Park environment risk priority management and control dynamic sorting system based on Bayesian network

Abstract

The invention provides a park environment risk priority management and control dynamic sorting system based on a Bayesian network. The system comprises a data acquisition layer, a data analysis layer and an output layer, the data acquisition layer is the basis of park environment risk assessment and is used for acquiring real-time and historical data from different sources; the data analysis layer constructs a dynamic index weight matrix by using a Bayesian network and is used for calculating a dynamic risk score RSI of an enterprise; the output layer is used for calculating the comprehensive risk score of the enterprise through the weight generated by the Bayesian network, dynamic weight adjustment and multi-source data fusion are utilized, an original static major / larger risk enterprise list is upgraded to a real-time dynamic sorting list, accurate resource allocation is achieved, and the risk assessment efficiency is improved. Through dynamic index fusion and Bayesian network weight adjustment, traditional static classification is upgraded to a real-time sorting priority management and control list, and an intelligent solution is provided for park environment risk management.

Description

Technical Field

[0001] The present invention relates to the technical field of park environmental risk control, and specifically to a dynamic ranking system for priority control of park environmental risks based on a Bayesian network. Background Art

[0002] Park environmental risk control is a systematic management method aimed at identifying, assessing, and controlling potential risks related to the environment to reduce negative environmental impacts and protect the sustainability of the environment. Specifically, environmental risk control includes risk identification, risk assessment, and risk management. Risk identification is to identify potential environmental risks related to specific industrial activities, projects, or locations, including pollutant emissions, waste generation, resource consumption, and damage to natural resources. Risk assessment is to quantitatively or qualitatively evaluate the identified risks to determine their potential impacts on the environment and ecosystem. Risk management is to reduce or eliminate potential environmental risks by formulating and implementing risk management plans and specifying monitoring, control, and corrective measures. Currently, park environmental risk assessment mainly relies on the expert static grading method: experts classify enterprises into three levels of major risks, relatively large risks, and general risks based on indicators such as enterprise production processes and hazardous substance storage amounts, forming a fixed priority control list.

[0003] However, the current assessment method has the following deficiencies: data lag, the assessment results rely on historical data and cannot reflect the current risk changes of enterprises in real time, including hidden danger rectification and recent accidents; strong subjectivity, the grading depends on expert experience and it is difficult to quantify dynamic factors, including public complaints and intelligent device alarms; resulting in waste of resources, major / relatively large risk enterprises occupy the priority list for a long time and other enterprises with short-term high-incidence risks cannot be dynamically identified. Summary of the Invention

[0004] The present invention provides a dynamic ranking system for priority control of park environmental risks based on a Bayesian network, which can effectively solve the following deficiencies in the above-mentioned background art: data lag, the assessment results rely on historical data and cannot reflect the current risk changes of enterprises in real time, including hidden danger rectification and recent accidents; strong subjectivity, the grading depends on expert experience and it is difficult to quantify dynamic factors, including public complaints and intelligent device alarms; resulting in waste of resources, major / relatively large risk enterprises occupy the priority list for a long time and other enterprises with short-term high-incidence risks cannot be dynamically identified.

[0005] To achieve the above object, the present invention provides the following technical solution: A dynamic ranking system for priority control of park environmental risks based on a Bayesian network, which upgrades the traditional static grading to a real-time ranking priority control list through dynamic index fusion and Bayesian network weight adjustment, providing an intelligent solution for park environmental risk management;

[0006] The system includes a data acquisition layer, a data analysis layer, and an output layer;

[0007] The data acquisition layer is the basis for the environmental risk assessment of the park, responsible for obtaining real-time and historical data from different sources for real-time monitoring and early warning;

[0008] The data analysis layer uses a Bayesian network to construct a dynamic index weight matrix for calculating the dynamic risk score RSI of an enterprise;

[0009] The output layer calculates the comprehensive risk score of an enterprise through the weights generated by the Bayesian network, and the comprehensive risk score is given by weighting all the indicators in the Bayesian network.

[0010] According to the above technical solution, the data acquisition layer integrates an intelligent system alarm module, an environmental protection complaint platform, an accident database, and a hidden danger investigation module. Among them, the intelligent system alarm module is an IoT sensor, and the hidden danger investigation module includes hidden dangers reported by enterprises and hidden dangers identified by AI;

[0011] By using the intelligent system alarm module, the environmental protection complaint platform, the accident database, and the hidden danger investigation module to construct a comprehensive data acquisition network, and obtaining data sources through the data acquisition network, it is possible to obtain the changes in the environmental risks in the park in real time, providing data support for subsequent risk assessment and analysis.

[0012] According to the above technical solution, the intelligent system alarm module collects real-time data of the park environment through the Internet of Things (IoT) sensors, specifically including gas concentration, liquid level, temperature and humidity, flow rate, and pressure. When the real-time environmental data in the park exceeds the safety threshold, an alarm is triggered;

[0013] The environmental protection complaint platform collects complaint information from residents, enterprise employees, and enterprises about environmental pollution problems in the park, and reflects the actual environmental problems and safety risks through the complaint information;

[0014] The accident database records various pollution and work safety accidents that have occurred in the park for analyzing potential risk points;

[0015] The hidden danger investigation module uses the means of enterprise reporting and AI identification to investigate hidden dangers. Enterprise reporting is through enterprises reporting safety hidden dangers online, and AI identification is using artificial intelligence (AI) technology to automatically analyze videos and pictures to identify potential risk hidden dangers and automatically give early warnings.

[0016] According to the above technical solution, the data analysis layer includes the following implementation steps:

[0017] A. Dynamic index normalization processing;

[0018] B. Construct a Bayesian network;

[0019] C. Perform dynamic sorting according to the rules.

[0020] According to the above technical solution, the said A is to standardize five key indicators: the number of alarms, the number of complaints, the number of accidents, the number of unrectified hidden dangers, and the rectification delay rate, ensuring that these indicator data can all be quantitatively processed, facilitating subsequent comparison of weight calculation and risk scoring, avoiding errors caused by large values of different units, and ensuring fairness.

[0021] According to the above technical solution, the said B, the Bayesian network is a graphical model used to qualitatively and quantitatively describe uncertainty and probabilistic dependence relationships, and can more flexibly handle complex risk assessment and prediction problems;

[0022] The constructed Bayesian network includes parent nodes, a conditional probability table CPT, and child nodes. When constructing the Bayesian network, each type of dynamic indicator serves as a parent node, and the initial weight is set through the conditional probability table CPT obtained by training with historical data. The risk level serves as the child node, and the weight will be adjusted in real time as the status changes on various indicators.

[0023] According to the above technical solution, the said parent nodes include five types of dynamic indicators: the number of alarms, the number of complaints, the number of accidents, the number of unrectified hidden dangers, and the rectification delay rate;

[0024] The said conditional probability table CPT trains the initial weight based on historical data. Each conditional probability table CPT is trained according to the historical data of its corresponding indicator. As the actual situation changes, especially the change in the indicator status, the weight will be adjusted in real time to ensure that the initial model is accurate enough;

[0025] The said child node is the risk level of the enterprise. The risk level is inferred by comprehensively considering the indicator situations of the above five parent nodes and is used to reflect the currently calculated dynamic risk score. The weight will be automatically adjusted in real time according to the change in the indicator status during the real-time analysis process.

[0026] According to the above technical solution, the said C dynamically generates the weight Wi through the Bayesian network and calculates the enterprise comprehensive risk score RSI based on this. The enterprise comprehensive risk score RSI is based on the latest data status of the five types of dynamic indicators in the current period;

[0027] The re-fitting of the dynamic weight and the calculation of the RSI score are run every 24 hours. By updating the weight and recalculating the RSI score of the enterprise, the accuracy and timeliness of the enterprise comprehensive risk score are ensured.

[0028] According to the above technical solution, the output layer converts the analysis result into actionable information through a combination of analysis and sorting, and outputs the evaluation structure of the enterprise dynamic risk, mainly including two contents:

[0029] One is the priority control list sorted in descending order of RSI;

[0030] The other is the risk heat map in the form of GIS visualization;

[0031] The priority control list sorted in descending order of RSI is based on the RSI scores, arranging each enterprise in descending order from high to low to generate the enterprise list for priority control. The enterprise list sorted in descending order of RSI helps the park management department identify the enterprises that need to be focused on and managed, and directly guides the park management department to take actions against high-risk enterprises;

[0032] The risk heat map in the form of GIS visualization is a visual display method using geographic information technology to show the risk distribution in different areas within the park, facilitating an intuitive understanding of the risk concentration areas within the park, providing an intuitive geographical distribution perspective, and facilitating the discovery and analysis of risk concentration in specific areas.

[0033] According to the above technical solution, the output layer further includes decision support. Specifically, based on the specified scoring criteria and the risk heat map, the park management platform can immediately make targeted management decisions, improving the efficiency and accuracy of problem handling.

[0034] Compared with the prior art, the beneficial effects of the present invention:

[0035] By integrating the static risk level and dynamic indicators through a Bayesian network, the weight is adaptively adjusted. At the same time, by integrating Internet of Things alarms, complaint texts, and hidden danger images into a unified evaluation framework, the fusion of multi-source heterogeneous data is achieved. Using dynamic weight adjustment and multi-source data fusion, the originally static list of major / larger risk enterprises is upgraded to a real-time dynamic ranking list, realizing precise resource allocation. Through the fusion of dynamic indicators and the adjustment of the Bayesian network weights, the traditional static classification is upgraded to a real-time sorted priority control list, providing an intelligent solution for the environmental risk management of the park;

[0036] Moreover, by using advanced AI technology to dynamically adjust the weight matrix and realizing visual display through GIS technology, the accuracy and timeliness of the park environmental management are effectively improved, helping enterprises and park managers better prevent and respond to environmental risks and ensuring environmental safety within the park. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0038] Figure 1 is the overall system architecture diagram of the present invention;

[0039] Figure 2 is the flowchart of data analysis of the present invention. Specific Embodiments

[0040] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0041] As Figure 1 shown, the present invention provides a technical solution, a dynamic ranking and priority control system for park environmental risks based on Bayesian networks. Through dynamic index fusion and Bayesian network weight adjustment, the traditional static grading is upgraded to a real-time ranking and priority control list, providing an intelligent solution for park environmental risk management;

[0042] The system includes a data acquisition layer, a data analysis layer and an output layer;

[0043] The data acquisition layer is the basis for park environmental risk assessment, responsible for obtaining real-time and historical data from different sources for real-time monitoring and early warning;

[0044] The data analysis layer uses Bayesian networks to construct a dynamic index weight matrix for calculating the dynamic risk score RSI of enterprises;

[0045] The output layer calculates the comprehensive risk score of enterprises through the weights generated by Bayesian networks. The comprehensive risk score is given by weighting all the indicators in the Bayesian network.

[0046] Based on the above technical solution, the data acquisition layer integrates an intelligent system alarm module, an environmental protection complaint platform, an accident database, and a hidden danger investigation module. Among them, the intelligent system alarm module is an IoT sensor, and the hidden danger investigation module includes hidden dangers reported by enterprises and hidden dangers identified by AI;

[0047] By using the intelligent system alarm module, the environmental protection complaint platform, the accident database, and the hidden danger investigation module to construct a comprehensive data acquisition network, and obtaining data sources through the data acquisition network, the change of environmental risks in the park can be obtained in real time, providing data support for subsequent risk assessment and analysis.

[0048] Based on the above technical solution, the intelligent system alarm module collects real-time data on the park environment through IoT sensors, specifically including gas concentration, liquid level, temperature and humidity, flow rate, and pressure, and triggers an alarm when the real-time environmental data in the park exceeds the safety threshold;

[0049] The environmental protection complaint platform collects complaint information from residents, enterprise employees, and enterprises regarding environmental pollution problems in the park, and reflects the actual environmental problems and safety risks through the complaint information;

[0050] The accident database records various pollution and work safety accidents that have occurred in the park for analyzing potential risk points;

[0051] The hidden danger investigation module investigates hidden dangers through means of enterprise reporting and AI recognition. Enterprise reporting is through enterprises reporting safety hidden dangers online, and AI recognition uses artificial intelligence technology to automatically analyze videos and pictures to identify potential risk hidden dangers and give automatic warnings.

[0052] As Figure 2 shown, based on the above technical solution, the data analysis layer includes the following implementation steps:

[0053] A. Dynamic index normalization processing;

[0054] B. Construct a Bayesian network;

[0055] C. Perform dynamic sorting according to rules.

[0056] Based on the above technical solution, A is to perform standardization processing on five key indicators: the number of alarms, the number of complaints, the number of accidents, the number of unrectified hidden dangers, and the rectification delay rate, ensuring that these indicator data can be quantitatively processed, facilitating subsequent weight calculation and risk score comparison, avoiding errors caused by large values of different units, and ensuring fairness.

[0057] Based on the above technical solution, B, the Bayesian network is a graphical model used to qualitatively and quantitatively describe uncertainty and probabilistic dependence relationships, and can more flexibly handle complex risk assessment and prediction problems;

[0058] The constructed Bayesian network includes parent nodes, a conditional probability table CPT, and child nodes. When constructing the Bayesian network, each type of dynamic indicator serves as a parent node, and the initial weight is set through the conditional probability table CPT obtained by training with historical data. The risk level serves as a child node, and the weight will be adjusted in real time as the status changes on various indicators.

[0059] Based on the above technical solution, the parent nodes include five types of dynamic indicators: the number of alarms, the number of complaints, the number of accidents, the number of unrectified hidden dangers, and the rectification delay rate;

[0060] The conditional probability table CPT trains the initial weights based on historical data. Each conditional probability table CPT is trained according to the historical data of its corresponding indicators. As the actual situation changes, especially the change of the indicator status, the weights will be adjusted in real time to ensure that the initial model is accurate enough;

[0061] The child node is the risk level of the enterprise, which is inferred by comprehensively considering the indicator situations of the above five parent nodes and is used to reflect the currently calculated dynamic risk score. The weights will be automatically adjusted in real time according to the change of the indicator status during the real-time analysis process.

[0062] Based on the above technical solution, C, dynamically generates weights Wi through the Bayesian network and calculates the comprehensive risk score RSI of the enterprise based on this. The comprehensive risk score RSI of the enterprise is based on the latest data status of five types of dynamic indicators in the current period;

[0063] The refitting of the dynamic weights and the calculation of the RSI score are run every 24 hours. By updating the weights and recalculating the RSI score of the enterprise, the accuracy and timeliness of the comprehensive risk score of the enterprise are ensured.

[0064] Based on the above technical solution, the output layer converts the analysis results into actionable information by combining analysis and sorting, and outputs the evaluation structure of the enterprise's dynamic risk, mainly including two contents:

[0065] One is the priority control list arranged in descending order of RSI;

[0066] The other is the risk heat map in the form of GIS visualization;

[0067] The priority control list arranged in descending order of RSI is based on the RSI score, arranging each enterprise in descending order from high to low to generate a list of enterprises for priority control. The list of enterprises arranged in descending order of RSI helps the park management department determine the enterprises that need to be focused on and managed, and directly guides the park management department to take actions against high-risk enterprises;

[0068] The risk heat map in the form of GIS visualization is a visual display method using geographic information technology to show the risk distribution in different areas of the park, which is convenient for intuitively understanding the risk aggregation areas in the park, providing an intuitive geographical distribution perspective and facilitating the discovery and analysis of risk aggregation situations in specific areas.

[0069] Based on the above technical solution, the output layer also includes decision support. Specifically, according to the specified scoring criteria and the risk heat map, the park management platform can make targeted management decisions immediately, improving the efficiency and accuracy of problem handling.

[0070] In the present specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other.

[0071] In this article, specific examples are used to elaborate on the principles and implementation manners of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation on the present invention.

Claims

1. A dynamic ranking system for priority control of park environmental risks based on Bayesian networks, characterized in that: By integrating dynamic indicators and adjusting the weights of Bayesian networks, the traditional static grading is upgraded to a real-time ranking and prioritized control list, providing an intelligent solution for environmental risk management in the park; The system includes a data acquisition layer, a data analysis layer, and an output layer; The data acquisition layer is the basis for environmental risk assessment in the park, responsible for obtaining real-time and historical data from different sources for real-time monitoring and early warning; The data analysis layer uses a Bayesian network to construct a dynamic indicator weight matrix for calculating the dynamic risk score RSI of enterprises; The output layer calculates the comprehensive risk score of enterprises through the weights generated by the Bayesian network. The comprehensive risk score is given by weighting all the indicators in the Bayesian network.

2. The dynamic ranking system for priority control of park environmental risks based on Bayesian network according to claim 1, wherein: The data acquisition layer integrates an intelligent system alarm module, an environmental protection complaint platform, an accident database, and a hidden danger investigation module. Among them, the intelligent system alarm module is an IoT sensor, and the hidden danger investigation module includes hidden dangers reported by enterprises and hidden dangers identified by AI; An integrated data acquisition network is constructed through the intelligent system alarm module, the environmental protection complaint platform, the accident database, and the hidden danger investigation module. By obtaining data sources through the data acquisition network, the changes in environmental risks in the park can be obtained in real time, providing data support for subsequent risk assessment and analysis.

3. The dynamic ranking system for priority control of park environment risks based on Bayesian network according to claim 2, characterized in that: The intelligent system alarm module collects real-time data on the park environment through IoT sensors, specifically including gas concentration, liquid level, temperature and humidity, flow rate, and pressure. When the real-time environmental data in the park exceeds the safety threshold, an alarm is triggered; The environmental protection complaint platform collects complaint information from residents, enterprise employees, and enterprises regarding environmental pollution problems in the park, and reflects actual environmental problems and safety risks through the complaint information; The accident database records various pollution and work safety accidents that have occurred in the park for analyzing potential risk points; The hidden danger investigation module investigates hidden dangers through the means of enterprise reporting and AI identification. Enterprise reporting is through enterprises reporting safety hidden dangers online, and AI identification uses artificial intelligence AI technology to automatically analyze videos and pictures, identify potential risk hidden dangers, and automatically give early warnings.

4. The dynamic ranking system for priority control of park environment risks based on Bayesian network according to claim 1, characterized in that: The data analysis layer includes the following implementation steps: A. Normalize dynamic indicators; B. Construct a Bayesian network; C. Perform dynamic ranking according to rules.

5. The dynamic ranking system for priority control of park environmental risks based on Bayesian network according to claim 4, wherein: In step A, standardize the five key indicators: the number of alarms, the number of complaints, the number of accidents, the number of unrectified hidden dangers, and the rectification delay rate, ensuring that these indicator data can be quantified for subsequent weight calculation and risk score comparison.

6. The dynamic ranking system for priority control of park environment risks based on Bayesian network according to claim 4, characterized in that: In step B, the Bayesian network is a graphical model used to qualitatively and quantitatively describe uncertainty and probabilistic dependence relationships, which can more flexibly handle complex risk assessment and prediction problems; The constructed Bayesian network includes parent nodes, a conditional probability table CPT, and child nodes. When constructing the Bayesian network, each type of dynamic indicator serves as a parent node, and the initial weights are set through the conditional probability table CPT obtained by training with historical data. The risk level serves as a child node, and the weights will be adjusted in real time as the status changes on various indicators.

7. The dynamic ranking system for priority control of park environment risks based on Bayesian network according to claim 6, characterized in that: The parent node includes five dynamic indicators: the number of alarms, the number of complaints, the number of accidents, the number of unrectified hidden dangers, and the rectification delay rate; The conditional probability table CPT is trained with initial weights based on historical data. Each conditional probability table CPT is trained according to the historical data of its corresponding indicator. As the actual situation changes, especially the change of the indicator status, the weights will be adjusted in real time to ensure that the initial model is accurate enough; The child node is the risk level of the enterprise. The risk level is inferred by comprehensively considering the above five parent node indicators and is used to reflect the currently calculated dynamic risk score. The weights will be automatically adjusted in real time according to the change of the indicator status during the real-time analysis process.

8. The dynamic ranking system for priority control of park environment risks based on Bayesian network according to claim 4, characterized in that: In C, the weights Wi are dynamically generated through the Bayesian network, and based on this, the comprehensive risk score RSI of the enterprise is calculated. The comprehensive risk score RSI of the enterprise is based on the latest data status of the five dynamic indicators in the current period; The re-fitting of the dynamic weights and the calculation of the RSI score are run every 24 hours. By updating the weights, the RSI score of the enterprise is recalculated.

9. The dynamic ranking system for priority control of park environment risks based on Bayesian network according to claim 1, wherein: The output layer converts the analysis results into actionable information through a combination of analysis and sorting, and outputs the evaluation structure of the enterprise's dynamic risk, which mainly includes two contents: One is the priority control list sorted in descending order of RSI; The other is the risk heat map in the form of GIS visualization; The priority control list sorted in descending order of RSI is based on the RSI score. The enterprises are sorted from high to low in descending order to generate a list of enterprises that need to be prioritized for control. The list of enterprises sorted in descending order of RSI helps the park management department determine the enterprises that need to be focused on and managed, and directly guides the park management department to take actions against high-risk enterprises; The risk heat map in the form of GIS visualization is a visual display method using geographic information technology to show the risk distribution in different areas of the park, which is convenient for intuitively understanding the risk aggregation areas in the park, provides an intuitive geographical distribution perspective, and is convenient for discovering and analyzing the risk aggregation in specific areas.

10. The dynamic ranking system for priority control of park environment risks based on Bayesian network according to claim 9, characterized in that: The output layer also includes decision support. Specifically, according to the specified scoring criteria and the risk heat map, the park management platform can immediately make targeted management decisions.

Images