure... The objectivity and authority of the value; It is no longer an abstract risk score, but a reviewable and verifiable multiple relative to the statutory safety benchmark; this makes the risk assessment results of the present invention highly credible and universally applicable in the industry, and solves the problem of different systems being unable to interoperate due to the lack of transparency in risk algorithms.

[0136] Example 3:

[0137] S3 includes:

[0138] The equipment protection index and the agent's own state index are obtained by analyzing the multi-source state data of the agent;

[0139] Based on the equipment protection index and the agent's own state index, the comprehensive protection capability factor is determined through weighted calculation.

[0140] S4 includes:

[0141] From the task scheduling instructions, the task priority and task consequence index are obtained;

[0142] Based on task priority and task consequence index, the key factors of the task are determined through weighted calculation.

[0143] As described in Example 1, the hierarchical alarm threshold dynamic adjustment control method based on electronic fences provides a specific weighted calculation model for the quantization generation steps of S3 and S4:

[0144] S3 includes:

[0145] The equipment protection index is obtained by analyzing the multi-source state data of the intelligent agent. With the agent's own state index ;

[0146] To ensure dimensional consistency, the system will For example, Grade A protective clothing = 1.0, Grade C = 0.5, and For example, the percentage of electricity can be converted into a dimensionless quantity using a lookup table or a maximum / minimum normalization method. and All are in the range of 0-1;

[0147] based on and The comprehensive protection capability factor is determined through weighted calculation. The weighted calculation formula used in this embodiment is as follows:

[0148] ;

[0149] in: The comprehensive protection capability factor is dimensionless and is output to S5; These are the normalized equipment protection index and the normalized agent's own state index, respectively. The weights are dimensionless and their origin is determined by safety regulations. ,For example ;

[0150] This model is designed to prevent a well-equipped but poorly maintained device from being used. high, The deployment of low-risk intelligent agents to high-risk areas provides a more realistic assessment of protection capabilities.

[0151] S4 includes:

[0152] The task priority is obtained by parsing the task scheduling instructions. With task consequences index ;

[0153] Similarly, the system will For example, levels 1-10, and For example, the contribution assessed is normalized into a dimensionless score. and For example, mapping to the 0-1 range;

[0154] based on and By using weighted calculations, the key factors of the task are determined. ;

[0155] The weighted calculation formula used in this embodiment is:

[0156] ;

[0157] in: This is a task-critical factor, which is dimensionless, and is output to S5; These are the normalized task priority and the normalized task consequence index, respectively; The weights are dimensionless and their origin is determined by the expert scoring method for emergency response plans. ;

[0158] This embodiment provides implementable and specific mathematical models for S3 and S4; through normalization and weighted summation, inputs from multiple sources and units, such as protective clothing level, power consumption, and task priority, are uniformly converted into dimensionless quantities. and Factors ensure that they can be used correctly in the S5 multiplication formula; and improve the comprehensiveness of the evaluation. The calculations also take into account the equipment. and state , The calculation takes priority into account. and consequences This multi-dimensional quantification method makes the evaluation results far more robust and reasonable than those based on a single indicator.

[0159] Example 4:

[0160] S5 includes:

[0161] Based on the comprehensive protection capability factor and the preset first weight coefficient, a protection capability product term is generated;

[0162] Based on the task criticality factor and the preset second weight coefficient, a task criticality product term is generated;

[0163] Multiply the product of the basic risk threshold and the protection capability to obtain the intermediate product;

[0164] The intermediate product is multiplied by the task-critical product term to determine the individualized risk tolerance threshold.

[0165] The method for dynamic adjustment control of hierarchical alarm thresholds based on electronic fences as described in Example 1 is characterized by the following: The calculation in S5... The process provides clear steps for multiplication operations;

[0166] In S5 of this embodiment, the individual threshold calculation unit strictly follows the following logic: it is based on the comprehensive protection capability factor. Originating from S3, and with the preset first weighting coefficient This generates a protection capability product term, which, in this embodiment, is constructed as follows: Based on task criticality factors It originates from S4 and is related to the preset second weighting coefficient. Generate a task-critical product term; in this embodiment, this term is constructed as follows: Then adjust the basic risk threshold. Multiply the product term with the protection capability to obtain an intermediate product; then multiply the intermediate product with the product term of mission criticality to determine the final individualized risk tolerance threshold.

[0167] The above steps correspond completely to mathematically This core formula;

[0168] This embodiment clarifies and It uses multiplication, not addition, with the base threshold. The combination of these elements, along with this multiplicative coupling structure, is a major innovation of this invention.

[0169] Its beneficial effect is as follows: and As the amplification factor To have an impact; to use The structure ensures that even in [the context of]... No protection, and In the baseline case without a task, the agent's threshold remains unchanged. The basic threshold, rather than collapsing to 0 due to multiplication, ensures the stability and robustness of the system mathematically, guaranteeing a basic safety baseline.

[0170] Example 5:

[0171] S6 includes:

[0172] If the normalized risk index is not greater than the individualized risk tolerance threshold, the current alarm status is determined to be a safe status.

[0173] If the normalized risk index is greater than the individualized risk tolerance threshold, but not greater than the product of the individualized risk tolerance threshold and the safety margin coefficient, then the current alarm status is determined to be a level 1 alarm status.

[0174] If the normalized risk index is greater than the product of the individualized risk tolerance threshold and the safety margin coefficient, the current alarm status is determined to be a level 2 alarm status.

[0175] The hierarchical alarm threshold dynamic adjustment control method based on electronic fence as described in Example 1 is characterized in that it provides clear and unambiguous three-level division rules for the alarm status determination logic of S6.

[0176] In S6 of this embodiment, the alarm status determination unit obtains the normalized risk index calculated in S2. Individualized risk tolerance threshold calculated by S5 and the preset safety margin coefficient Then, the following strict comparison logic is executed: if the normalized risk index Not greater than the individualized risk tolerance threshold ,Right now If the normalized risk index is... Greater than the individualized risk tolerance threshold And not greater than the individualized risk tolerance threshold and safety margin coefficient. The product of, i.e. If the normalized risk index is... Greater than the individualized risk tolerance threshold and safety margin coefficient The product of, i.e. If so, the current alarm status is determined to be a level 2 alarm status;

[0177] This embodiment introduces... It defines three clear levels: safety, level one, and level two, with judgment ranges, providing a level of precision far exceeding the traditional binary system of safety / danger and alarms;

[0178] It avoids alarm fatigue; the distinction between level 1 alarms (warnings) and level 2 alarms (emergency) allows the system to execute commensurate and differentiated responses; for example, level 1 alarms only trigger path corrections without interrupting the task, thereby greatly improving the continuity and efficiency of operations while ensuring safety margins.

[0179] Example 6:

[0180] S8 includes:

[0181] In response to a risk tolerance deviation greater than zero, a threshold tightening factor is calculated and generated based on the risk tolerance deviation and a preset correction coefficient.

[0182] The individualized risk tolerance threshold and the threshold tightening factor are calculated to determine the individualized risk tolerance threshold for the next cycle.

[0183] If the risk tolerance deviation is not greater than zero, the individualized risk tolerance threshold will be determined as the individualized risk tolerance threshold for the next period.

[0184] The hierarchical alarm threshold dynamic adjustment control method based on electronic fence described in Example 1 provides clear conditional logic for the closed-loop correction step of S8.

[0185] In S8 of this embodiment, the threshold correction unit obtains the value calculated in S7. Then, execute the following logic: in response to risk tolerance deviation. A value greater than zero indicates that the agent is in an over-exposure state: the system is based on the risk tolerance deviation. With the preset correction coefficient A threshold tightening factor is calculated and generated. In this embodiment, the tightening factor is constructed as follows: and will individualize risk tolerance thresholds. The threshold tightening factor is used in a multiplication operation in this case to determine the individualized risk tolerance threshold for the next period. ,Right now: In response to a risk tolerance deviation not exceeding zero, i.e., the agent is in a safe state, the individualized risk tolerance threshold will be determined. This is determined as the individualized risk tolerance threshold for the next cycle, namely: ;

[0186] In this embodiment, the above two logics are unified by a more robust mathematical formula:

[0187] ;

[0188] The formula is in hour, The item is 0. Automatic equals It implements the full functionality of S8;

[0189] This embodiment clarifies the triggering conditions for negative feedback, only when... Timing and execution method, threshold tightening;

[0190] It ensures the stability and one-way penalty of closed-loop control; the system only tightens the threshold when a dangerous exposure occurs, while keeping the threshold unchanged when it is safe, and will not erroneously raise the threshold because of safety; this design conforms to the basic principle of safety management, that is, to cumulatively penalize risk exposure, ensuring the convergence and stability of the control system and preventing threshold oscillation or runaway.

[0191] Example 7:

[0192] S9 includes:

[0193] If the current alarm status is a safe state, the agent is allowed to perform the predetermined task.

[0194] If the current alarm status is a Level 1 alarm status, then based on the dynamic risk distribution field, the lowest risk path is planned and pushed.

[0195] If the current alarm status is a level 2 alarm status, the individualized risk tolerance threshold for the next cycle will be forcibly adjusted to the preset minimum value, and the fastest evacuation route will be planned and pushed.

[0196] The hierarchical alarm threshold dynamic adjustment control method based on electronic fence provides specific execution actions that correspond one-to-one with the alarm states in Example 5 for the hierarchical correction strategy of S9.

[0197] In S9 of this embodiment, the policy execution unit strictly executes the following policy: In response to the current alarm state being a safe state, the determination is based on: The system allows the agent to perform predetermined tasks without intervention; the current alarm status is a Level 1 alarm status, determined based on: The system triggers a first-level correction; based on the dynamic risk distribution field. Originating from S1, it plans and pushes the lowest-risk path; this is intended to bypass risks, not to stop the task; specifically, the system activates the path planning algorithm and calculates a risk score. Find the lowest possible, yet still attainable, path to the task point, and push that path to the agent terminal in the form of a virtual pipeline;

[0198] The current alarm status is classified as a level 2 alarm, and the determination is based on the following criteria: The system triggers a secondary correction; this action aims to immediately terminate the risk exposure; specifically, the system forcibly adjusts the individualized risk tolerance threshold used for the next cycle. To preset the minimum value, for example The purpose of this action is to immediately revoke all access permissions of the agent at the digital level to prevent it from re-entering in the next cycle. At the same time, the system plans and pushes the fastest evacuation route, which points to the nearest safe exit, and monitors its evacuation throughout the process. The fastest evacuation route can also be based on the above-mentioned gridded graph, but the edge weights are modified to the estimated time required to pass through the grid, that is, the grid path length / the agent's maximum safe speed, and the shortest time path to the nearest safe exit and preset nodes is calculated using Dijkstra's algorithm.

[0199] This embodiment provides specific and significantly technically different execution methods for different alarm levels determined by S6;

[0200] Handling Level 1 Alarm Status: This invention does not employ the one-size-fits-all approach of stopping operations or in-situ evasion commands found in existing technologies, which would lead to task failure. Instead, it innovatively provides a minimum-risk path replanning strategy; this strategy ensures safety while bypassing high-risk obstacles. The optimal balance between area and efficiency, and the continued execution of tasks, is the key execution logic of this invention for achieving synergistic optimization of safety and efficiency.

[0201] Meanwhile, the forced threshold clearing strategy for level 2 alarms provides a stronger and more fundamental security guarantee than simply pushing evacuation routes, ensuring that high-risk intelligent agents cannot re-enter the danger zone.

[0202] Example 8:

[0203] Please see Figure 2 A hierarchical alarm threshold dynamic adjustment control system based on electronic fences, comprising:

[0204] The risk field construction unit is used to construct a dynamic risk distribution field based on the collected multi-dimensional environmental data;

[0205] The risk index calculation unit is used to calculate the normalized risk index of the agent's location based on the dynamic risk distribution field.

[0206] The protection capability quantification unit is used to quantify and generate the comprehensive protection capability factor of the agent based on the collected multi-source state data of the agent.

[0207] The task quantization unit is used to quantify the key factors of the task in generating the intelligent agent based on the collected task scheduling instructions.

[0208] The individual threshold calculation unit is used to calculate the individualized risk tolerance threshold of the generated agent based on a preset basic risk threshold, combined with a comprehensive protection capability factor and a task criticality factor.

[0209] The alarm status determination unit is used to compare the normalized risk index with the individualized risk tolerance threshold, and combine it with the preset safety margin coefficient to determine and generate the current alarm status.

[0210] The deviation calculation unit is used to calculate the difference between the normalized risk index and the individualized risk tolerance threshold to obtain the risk tolerance deviation.

[0211] The threshold correction unit is used to generate an individualized risk tolerance threshold for the next cycle by closed-loop correction based on the risk tolerance deviation.

[0212] The strategy execution unit is used to execute hierarchical correction strategies based on the current alarm status.

[0213] In this embodiment, the system can be a physical server cluster deployed in the safety control room of a petrochemical plant, or a SaaS platform deployed in the cloud; including:

[0214] The risk field construction unit is a data acquisition and modeling module used to execute S1, based on the acquired multi-dimensional environmental data, such as... Dynamic risk distribution field is constructed using CFD or Gaussian models. ;

[0215] The risk index calculation unit is a risk calculation engine used to execute S2, based on a dynamic risk distribution field. ,pass The formula calculates the normalized risk index of the agent's location. ;

[0216] The protection capability quantification unit is an agent state analysis module used to execute S3, based on the collected multi-source agent state data, such as... ,pass Formula, quantifying the comprehensive protective capability factor of the generated intelligent agent. ;

[0217] The task quantization unit, which is a task instruction parsing module, is used to execute S4 based on the collected task scheduling instructions, such as... ,pass Formulas quantify the key factors for generating intelligent agents in the task. ;

[0218] The individual threshold calculation unit, which is a threshold decision engine, is used to execute S5 based on a preset base risk threshold. , combined and ,pass Formula for calculating the individualized risk tolerance threshold of the generated intelligent agent. ;

[0219] The alarm status determination unit is a status comparator used to execute S6 and compare... and And combined with the safety margin factor Determine and generate the current alarm status;

[0220] The deviation calculation unit, which is an error signal generator, is used to execute S7 and calculate... The risk tolerance deviation was obtained. ;

[0221] The threshold correction unit, which is a closed-loop feedback controller, executes S8 based on the risk tolerance deviation. ,pass The formula, through closed-loop correction, generates an individualized risk tolerance threshold for the next cycle. ;

[0222] And the strategy execution unit, which is an action execution and path planning module, is used to execute S9 and execute the hierarchical correction strategy according to the current alarm status;

[0223] In this embodiment, the above-mentioned units are all logical functional units, which can be implemented on one or more central processing units (CPUs) or graphics processing units (GPUs) running on a server by means of software programming.

[0224] This embodiment provides a feasible system architecture. Through this modular design, the system can automate, real-time, and parallelize complex security management processes.

[0225] This system can process data from massive amounts of sensors in real time and build... Field, and a large number of intelligent agents, computing The high-concurrency data stream completes the perception-assessment-decision-response (S1 to S9) closed loop within milliseconds; this architectural support is the technical prerequisite for realizing the application of the individualized and precise risk navigation described in this invention in large-scale industrial scenarios, and has high scalability, high reliability and high real-time performance.

[0226] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A hierarchical alarm threshold dynamic adjustment control method based on electronic fences, characterized in that, include: S1, based on the collected multi-dimensional environmental data, constructs a dynamic risk distribution field; S2, based on the dynamic risk distribution field, calculates the normalized risk index of the agent's location; S3, based on the collected multi-source state data of the intelligent agent, quantifies and generates the comprehensive protection capability factor of the intelligent agent; S4, based on the collected task scheduling instructions, quantifies and generates the key factors of the intelligent agent's tasks; S5 calculates the individualized risk tolerance threshold of the generated agent based on the preset basic risk threshold, combined with the comprehensive protection capability factor and the mission criticality factor. S6. Compare the normalized risk index with the individualized risk tolerance threshold, and combine them with the preset safety margin coefficient to determine the current alarm status. S7, calculate the difference between the normalized risk index and the individualized risk tolerance threshold to obtain the risk tolerance deviation; S8, based on the risk tolerance deviation, uses closed-loop correction to generate an individualized risk tolerance threshold for the next cycle; S9, based on the current alarm status, execute the graded correction strategy; S3 includes: The equipment protection index and the agent's own state index are obtained by analyzing the multi-source state data of the agent; Based on the equipment protection index and the agent's own state index, the comprehensive protection capability factor is determined through weighted calculation. S4 includes: From the task scheduling instructions, the task priority and task consequence index are obtained; Based on task priority and task consequence index, the key factors of the task are determined through weighted calculation; S5 includes: Based on the comprehensive protection capability factor and the preset first weight coefficient, a protection capability product term is generated; Based on the task criticality factor and the preset second weight coefficient, a task criticality product term is generated; Multiply the product of the basic risk threshold and the protection capability to obtain the intermediate product; The intermediate product is multiplied by the task-critical product term to determine the individualized risk tolerance threshold; pass Formula for calculating the individualized risk tolerance threshold of the generated intelligent agent. ; in, Basic risk threshold, As a comprehensive protective factor, As a key factor in the mission, These are the weighting coefficients; S8 includes: In response to a risk tolerance deviation greater than zero, a threshold tightening factor is calculated and generated based on the risk tolerance deviation and a preset correction coefficient. The individualized risk tolerance threshold and the threshold tightening factor are calculated to determine the individualized risk tolerance threshold for the next cycle. If the risk tolerance deviation is not greater than zero, the individualized risk tolerance threshold will be determined as the individualized risk tolerance threshold for the next period.

2. The hierarchical alarm threshold dynamic adjustment control method based on electronic fence according to claim 1, characterized in that, S2 include: Obtain the real-time or predicted gas concentration at the location of the intelligent agent; Obtain the preset danger threshold corresponding to the real-time or predicted gas concentration; The normalized risk index is obtained by normalizing the real-time or predicted gas concentration with the preset danger threshold.

3. The hierarchical alarm threshold dynamic adjustment control method based on electronic fence according to claim 1, characterized in that, S6 include: If the normalized risk index is not greater than the individualized risk tolerance threshold, the current alarm status is determined to be a safe status. If the normalized risk index is greater than the individualized risk tolerance threshold, but not greater than the product of the individualized risk tolerance threshold and the safety margin coefficient, then the current alarm status is determined to be a level 1 alarm status. If the normalized risk index is greater than the product of the individualized risk tolerance threshold and the safety margin coefficient, the current alarm status is determined to be a level 2 alarm status.

4. The hierarchical alarm threshold dynamic adjustment control method based on electronic fence according to claim 3, characterized in that, S9 includes: If the current alarm status is a safe state, the agent is allowed to perform the predetermined task. If the current alarm status is a Level 1 alarm status, then based on the dynamic risk distribution field, the lowest risk path is planned and pushed. If the current alarm status is a level 2 alarm status, the individualized risk tolerance threshold for the next cycle will be forcibly adjusted to the preset minimum value, and the fastest evacuation route will be planned and pushed.

5. A hierarchical alarm threshold dynamic adjustment control system based on electronic fences, based on the hierarchical alarm threshold dynamic adjustment control method based on electronic fences according to any one of claims 1-4, characterized in that, include: The risk field construction unit is used to construct a dynamic risk distribution field based on the collected multi-dimensional environmental data; The risk index calculation unit is used to calculate the normalized risk index of the agent's location based on the dynamic risk distribution field. The protection capability quantification unit is used to quantify and generate the comprehensive protection capability factor of the agent based on the collected multi-source state data of the agent. The task quantization unit is used to quantify the key factors of the task in generating the intelligent agent based on the collected task scheduling instructions. The individual threshold calculation unit is used to calculate the individualized risk tolerance threshold of the generated agent based on a preset basic risk threshold, combined with a comprehensive protection capability factor and a task criticality factor. The alarm status determination unit is used to compare the normalized risk index with the individualized risk tolerance threshold, and combine it with the preset safety margin coefficient to determine and generate the current alarm status. The deviation calculation unit is used to calculate the difference between the normalized risk index and the individualized risk tolerance threshold to obtain the risk tolerance deviation. The threshold correction unit is used to generate an individualized risk tolerance threshold for the next cycle by closed-loop correction based on the risk tolerance deviation. The strategy execution unit is used to execute hierarchical correction strategies based on the current alarm status.

Images