Safety production hidden danger closed-loop management and control supervision system and risk assessment early warning method

Abstract

The invention relates to the technical field of safety production management, and discloses a safety production hidden danger closed-loop management and control supervision system and a risk assessment early warning method, and the system comprises an environment modeling module which is used for building a digital model of a production unit and calculating a systematic instability coefficient based on historical data; the risk assessment module is used for quantifying the initial risk potential energy in combination with the coefficient and the hidden danger inherent attribute; the risk evolution module is used for dynamically updating risk potential energy according to a time and space coupling rule; the closed-loop verification module is used for judging the actual effectiveness of the intervention measure by comparing the management entropy before and after the intervention after the intervention measure is implemented; and the self-adaptive feedback module is used for automatically updating the systematic instability coefficient in the environment model according to the judgment result. According to the method, risk quantification, dynamic early warning, effect verification and model self-optimization are integrated, a complete closed-loop management process is formed, and the scientificity and foresight of safety production management are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of safety production management technology, specifically to a closed-loop control and monitoring system for safety production hazards and a risk assessment and early warning method. Background Technology

[0002] Safety in production is the lifeline of all industrial enterprises, especially high-risk industries such as process industries, construction, and mining. To prevent and reduce production safety accidents, the core of safety management work is to investigate, assess, manage, and verify potential safety hazards in the production process.

[0003] Currently, commonly used risk assessment methods in the industry, such as Job Condition Hazard Analysis (LEC) and Risk Matrix (LS), provide basic guidance for safety management. However, these traditional methods have gradually revealed their inherent limitations in practice. Existing risk assessment technologies rely heavily on the experience and subjective judgment of assessors, making it difficult to guarantee the consistency and accuracy of assessment results. Furthermore, they often treat risk levels as relatively fixed values, ignoring the dynamic changes in the state of the production unit itself. Even when facing the same hidden danger, the likelihood and severity of an accident in a smoothly operating and well-managed production unit will be drastically different from that in a unit that is chronically unstable and chaotic, but traditional methods struggle to quantify the impact of this environmental "vulnerability."

[0004] Furthermore, traditional management methods typically view each safety hazard in isolation, lacking a systematic analysis of the interrelationships between hazards. In complex production systems, multiple seemingly low-risk hazards may interact due to process correlations or spatial proximity, forming a "risk cluster" that generates an amplifying effect far exceeding the sum of individual risks. This chain reaction of "small hazards" triggering "major accidents" is difficult for traditional methods to predict. In addition, there is a lack of quantitative prediction of the natural evolution trend of risks over time.

[0005] Most critically, existing safety management processes often end at the point of "hazard rectification completed," forming an "open-loop" management model. This model lacks an objective and quantifiable means to verify whether the implemented management interventions are truly effective, whether they have completely eliminated risks, or even whether the interventions themselves have introduced new, undetected negative disturbances to the system. Due to the lack of an effective post-evaluation and feedback mechanism, lessons learned from management practice cannot be distilled into iteratively optimized systematic knowledge, causing safety management to stagnate in a passive response mode and making it difficult to achieve a proactive, continuous, and closed-loop improvement. Summary of the Invention

[0006] The technical problem that this invention aims to solve is that in existing safety production management solutions, risk assessment is usually static, which fails to reflect the evolution of risks over time and ignores the coupling effect between multiple risk events. At the same time, it is insufficient in verifying the effectiveness of rectification measures and dynamically allocating management resources.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0008] The first aspect of this invention provides a closed-loop management and monitoring system for potential safety hazards in production, the system comprising:

[0009] The environment modeling module is configured to build digital models of each production unit within the production environment and to determine a systematic instability coefficient for each of the production units.

[0010] The risk assessment module, which is connected to the environment modeling module, is configured to calculate the initial risk potential of a safety hazard when a safety hazard is identified in a certain production unit, based on the inherent properties of the safety hazard and the systematic instability coefficient of the production unit.

[0011] The risk evolution module, which is connected to the risk assessment module, is configured to dynamically update the risk potential energy of the safety production hazard according to preset time evolution rules and risk coupling rules.

[0012] The closed-loop verification module is configured to determine the effectiveness of the management intervention measures by comparing the management entropy before and after the intervention after the implementation of the management intervention measures for the safety production hazards.

[0013] An adaptive feedback module, which is connected to the environment modeling module and the closed-loop verification module, is configured to update the systematic instability coefficient of the production unit based on the validity result determined by the closed-loop verification module.

[0014] In one specific embodiment, the environment modeling module is specifically configured to perform the following operations: calculate the background management entropy based on the historical operating data of each production unit; wherein, for a production unit j, its background management entropy EME is... j The calculation formula is:

[0015]

[0016] Where, p j (k) represents the probability that a certain operating parameter of the production unit j is in state k, and n is the total number of states.

[0017] Subsequently, the environment modeling module generates the systematic instability coefficient based on the background management entropy.

[0018] Preferably, the risk evolution module is specifically configured to perform one or a combination of the following operations:

[0019] Based on a preset time decay constant, the risk potential energy increases over time, and its calculation formula is as follows:

[0020]

[0021] Among them, PRE i (t new The updated risk potential energy of hazard i is RPE. i (t old Let λ be the risk potential energy of the previously recorded hazard i, e be the base of the natural logarithm, and λ be the risk potential energy of the hazard i. i Let be the time evolution coefficient corresponding to the hidden danger i, and Δt be the time elapsed since the last update;

[0022] When a correlation is identified among multiple safety hazards, the risk potential energy of these hazards is aggregated and enhanced based on a preset coupling coefficient. The calculation formula is as follows:

[0023]

[0024] Among them, RPE cluster (t) represents the total potential energy of the risk cluster at time t, K c H is the coupling coefficient. cluster This refers to the set of potential safety hazards contained within the risk cluster.

[0025] In one specific embodiment, the system further includes:

[0026] The decision support module, which is connected to the risk evolution module, is configured to pre-determine and calculate the management intervention risk ripples caused by the management intervention measures when planning management intervention measures for the safety production hazards.

[0027] Preferably, the decision support module is further configured to perform the following operations:

[0028] Based on the risk potential energy that the management intervention measures are expected to eliminate and the risk ripple of the management intervention, the net risk benefit of one or more alternative management intervention schemes is calculated.

[0029] Based on the net risk-reward ratio, the optimal management intervention plan is recommended.

[0030] Furthermore, the implementation of the management intervention measures for potential safety hazards is based on the optimal management intervention plan recommended by the decision support module.

[0031] In one specific embodiment, the closed-loop verification module is specifically configured to perform the following operations:

[0032] The operating parameters of the production unit before the implementation of the management intervention measures are collected in order to calculate the management entropy before the intervention.

[0033] The operating parameters of the production unit after the implementation of the management intervention measures are collected in order to calculate the management entropy after the intervention;

[0034] When the management entropy after the intervention is less than the management entropy before the intervention, the management intervention measure is deemed effective.

[0035] Preferably, the adaptive feedback module is specifically configured to perform the following operations:

[0036] When the management intervention is determined to be effective, the background management entropy of the production unit is reduced; or when the management intervention is determined to be ineffective, the background management entropy of the production unit is increased.

[0037] Based on the updated background management entropy, the systematic instability coefficient of the production unit is recalculated and updated.

[0038] In one specific embodiment, the system further includes:

[0039] The risk visualization module, which is connected to the risk evolution module, is configured to integrate the risk potential energy of all safety production hazards, generate a global risk field view, and dynamically display it on the user interface.

[0040] A second aspect of the present invention provides a risk assessment and early warning method, the method comprising the following steps:

[0041] Step 1: Establish a digital model of each production unit within the production environment, and determine a systematic instability coefficient for each production unit;

[0042] Step 2: When a safety hazard is identified in a certain production unit, the initial risk potential of the safety hazard is calculated based on the inherent properties of the safety hazard and the systematic instability coefficient of the production unit.

[0043] Step 3: Dynamically update the risk potential energy of the potential safety hazards according to the preset time evolution rules and risk coupling rules;

[0044] Step 4: After implementing management intervention measures for the aforementioned safety hazards, the effectiveness of the management intervention measures is determined by comparing the management entropy before and after the intervention.

[0045] Step 5: Update the systematic instability coefficient of the production unit based on the validity result of the determination.

[0046] This invention provides a closed-loop management and monitoring system for potential safety hazards in production, as well as a risk assessment and early warning method. It has the following beneficial effects:

[0047] 1. This invention calculates the initial risk potential energy by introducing a systematic instability coefficient and dynamically updates the risk potential energy by combining time evolution and risk coupling rules. This enables the risk assessment results of potential safety hazards to reflect the stability of the environment in which they are located, the urgency of their changes over time, and their correlation with other hazards. This provides a dynamic and systematic description of the risk situation and overcomes the problems of static and isolated risk assessment in the prior art.

[0048] 2. This invention, by setting up a decision support module, pre-analyzes the potential ripple effects of management intervention risks before implementation and recommends intervention plans based on net risk-return. This approach quantifies the potential negative impacts of management decisions, enabling the allocation of management resources to be based on predictions of the overall future risk status of the system, rather than solely on the current level of individual risks, thus providing objective data support for decision-making.

[0049] 3. This invention provides an objective, data-driven standard for determining the effectiveness of intervention measures by comparing the management entropy before and after the intervention, overcoming the limitations of relying on subjective judgment or process status markers. Furthermore, by using an adaptive feedback module to update the systemic instability coefficient based on the judgment results, a mechanism is established for dynamically adjusting system risk assessment parameters according to actual management effectiveness, giving the system adaptive capabilities. Attached Figure Description

[0050] Figure 1 This is a functional module block diagram of a closed-loop management and supervision system for safety production hazards according to an embodiment of the present invention;

[0051] Figure 2 This is a flowchart of a risk assessment and early warning method according to an embodiment of the present invention.

[0052] The modules are: 10. Environmental Modeling Module; 20. Risk Assessment Module; 30. Risk Evolution Module; 40. Closed-Loop Verification Module; and 50. Adaptive Feedback Module. Detailed Implementation

[0053] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0054] See attached document Figure 1 , Figure 1 This is a functional block diagram of a closed-loop management and monitoring system for safety production hazards according to an embodiment of the present invention. The present invention provides a closed-loop management and monitoring system for safety production hazards, which can be deployed on one or more server computers. The server computer includes at least one processor, a memory, and a network interface for communicating with other devices. The memory stores computer program instructions, and the processor is configured to execute the computer program instructions to implement the following functional modules.

[0055] The system may include: an environment modeling module 10, a risk assessment module 20, a risk evolution module 30, a closed-loop verification module 40, and an adaptive feedback module 50.

[0056] The environmental modeling module 10 is configured to build digital models of each production unit within the production environment and determine a systematic instability coefficient for each production unit. This module acquires data from sensors, monitoring systems, or manual input terminals on the production site via a data interface and stores it in the system's database.

[0057] The risk assessment module 20 interacts with the environmental modeling module 10. This module is configured to, upon receiving new information about a potential safety hazard, call the environmental modeling module 10 to obtain the systematic instability coefficient of the production unit where the hazard is located, and, combined with the inherent properties of the hazard, calculate its initial risk potential.

[0058] The risk evolution module 30 is connected to the risk assessment module 20. This module is configured to periodically or event-triggered dynamically update the risk potential of all existing but not closed hidden dangers in the system. The updates are based on preset time evolution rules and risk coupling rules.

[0059] The closed-loop verification module 40 is configured to initiate a data observation cycle after receiving a signal indicating the completion of a management intervention measure. During this cycle, the module determines the effectiveness of the management intervention measure by collecting and calculating the management entropy values ​​before and after the intervention.

[0060] The adaptive feedback module 50 is connected to the environment modeling module 10 and the closed-loop verification module 40. This module is configured to adjust the parameters in the environment modeling module 10 based on the judgment result output by the closed-loop verification module 40, specifically by updating the systematic instability coefficient of the affected production unit.

[0061] In one specific embodiment, the system may further include a decision support module 60. The decision support module 60 is connected to the risk evolution module 30 and is used to extrapolate and optimize the planning and management intervention measures, providing decision-making basis to users or automated execution systems.

[0062] In another specific embodiment, the system may further include a risk visualization module 70. The risk visualization module 70 is connected to the risk evolution module 30. This module is configured to acquire global risk potential data from the risk evolution module 30 and generate a graphical view, which is displayed through a user interface on a user terminal device. The user terminal device communicates with the server where the system is located through a network interface.

[0063] The environment modeling module 10 is configured to create a production space topology map, which is a digital model stored in the system database that describes the various entity elements within the production environment and their interrelationships.

[0064] In one specific embodiment, the production space topology graph is constructed as a weighted directed or undirected graph structure, denoted as G = (V, E), where V is the set of vertices and E is the set of edges.

[0065] Each vertex in the vertex set V corresponds to a specific entity element in the production environment. Each vertex is assigned a unique vertex identifier and stores a set of attributes. Vertex types include, but are not limited to: device vertices, used to represent a specific device, whose attributes may include device number, device type, and three-dimensional spatial coordinates of the installation location; region vertices, used to represent a well-defined work area, whose attributes may include the boundary coordinates of the area; and sensor vertices, used to represent a data acquisition point, whose attributes may include sensor type, the measured parameter, and the associated device vertex or region vertex.

[0066] Each edge in the edge set E represents a specific relationship between two vertices. Each edge stores the identifiers of the two vertices it connects, a relationship type, and a weight value. Relationship types include, but are not limited to: physical proximity, used to connect two entity elements whose physical spatial distance is less than a first preset threshold; process flow, used to connect two equipment vertices where there is material or energy flow, with the direction of the edge indicating the flow direction; and resource dependency, used to connect two vertices that need to share the same limited resource, such as sharing the same maintenance team or the same power supply circuit.

[0067] The weight of an edge is a numerical value used to quantify the closeness or importance of the relationship represented by that edge.

[0068] The environmental modeling module 10 receives the basic data used to construct the topology graph through a data input interface. This basic data can be imported in batches in a structured file format (e.g., a tabular file containing an equipment list and coordinates, or an engineering data file containing process flow information), or it can be manually entered by technicians through the graphical user interface provided by the system. After receiving the basic data, the environmental modeling module 10 parses it, generates corresponding vertices and edges, and determines the weight value of each edge according to preset calculation rules (e.g., the weight is the reciprocal function of physical distance). Finally, the constructed graph structure data is stored in the database.

[0069] The established production space topology map serves as the foundational data structure for subsequent calculations in this system. The risk evolution module 30 utilizes the edges in this map to identify potential hazards with coupling relationships, while the decision support module 60 uses this map to deduce the propagation paths of the potential impacts of management intervention measures.

[0070] The environment modeling module 10 is configured to calculate and update the background management entropy for each production unit. Background management entropy is a quantitative indicator used to characterize the degree of uncertainty or disorder in the operating state of a production unit over a long historical period.

[0071] In one specific embodiment, the process of calculating background management entropy by the environment modeling module 10 includes the following steps:

[0072] Step 1: The module obtains time-series data of one or more key operating parameters of a specified production unit within a preset historical time window from the system database or a connected external data source (e.g., a Manufacturing Execution System (MES) or a Supervisory Control System (SCADA)).

[0073] Key operating parameters can be continuous values, such as equipment temperature, pressure, and vibration frequency; or they can be discrete events, such as equipment alarm logs and records of human operational errors.

[0074] Step 2: For the acquired continuous parameter data, the environment modeling module 10 discretizes it, dividing it into n non-overlapping state intervals. For example, for temperature parameters, it can be divided into multiple temperature range intervals, each interval representing a discrete state. For discrete event data, each event type itself is a discrete state.

[0075] Step 3: The environment modeling module 10 counts the frequency with which the operating parameters of the production unit fall into each discrete state k within the historical time window, and calculates the corresponding probability p. j (k). This probability is the number of times state k occurs divided by the total number of observations.

[0076] Step 4: The environment modeling module 10 applies the formula for calculating information entropy to calculate the background management entropy (BME) of production unit j. j The calculation formula is as follows:

[0077]

[0078] Among them, BME j Let p be the background management entropy of production unit j. j (k) represents the probability that the operating parameter of the unit is in state k, and n is the total number of discrete states. If a production unit contains multiple key operating parameters, its total background management entropy can be a weighted sum of the entropy values ​​calculated for each parameter, with the weight values ​​pre-set according to the degree of influence of each parameter on the stability of the unit.

[0079] Calculated background management entropy BME j The value is stored as an attribute of the corresponding vertex j in the production space topology graph. This value is not static; the environment modeling module 10 periodically (e.g., every 30 days) re-executes the above calculation process to update this value. Furthermore, this value is also updated aperiodically according to instructions from the adaptive feedback module 50 to reflect the impact of management interventions on the long-term stability of the unit in real time.

[0080] The environment modeling module 10 calculates the background management entropy (BME) of production unit j. j Subsequently, it was configured to be further based on this BME j The value generates a systematic instability coefficient SIC. j The systemic instability coefficient SIC j It is a multiplier factor used to quantify the amplifying effect of the inherent instability of production unit j on the degree of risk of newly emerging hidden dangers.

[0081] In one specific embodiment, the systematic instability coefficient SIC j Defined as Background Management Entropy (BME)j A monotonically increasing function f sic To ensure that the more chaotic the historical operating state of a production unit (i.e., BME) the better. j The higher the value, the more significant its amplification effect on new risks. This function f sic It is pre-configured in the system and stored in the system configuration database.

[0082] The function f sic It can be specifically implemented as a piecewise linear function or an exponential function. In one embodiment, the function f... sic It is defined as a linear mapping, and its calculation formula is:

[0083] SIC j =1+α×BME j ;

[0084] Among them, SIC j Let BME be the systematic instability coefficient of production unit j. j Let α be the background management entropy of the unit, and let α be a pre-set, positive-zero system parameter used to adjust the sensitivity of the BME to the risk amplification effect. In this formula, the constant 1 ensures that even in the ideal state where the background management entropy approaches zero, the baseline value of the systematic instability coefficient is not less than 1.

[0085] In another embodiment, the function f sic It can be defined as an exponential mapping to reflect the amplification effect of non-linear growth brought about by higher BME values, and its calculation formula is as follows:

[0086]

[0087] Here, β is another pre-defined system parameter that is greater than zero.

[0088] Environment modeling module 10 calculates or updates the BME of a production unit j each time. j After the value is obtained, the preset function f will be called immediately. sic Calculate the new SIC j The value is stored or updated as an attribute of the corresponding vertex j in the production space topology graph. Subsequently, when assessing potential hazards occurring in unit j, the risk assessment module 20 will directly read this SIC. j The value is used to calculate the initial risk potential.

[0089] The risk assessment module 20 is configured to calculate an initial risk potential energy when a new safety hazard is entered into the system.

[0090] In one specific embodiment, when a safety hazard is identified and its information is entered into the system, the risk assessment module 20 is activated.

[0091] The input information must include at least:

[0092] The unique identifier j of the production unit where the potential hazard is located;

[0093] A numerical value S used to represent the inherent severity of the hazard;

[0094] And a numerical value P0 to represent the initial probability of the hazard occurring.

[0095] The values ​​S and P0 can be derived from a pre-defined risk assessment standard library, in which different types of hazards are assigned standardized severity and probability levels.

[0096] After receiving the above input information, the risk assessment module 20 first performs a query operation. It uses the production unit identifier j to query the environmental modeling module 10 or directly the system database to obtain the pre-calculated systematic instability coefficient SIC associated with that production unit j. j .

[0097] Next, the risk assessment module 20 calculates the baseline risk value for the hazard. This calculation is performed using a predefined baseline risk function f. risk Complete. The function takes the inherent severity S of the hazard and the initial probability of occurrence P0 as input.

[0098] In one embodiment, function f risk It can be a multiplication operation, i.e., f risk (S,P0) = S × P0. In another embodiment, the function f risk It can be a lookup operation, that is, in a preset two-dimensional risk matrix, using the levels of S and P0 as indexes, to find and return the corresponding baseline risk value.

[0099] After obtaining the baseline risk value and the systematic instability coefficient SIC j Then, the risk assessment module 20 calculates the final initial risk potential energy (RPE) of the hazard by multiplying the two. init The complete calculation formula is as follows:

[0100] RPE init =f risk (S,P0)×SIC j ;

[0101] Among them, RPE init f represents the initial risk potential energy of potential safety hazards in production. risk(S,P0) is the baseline risk value calculated based on the inherent severity S and the initial probability of occurrence P0, SIC j The system instability coefficient of production unit j where the potential hazard is located.

[0102] After the calculation is completed, the risk assessment module 20 will calculate the initial risk potential energy RPE. init The value, along with other relevant information about the hazard (such as description, location, reporting time, and reporter), is written into a new hazard data record in the database, and the record's status is marked as "pending processing." This RPE init The value will serve as the benchmark for subsequent dynamic evolution calculations in the risk evolution module 30.

[0103] The risk evolution module 30 is configured to perform time-based dynamic updates on the risk potential of all active (i.e. not closed) safety hazards in the system to reflect how the risk level of a hazard will increase over time if it is not addressed.

[0104] In one specific embodiment, the risk evolution module 30 performs this function through a background timed task that is activated at a preset period (e.g., hourly). Each time it is activated, the module first queries the system database to obtain a collection of hazard records for all active states.

[0105] For each hazard i in the set, the risk evolution module 30 reads its currently stored risk potential energy value RPE. i (t old ) and its last update timestamp t old The module then calculates the current time t. new Compared to the last update time t old The time difference between them is Δt = t new -t old .

[0106] Next, the risk evolution module 30 queries and obtains the corresponding time evolution coefficient λ from a preset parameter configuration library based on the type of hazard i. i The time evolution coefficient λ i This is a positive value, and its magnitude is related to the urgency of the hazard type. For example, a hazard type with immediate danger (such as a flammable gas leak) would be assigned a larger λ value. i A value is assigned to a slower-developing hazard type (such as structural corrosion), while a smaller λ value is assigned to it. i value.

[0107] After obtaining Δt and λ iSubsequently, the risk evolution module 30 uses an exponential growth model to calculate the new risk potential value RPE of hazard i at the current moment. i (t new The specific incremental update calculation formula is as follows:

[0108]

[0109] Among them, RPE i (t new The updated risk potential energy of hazard i is RPE. i (t old Let λ be the risk potential energy of the previously recorded hazard i, e be the base of the natural logarithm, and λ be the risk potential energy of the hazard i. i Let be the time evolution coefficient corresponding to the hidden danger i, and Δt be the time elapsed since the last update.

[0110] Calculate the new risk potential energy value RPE i (t new After that, the risk evolution module 30 will use the new value and the current timestamp t new Write back to the database and update the corresponding record for hazard i. Through this periodic update process, the risk potential value of all unclosed hazards will show a non-linear growth trend over time.

[0111] In addition to performing time-based updates, the risk evolution module 30 is also configured to identify and quantify the coupling effects between multiple safety hazards.

[0112] In one specific embodiment, this spatial coupling evolution process can be triggered each time a new safety hazard is entered into the system, or it can be executed as part of a periodic update task. When the process is activated, the risk evolution module 30 uses the production space topology G=(V,E) established by the environment modeling module 10 to identify a set of related hazards.

[0113] Specifically, the risk evolution module 30 first obtains a list of all active hazards in the system and their corresponding vertices (production units). Then, for each hazard, the module checks its corresponding vertex in the topology graph G. If two or more vertices with active hazards are directly or indirectly connected by one or more edges that meet preset conditions, these hazards are identified as a risk cluster. The preset conditions may include: the edge type is "process flow relationship", or the edge type is "physical proximity relationship" and its weight value is greater than a preset coupling threshold.

[0114] Once a risk cluster H cluster Once identified, the risk evolution module 30 calculates the aggregate risk potential (RPE) of the risk cluster. cluster(t). This calculation aims to reflect the enhancement effect resulting from the interaction of multiple hidden dangers within a cluster. The calculation formula is as follows:

[0115]

[0116] Among them, RPE cluster H(t) represents the total potential energy of the risk cluster at time t. cluster RPE is the set of safety hazards contained within a risk cluster. i (t) is the set H cluster The risk potential energy value of the i-th hidden danger at time t, after time evolution and update.

[0117] K c K is the coupling coefficient of the risk cluster. This coefficient is a value greater than 1, and its specific value depends on the nature of the connections between the hazards constituting the risk cluster. In one embodiment, K c The value can be derived from the properties of the edges connecting these potential risks. For example, a risk cluster formed by edges connecting "process flow relationships" has a K value. c The value is set to 1.5; and a risk cluster formed by edges connected by "physical proximity" has a K value. c The value is set to 1.2.

[0118] Calculated aggregation risk potential RPE cluster (t) is stored as a new data entity in the system database and associated with all individual hazards within that risk cluster. This aggregated risk potential value does not replace the risk potential value of an individual hazard, but exists as an independent, higher-level risk indicator to highlight complex risk areas with higher hazard levels formed by the interaction of multiple hazards in risk ranking and visualization.

[0119] The decision support module 60 is configured to pre-analyze and quantify the secondary risks that the measure itself may introduce when planning management intervention measures for one or more safety hazards, namely the management intervention risk ripple (IRR).

[0120] In one specific embodiment, the simulation process is triggered when a user constructs an alternative management intervention plan in the system interface. The alternative plan is a structured data object that includes at least: identifiers of one or more target hazards, a specific intervention action (e.g., equipment shutdown for maintenance, component replacement, area isolation), and the expected duration of the action.

[0121] After receiving the alternative solution, the decision support module 60 first performs the following steps:

[0122] Step 1: In the production space topology graph G=(V,E) established by the environment modeling module 10, locate one or more vertices that the intervention action directly affects. These vertices are called intervention origin vertices. For example, if a pre-action is "to shut down and maintain equipment A", then the vertex corresponding to equipment A is the intervention origin vertex.

[0123] Step 2, the decision support module 60 resolves the intervention action as one or more temporary negative states imposed on the origin vertex of the intervention. For example, "downtime maintenance" is resolved as the vertex's "availability" state being "no" and "output" state being "zero" for the expected duration.

[0124] Step 3: Starting from the origin vertex of the intervention, the decision support module 60 traverses the edges of the production space topology graph G to identify all associated vertices affected by the negative state. This traversal process is based on the type of edge; for example, it traces downstream along edges related to "process flow relationships," or it traces all other vertices sharing the same resource along edges related to "resource dependencies." All traversed vertices are identified as affected vertices.

[0125] Step 4: For each vertex k determined to be affected, the decision support module 60 calculates a risk ripple score (IRR). k This score is based on a preset impact function f. impact The calculated function takes into account the nature of the intervention action, the path properties from the intervention origin to vertex k, and the properties of vertex k itself.

[0126] Finally, the risk ripple scores of all affected vertices are summed to obtain the total management intervention risk ripple value (IRR) for the alternative management intervention plan. total The calculation formula is as follows:

[0127]

[0128] Among them, IRR total V represents the total risk ripple value of this alternative solution. affected For the set of all affected vertices; I represents the original intervention action; path(V) origin ,V k Attr(V) represents path information from the origin vertex of the intervention to the affected vertex k, such as path length and the weights of edges along the path; k ) represents the properties of the affected vertex k itself, such as its inherent importance level.

[0129] The calculated IRR total The value is stored as a quantitative attribute of the alternative management intervention plan for subsequent optimal solution decision-making.

[0130] After calculating the respective ripple risk (IRR) of management intervention for multiple alternative management intervention options, the decision support module 60 is further configured to execute a decision model to identify and recommend an optimal option from these alternatives.

[0131] In one specific embodiment, the decision-making process is activated when two or more alternative management intervention options exist on the system interface. The decision support module 60 first calculates the net risk benefit (RNB) for each alternative. This calculation includes the following steps:

[0132] Step 1: For a given alternative, the decision support module 60 identifies one or more target hazards to be addressed. The module then queries the risk evolution module 30 or directly from the system database to obtain the risk potential energy (RPE) values ​​of these target hazards at the current moment, and sums these values ​​to obtain the expected risk reduction amount (RPE) of the proposed solution. eliminated .

[0133] Step 2: The decision support module 60 obtains the total management intervention risk ripple value (IRR) calculated in the previous stage for the alternative solution. total .

[0134] Step 3: The decision support module 60 calculates the net risk benefit RNBRNB of the alternative by subtracting the total management intervention risk ripple value from the expected risk elimination amount. The calculation formula is as follows:

[0135] RNB = RPE eliminated -IRR total ;

[0136]

[0137] Wherein, RNB represents the net risk-reward of the alternative management intervention, and RPE... eliminated H represents the total potential energy of risk that the plan is expected to eliminate. target RPE is the set of target hazards addressed by this solution. i (t) represents the risk potential energy of the target hazard i at the current time t, and IRR total This represents the total management intervention risk ripple value for the plan.

[0138] The decision support module 60 repeats the above steps for all alternative management intervention options, and calculates a corresponding net risk return (RNB) value for each option.

[0139] Finally, the decision support module 60 compares all calculated RNB values ​​and identifies the alternative management intervention with the highest RNB value as the optimal solution. The identifier of this optimal solution is then output. This output can be used to highlight or prioritize this optimal solution on the user interface, providing managers with a decision-making basis based on quantitative data comparison, enabling them to choose the appropriate solution.

[0140] The closed-loop verification module 40 is configured to quantitatively evaluate the actual effect of a management intervention by calculating and comparing the management entropy (ME) before and after the intervention after the intervention is marked as "completed" in the system.

[0141] In one specific embodiment, when the closed-loop verification module 40 is activated, it first identifies the set of production units associated with the completed management interventions. This set includes production units directly affected by the interventions, as well as other related production units that were deduced to be affected by the management intervention risk ripple (IRR) during the decision support phase. This set of production units is defined as the "observation domain" for this verification.

[0142] Subsequently, the closed-loop verification module 40 performs the data acquisition step. It queries and retrieves data from the system database within two independent time windows. The first time window is the "pre-intervention window," defined as a preset time period (e.g., 7 days) immediately preceding the start of the management intervention. The second time window is the "post-intervention window," defined as a time period of the same length immediately preceding the start of the management intervention after its completion. For both time windows, the module collects time-series data of one or more key operating parameters for all production units within the observation domain.

[0143] Next, the closed-loop verification module 40 calculates the corresponding management entropy for the datasets in these two time windows. This is used to calculate the management entropy ME before intervention. pre For example, the process includes: summarizing all data collected within the pre-intervention window and discretizing it into n non-overlapping state intervals; counting the frequency of the entire observation domain falling into each discrete state k and calculating its corresponding probability p. pre (k); Finally, the information entropy formula is applied for calculation. The exact same process is repeated for the post-intervention window data to calculate the post-intervention management entropy ME. post .

[0144] The calculation formula is as follows:

[0145]

[0146] Among them, ME pre and ME postThese represent the management entropy before and after the intervention; p pre (k) and p post (k) represents the probability that the operating state of the observation domain is in state k within the windows before and after the intervention, respectively; n is the total number of discrete states.

[0147] After the calculation is completed, the closed-loop verification module 40 will calculate the ME. pre and ME post Two numerical values ​​are stored as a data pair and then output. This pair of values ​​provides an objective and quantitative data foundation for subsequent validity determination and adaptive feedback.

[0148] Calculate the management entropy ME before intervention. pre Management Entropy (ME) after Intervention post Subsequently, the closed-loop verification module 40 is further configured to execute a judgment logic to determine whether the corresponding management intervention measures have achieved a "deep closed loop".

[0149] In one specific embodiment, the core of this determination logic is to compare ME. pre and ME post The value of . The closed-loop verification module 40 first calculates the difference ΔME between the two management entropy values, and its calculation formula is:

[0150] ΔME=ME pre -ME post ;

[0151] Next, the closed-loop verification module 40 compares the calculated ΔME value with two pre-configured thresholds in the system. These two thresholds are: a success threshold θ... success , which is a value greater than zero; and a failure threshold θ. failure It is a value less than or equal to zero.

[0152] The decision logic includes the following three branches:

[0153] 1. If ΔME > θ success This indicates a significant reduction in management entropy after the intervention compared to before. This suggests that the management intervention not only eliminated existing safety hazards but also made the overall operational status of the affected observation domain more orderly and stable. In this case, the closed-loop verification module 40 classifies the result of this intervention as "deep closed loop."

[0154] 2. If θ failure ≤ΔME≤θ syccessThis indicates that the management entropy after the intervention has not changed significantly compared to before the intervention, or has only decreased or increased slightly. This suggests that the management intervention has achieved its intended goal (i.e., eliminated the hidden danger), but has failed to have a significant positive impact on the overall stability of the observation domain, or has introduced negligible negative disturbances. In this case, the closed-loop verification module 40 determines the result of this intervention as "shallow closed loop".

[0155] 3. If ΔME < θ failure This indicates that the management entropy after the intervention was significantly higher than before the intervention. This suggests that while the management intervention may have resolved the apparent hidden problems, its implementation caused significant negative disturbances to the system, making the overall operating state of the observation domain more chaotic and unstable. In other words, the predicted management intervention risk ripple (IRR) was confirmed and had a substantial impact. Under these circumstances, the closed-loop verification module 40 classifies the result of this intervention as a "negative closed loop."

[0156] After the determination is completed, the closed-loop verification module 40 transmits the final determination result (i.e., "deep closed loop", "shallow closed loop" or "negative closed loop") as an output signal to the adaptive feedback module 50 to trigger subsequent parameter adjustment actions.

[0157] The adaptive feedback module 50 is configured to adjust the core parameters in the environmental modeling module 10 based on the judgment results output by the closed-loop verification module 40, so that the system model can reflect the real effects of management intervention measures.

[0158] In one specific embodiment, the adaptive feedback module 50 is activated upon completion of the verification process for a management intervention. This module receives a data packet from the closed-loop verification module 40, which includes at least: a set of production units identified by the intervention, i.e., an "observation domain"; and a management entropy difference ΔME calculated for the observation domain.

[0159] The core function of the adaptive feedback module 50 is to use the received ΔME value to calculate the background management entropy BME for each production unit j within the observation domain. j Adjustments are made. This adjustment process aims to feed back and solidify the effects of a single intervention into the parameters characterizing the long-term stability of the production unit.

[0160] Specifically, for each production unit j in the observation domain, the adaptive feedback module 50 performs the following update operation. The module first reads the current background management entropy (BME) of that unit from the database. j,old Then, the module calculates its new background management entropy (BME) based on a preset adjustment function. j,newIn one embodiment, the adjustment function is defined as a linear update rule, calculated as follows:

[0161] BME j,new =BME j,old -η×ΔME;

[0162] Among them, BME j,new For the new background management entropy adjusted for production unit j, BME j,old ΔME is the original background management entropy before production unit j is adjusted, ΔME is the management entropy difference calculated by the closed-loop verification module 40, and η is a pre-set system parameter that is greater than zero, called the learning rate. This parameter is used to control the magnitude of the impact of a single verification result on the long-term stability parameter.

[0163] After calculating the new background management entropy BME j,new Next, the adaptive feedback module 50 writes the new value to the database, updating the vertex attributes corresponding to production unit j. Subsequently, the module triggers the environment modeling module 10, or calls the corresponding function itself, based on the updated BME. j,new The value is obtained by using the preset function f. sic (e.g., SIC) j =1+α×BME j Recalculate the systematic instability coefficient SIC of this unit. j In this way, a successful intervention (ΔME > 0) will reduce the BME and SIC values ​​of the relevant units, leading the system model to deem it more stable. Conversely, a negative intervention (ΔME < 0) will increase these values, reflecting an increase in instability.

[0164] The risk visualization module 70 is configured to generate and present an integrated, interactive risk topology map to the user. This map comprehensively displays abstract risk data in an intuitive graphical way, providing managers with a global, dynamically updated risk situation awareness interface.

[0165] In one specific embodiment, the generation process of the integrated risk topology map is as follows. First, the risk visualization module 70 reads the structural data of the production space topology map G = (V, E) established by the environment modeling module 10. The module renders the vertices (production units) of the topology map as graphical nodes on the screen and renders the edges as lines connecting the corresponding nodes. The layout of the nodes can be mapped according to the actual coordinates of the physical space, or it can be automatically arranged by a force-guided layout algorithm, so that closely related nodes appear closer together visually.

[0166] Based on this, the risk visualization module 70 overlays various risk-related data calculated by the system as multiple layers onto the topology map.

[0167] Background stability layer: The background color of each node is dynamically colored according to its systematic instability coefficient (SIC) value. For example, a color spectrum from green to red is used, where green represents a very low SIC value (highly stable) and bright red represents a very high SIC value (highly unstable), thus making the inherent vulnerabilities in the entire production environment immediately apparent.

[0168] Active Hazard Layer: For each active safety hazard, the module overlays a unique risk marker on or next to its associated node. The visual attributes of this marker are used to encode the hazard's Risk Potential (RPE). For example, the marker size is proportional to its RPE value; a hazard with a high RPE value will be displayed with a large marker to attract user attention. Furthermore, the marker's shape or icon can be used to distinguish different hazard types; for example, a flame icon can represent a fire hazard.

[0169] Risk Cluster Layer: When the risk evolution module 30 identifies a risk cluster consisting of multiple coupled vulnerabilities, the risk visualization module 70 draws a semi-transparent highlighted area around all nodes belonging to that cluster. The color and transparency of this area are related to its aggregate risk potential (RPE). cluster In conjunction with this, a risk cluster with high aggregate risk potential will be surrounded by a striking, dark area, visually marking the most dangerous complex risk area in the system.

[0170] This integrated risk topology map is interactive. Users can pan and zoom using the mouse to browse different areas of the entire topology map. When the mouse hovers over any node or risk marker, the system will pop up an information dialog box displaying detailed data for that element, such as the name of the production unit, the precise SIC value, a detailed description of the hazard, and the real-time RPE value.

[0171] See attached document Figure 2 , Figure 2 This is a flowchart of a risk assessment and early warning method according to an embodiment of the present invention. The present invention provides a method for risk assessment and early warning of closed-loop management and supervision of safety production hazards, the specific implementation of which is as follows:

[0172] This method first establishes digital models of each production unit within the production environment through an environmental modeling module, and assigns a systematic instability coefficient to each production unit. In this step, all equipment, devices, areas, and other production units on-site are identified, and a digital production space topology map is constructed based on their technological relationships, material handling relationships, and physical proximity relationships. Subsequently, for each production unit, the system retrieves historical data of its key operating parameters over a long period. By analyzing the dispersion and disorder of these data fluctuations over time, the inherent stability of the unit is quantified, and this is calculated as a background management entropy value. Finally, according to preset conversion rules, the system transforms this background management entropy value into a systematic instability coefficient, which directly reflects a production unit's inherent ability to amplify potential risks.

[0173] When a safety hazard is identified within a production unit, this method uses a risk assessment module to calculate the initial risk potential of the safety hazard based on its inherent attributes and the systematic instability coefficient of the production unit. In this step, when a new safety hazard is reported to the system, the system first matches the inherent severity level and initial probability level of the hazard type from its built-in risk knowledge base. The system multiplies these two levels to obtain a baseline risk value that does not consider environmental factors. Next, the system obtains the pre-calculated systematic instability coefficient of the production unit where the hazard is located. Finally, by multiplying the baseline risk value by this systematic instability coefficient, the initial risk potential of the hazard, taking into account environmental amplification effects, is obtained.

[0174] This method uses a risk evolution module to dynamically update the risk potential energy of safety hazards based on preset time evolution rules and risk coupling rules. This step includes two aspects. First, the system automatically increases the risk potential energy values ​​of all unaddressed hazards at fixed time intervals, with the rate of increase determined by the type of hazard, to simulate the objective law of risk deterioration over time. Second, the system uses a production space topology map to automatically identify multiple hazards that are interconnected in terms of process or space, defining them as a risk cluster. The system calculates the sum of the risk potential energy of all hazards within this risk cluster and multiplies it by a coupling coefficient greater than one, thereby obtaining a higher aggregated risk potential energy to quantify the aggravating effect of the interaction of multiple hazards.

[0175] After implementing management intervention measures to address safety hazards, this method uses a closed-loop verification module to compare the management entropy before and after the intervention to determine the effectiveness of the intervention measures. In this step, once a management intervention measure is marked as completed in the system, the system collects key operating parameters of all affected production units within two time windows: before and after the measure's implementation. The system calculates the degree of disorder in the overall operating status of the affected area within these two time windows, obtaining two values: the management entropy before and after the intervention. By comparing the difference between these two values, if the post-intervention management entropy is significantly lower than before the intervention, the measure is considered excellent; if there is no significant change, the effect is considered average; if it increases, a negative impact is considered to have occurred.

[0176] This method uses an adaptive feedback module to update the systematic instability coefficient of production units based on the validity of the judgment. In this step, the system feeds back the judgment result from the previous step to the initial model. If the judgment result is excellent, the system will automatically lower the background management entropy value of the relevant production unit; conversely, if the judgment result is poor or has a negative impact, the background management entropy value will be increased accordingly. Subsequently, the system recalculates and updates the systematic instability coefficient of these production units based on the adjusted background management entropy. In this way, the system model can learn from each management practice, making its assessment of production environment stability increasingly closer to reality, thus forming a continuously optimizing adaptive closed loop.

[0177] Example:

[0178] To more clearly illustrate the technical solution of the present invention, the present invention will be described in detail below with reference to a specific application case in a chemical production scenario.

[0179] The application scenario of this embodiment is a chemical synthesis workshop. The core production process of the workshop includes a reactor R-101, a pump P-102 for conveying materials, and a downstream separation tower T-103.

[0180] Step 1: Production space modeling and initial parameter calculation

[0181] First, the environment modeling module 10 is used to create a topological map of the production space of the workshop. In this map, reactor R-101, pump P-102, and separation tower T-103 are created as three vertices. According to the process flow, an edge from P-102 to R-101 and an edge from R-101 to T-103 are created, and the type of the edges is labeled as "process flow relationship".

[0182] The environmental modeling module 10 collects the temperature records of reactor R-101 over the past 90 days and discretizes them into three states: normal (60 to 80 degrees Celsius), slightly high (81 to 90 degrees Celsius), and extremely high (above 91 degrees Celsius).

[0183] Statistics show that the probabilities of these three states are p(normal) = 0.70, p(high) = 0.20, and p(very high) = 0.10, respectively.

[0184] Based on this, the background management entropy (BME) of R-101 is calculated. R-101 The value is 1.29. Assuming the formula for calculating the systematic instability coefficient is SIC = 1 + 0.5 × BME, then the systematic instability coefficient SIC of R-101 is... R-101 It was calculated to be 1.645. Meanwhile, pump P-102, due to its stable operation, had a SiC... P-102 It was calculated as 1.25.

[0185] Step 2: Hazard entry and initial risk potential assessment

[0186] An inspector discovered a minor leak in the agitator seal of reactor R-101 and recorded this hazard in the system. Risk assessment module 20 was activated. The inherent severity S of this type of hazard in the standard library is 8 (out of a maximum of 10), and the initial probability of occurrence P0 is 0.3. The baseline risk value was calculated as S × P0 = 8 × 0.3 = 2.4.

[0187] Risk assessment module 20 then reads the SIC of R-101. R-101 The value is 1.645. Ultimately, the initial risk potential energy (RPE) for this hazard is... init Calculated as:

[0188] RPE init =2.4 × 1.645 = 3.948;

[0189] The potential hazard was marked as "pending action" and stored in the database, with an initial risk potential energy of 3.948.

[0190] Step 3, Dynamic Evolution of Risk Potential Energy

[0191] If a hazard remains unaddressed 24 hours after being recorded, the risk evolution module 30 performs a time-based update. For this "sealing leak" type of hazard, the time evolution coefficient λ is set to 0.01 (per hour). After 24 hours, the risk potential energy (RPE) of this hazard increases as follows:

[0192] RPE new =3.948×e 0.01×24 ≈5.018;

[0193] At this point, a new potential hazard was recorded in the system: the outlet pressure of pump P-102 exhibited abnormal fluctuations, with a calculated current risk potential energy of 2.0. Since R-101 and P-102 are connected in the topology diagram via a "process flow relationship" edge, the risk evolution module 30 identified this as a risk cluster. The coupling coefficient K for this type of connection... c It is set to 1.5. Therefore, the aggregate risk potential RPE is... cluster The calculation is as follows:

[0194] RPE cluster =1.5×(5.018+2.0)=10.527;

[0195] In the topology diagram presented by the risk visualization module 70, nodes R-101 and P-102 are surrounded by a prominent highlighted area to alert managers to the complex risks present there.

[0196] Step 4, Decision on the optimal intervention plan

[0197] In response to the two related potential risks mentioned above, the administrator created two alternative management intervention plans in the system.

[0198] Option A: Immediately shut down reactor R-101 and replace its agitator seal. This option completely eliminates the potential hazard of R-101 (expected risk reduction of 5.018), but the shutdown will cause a feed interruption in downstream separation tower T-103. Decision support module 60 calculates the management intervention risk ripple (IRR) for this option. A It is version 3.0.

[0199] Option B: Implement online live leak sealing for R-101 and assign dedicated personnel to strengthen the monitoring and adjustment of P-102. This option can address two potential risks simultaneously (expected risk reduction is 5.018 + 2.0 = 7.018), with minimal impact on production and a low management intervention risk ripple effect (IRR). B It is estimated to be 0.8.

[0200] Decision support module 60 calculates the net risk return (RNB) for both options:

[0201] RNB A =5.018 - 3.0 = 2.018;

[0202] RNB B =7.018 - 0.8 = 6.218;

[0203] Due to RNB B RNB A The system marks solution B as the optimal solution and recommends it to the administrator.

[0204] Step 5, Closed-loop verification and adaptive feedback

[0205] The manager adopted and implemented Option B. One week later, the intervention was marked as "completed." The closed-loop verification module 40 was activated. The module collected the operating parameters of reactor R-101 and pump P-102 (i.e., the observation field) for one week before and after the implementation of Option B. The management entropy ME before the intervention was calculated. pre =1.80, post-intervention management entropy ME post =1.10. Management entropy difference ΔME = 1.80 - 1.10 = 0.7.

[0206] Assuming the system's success threshold θ success The value is 0.5. Since ΔME = 0.7 > 0.5, this intervention is judged to be a "deep closed loop".

[0207] The adaptive feedback module 50 receives this result. Assuming the learning rate η is 0.1, the module adjusts the original background management entropy of reactor R-101 as follows:

[0208] BME R-101,new =1.29 - 0.1 × 0.7 = 1.22;

[0209] Subsequently, based on the new BME value, the systematic instability coefficient SIC of R-101 was calculated. R-101 It has been updated to version 1.61. This indicates that, within the system model, a successful management intervention improved the long-term operational stability of reactor R-101, and correspondingly reduced its amplification effect on future risks. The entire management process forms a complete, self-optimizing closed loop.

[0210] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A closed-loop management and monitoring system for safety production hazards, characterized in that: include: The environment modeling module is used to establish digital models of each production unit within the production environment and to determine a systematic instability coefficient for each production unit. The risk assessment module is used to calculate the initial risk potential of a safety hazard when a safety hazard is identified in a certain production unit, based on the inherent properties of the safety hazard and the systematic instability coefficient of the production unit. The risk evolution module is used to dynamically update the risk potential energy of the safety production hazards according to preset time evolution rules and risk coupling rules; The closed-loop verification module is used to determine the effectiveness of the management intervention measures after they are implemented in response to the safety production hazards by comparing the management entropy before and after the intervention. An adaptive feedback module is used to update the systematic instability coefficient of the production unit based on the validity result determined by the closed-loop verification module.

2. The closed-loop management and supervision system for safety production hazards according to claim 1, characterized in that, The environment modeling module is specifically used for: Based on the historical operating data of each production unit, the background management entropy, which characterizes the degree of uncertainty, is calculated. The systemic instability coefficient is generated based on the background management entropy.

3. The closed-loop management and supervision system for safety production hazards according to claim 1, characterized in that, The risk evolution module is specifically used for: Based on a preset time decay constant, the risk potential energy increases over time; When a correlation is identified among multiple safety hazards, the risk potential energy of the multiple safety hazards is aggregated and enhanced based on a preset coupling coefficient.

4. The closed-loop management and supervision system for safety production hazards according to claim 1, characterized in that, Also includes: The decision support module is used to pre-determine and calculate the management intervention risk ripples caused by the management intervention measures when planning management intervention measures for the aforementioned safety production hazards.

5. The closed-loop management and supervision system for safety production hazards according to claim 4, characterized in that, The decision support module is also used for: Based on the risk potential energy that the management intervention measures are expected to eliminate and the risk ripple of the management intervention, the net risk benefit of one or more alternative management intervention schemes is calculated. Based on the net risk-reward ratio, the optimal management intervention plan is recommended.

6. The closed-loop management and supervision system for safety production hazards according to claim 5, characterized in that, The implementation of the management intervention measures for potential safety hazards is based on the optimal management intervention plan recommended by the decision support module.

7. The closed-loop management and supervision system for safety production hazards according to claim 1, characterized in that, The closed-loop verification module is specifically used for: The operating parameters of the production unit before the implementation of the management intervention measures are collected in order to calculate the management entropy before the intervention. The operating parameters of the production unit after the implementation of the management intervention measures are collected in order to calculate the management entropy after the intervention; When the management entropy after the intervention is less than the management entropy before the intervention, the management intervention measure is deemed effective.

8. The closed-loop management and supervision system for safety production hazards according to claim 7, characterized in that, The adaptive feedback module is specifically used for: When the management intervention is determined to be effective, the background management entropy of the production unit is reduced; or when the management intervention is determined to be ineffective, the background management entropy of the production unit is increased. Based on the updated background management entropy, the systematic instability coefficient of the production unit is recalculated and updated.

9. The closed-loop management and supervision system for safety production hazards according to claim 1, characterized in that, Also includes: The risk visualization module is used to integrate the potential risks of all potential safety hazards in production, generate a global risk field view, and dynamically display it on the user interface.

10. A risk assessment and early warning method, based on the system described in any one of claims 1-9, characterized in that, Includes the following steps: Establish a digital model of each production unit within the production environment, and determine a systematic instability coefficient for each production unit; When a safety hazard is identified in a certain production unit, the initial risk potential of the safety hazard is calculated based on the inherent properties of the safety hazard and the systematic instability coefficient of the production unit. The risk potential energy of the safety production hazards is dynamically updated according to the preset time evolution rules and risk coupling rules; After implementing management intervention measures to address the aforementioned safety hazards, the effectiveness of the management intervention measures is determined by comparing the management entropy before and after the intervention. Based on the validity result of the determination, the systematic instability coefficient of the production unit is updated.

Images