CN-121981531-A - Toxin agent early warning monitoring management system

CN121981531ACN 121981531 ACN121981531 ACN 121981531ACN-121981531-A

Abstract

The invention discloses a toxic agent early warning monitoring management system, which belongs to the technical field of toxic agent monitoring and comprises a data acquisition module, a data processing module, a data analysis module, an early warning module and a warning module, wherein the data acquisition module is used for acquiring real-time toxic agent monitoring data output by toxic agent monitoring equipment, the data processing module is used for preprocessing the real-time toxic agent monitoring data, classifying and integrating the preprocessed data and transmitting the preprocessed data to the data analysis module based on a wireless protocol mode, the data analysis module is used for inputting the received preprocessed data into a pre-trained risk assessment model for analysis and determining a risk level, the early warning module is used for sending an early warning prompt when the risk level is greater than or equal to a preset risk level threshold value, and analyzing the toxic agent monitoring data based on the risk assessment model, so that the risk level of the toxic agent can be accurately determined, and scientific basis is provided for subsequent decisions.

Inventors

WANG YANG
YANG ZE
YANG SHUO
LI ZHUANG
ZHANG YONGFEI
HE DAPENG
ZHANG LIANSHENG

Assignees

营口世纪电子仪器股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260107

Claims (9)

1. A toxicant pre-warning monitoring management system, comprising: The data acquisition module is used for acquiring real-time toxin agent monitoring data output by the toxin agent monitoring equipment; The data processing module is used for preprocessing the real-time toxin agent monitoring data, classifying and integrating the preprocessed data and transmitting the preprocessed data to the data analysis module based on a wireless protocol mode; The data analysis module is used for inputting the received preprocessed data into a pre-trained risk assessment model for analysis and determining a risk level; and the early warning module is used for sending out early warning prompt when the risk level is determined to be greater than or equal to a preset risk level threshold value.
2. The toxicant pre-warning monitoring management system of claim 1, wherein said data processing module comprises: the pretreatment sub-module is used for comparing, warehousing, digitizing and integrating analysis processing on the real-time toxin agent monitoring data, wherein the integrating analysis processing comprises integrity analysis, validity analysis, trend analysis and abnormality analysis; And the data transmission sub-module is used for transmitting the preprocessed data to the data analysis module based on a wireless protocol mode.
3. The toxicant pre-warning monitoring management system of claim 2, wherein the pre-processing sub-module comprises: the data verification unit is used for carrying out analysis, statistics, query and visualization operation on the real-time toxin agent monitoring data, and eliminating interference abnormal data through verifying the real-time toxin agent monitoring data; The marking unit is used for acquiring real-time toxin agent monitoring data after interference abnormality is eliminated, dividing the real-time toxin agent monitoring data according to the attribute of the real-time toxin agent monitoring data, and obtaining a real-time toxin agent monitoring data distribution diagram; the instruction determining unit is used for determining a plurality of different data integration rules for the real-time toxin agent monitoring data distribution diagram according to a plurality of data integration requirements, and establishing a dynamic data integration instruction based on the different data integration rules; and the data integration unit is used for dynamically integrating the verified real-time toxin agent monitoring data based on the dynamic data integration instruction to obtain preprocessed data.
4. The toxicant pre-warning monitoring management system of claim 3, wherein said instruction determination unit comprises: The acquisition subunit is used for acquiring a multisource data integration requirement set and a real-time toxin agent monitoring data distribution diagram basic attribute, wherein the multisource data integration requirement set comprises a rigid constraint condition and a flexible preference condition, the rigid constraint condition comprises a real-time threshold value, a precision requirement and a data integrity requirement of toxin agent monitoring data, the flexible preference condition comprises a scene priority preference and a data dimension preference, and the real-time toxin agent monitoring data distribution diagram basic attribute comprises a spatial resolution, a time update frequency and a data state characteristic; The system comprises a decomposition subunit, a concentration dimension quantization index, a decomposition subunit, a processing subunit and a processing unit, wherein the decomposition subunit is used for carrying out hierarchical analysis and quantization mapping on the multi-source data integration requirement set through a built-in requirement analysis engine, decomposing the multi-source data integration requirement set into three core dimension quantization indexes of space, time and concentration, and marking priority weights of all dimension indexes; The system comprises a construction subunit, a data integration rule, a concentration integration rule and a concentration integration rule, wherein the construction subunit is used for constructing a plurality of different data integration rules based on the analyzed quantization index and the priority weight, the data integration rule comprises a basic integration rule and the composite integration rule, the basic integration rule comprises a space integration rule, a time integration rule and the concentration integration rule, the space integration rule comprises a coarse grid aggregation rule, a fine grid retention rule and an edge sharpening rule, the time integration rule comprises a sliding window average rule, a real-time sampling non-integration rule and a historical data fitting rule, the concentration integration rule comprises a hierarchical threshold clustering rule and an abnormal value removing rule, and the composite integration rule is based on the dynamic combination of the priority weight; The generation subunit is configured to generate a dynamic data integration instruction carrying a trigger condition, an execution parameter, a switching logic and a termination condition based on the constructed data integration rule set, where the trigger condition is a scene threshold value enabled by the rule, the execution parameter is a quantization parameter corresponding to the integration rule, the switching logic is a switching trigger condition between different rules, and the termination condition is a stop threshold value executed by the rule.
5. The toxicant pre-warning monitoring management system of claim 1, wherein said data analysis module comprises: The data receiving sub-module is used for establishing a data transmission channel with each toxic agent monitoring device and receiving real-time toxic agent monitoring data and preprocessed data; The data fitting sub-module is used for storing the received real-time toxin agent monitoring data into the storage chip, fitting the real-time toxin agent monitoring data and analyzing the toxin agent level trend to obtain real-time toxin agent monitoring fitting data and toxin agent level trend data; The data analysis sub-module is used for inputting real-time toxin agent monitoring fitting data and toxin agent level trend data into a pre-trained risk assessment model for analysis, and determining risk levels.
6. The toxicant pre-warning monitoring management system of claim 5, wherein said data receiving sub-module comprises: A data receiving unit for receiving the poison monitoring data generated by each poison monitoring apparatus in a unit time, wherein the unit time is 50ms; the gas characteristic coefficient acquisition module is used for calculating a gas characteristic coefficient corresponding to the toxic agent monitoring device based on the toxic agent monitoring data, wherein the gas characteristic coefficient is acquired through the following formula: ; Wherein C represents a gas characteristic coefficient, Indicating real-time monitoring of air pressure; Represents standard atmospheric pressure; Representing a real-time temperature difference; Representing a reference temperature; 、 is a calibration constant.
7. The toxicant pre-warning monitoring management system of claim 5, wherein the method for constructing the risk assessment model comprises: acquiring a training data set, wherein the training data set comprises historical toxic agent monitoring data, temperature and humidity data, positioning information, toxic agent attribute data and historical risk tag data, and the historical toxic agent monitoring data comprises toxic gas concentration data and toxic agent level trend data; Extracting features of the training data set to obtain multi-mode training features, wherein the multi-mode training features comprise static features and dynamic features; Fusing the static features and the dynamic features, and mapping all features to a [0,1] interval by adopting Min-Max standardization to obtain standardized multi-mode training features; Building a risk assessment model, wherein the risk assessment model comprises a multi-mode feature encoder, a layered attention fusion layer, a dynamic risk calculation module and a dynamic threshold calibration module, and the risk assessment model comprises the following components: The multi-mode feature encoder comprises a numerical feature branch, a time sequence feature branch and a space feature branch, wherein the numerical feature branch is a 2-layer full-connection network, an activation function is a ReLU, and a 12-dimensional feature vector is converted into a 64-dimensional feature vector; the time sequence feature branch is a 1D-CNN+LSTM combined structure, and converts the time sequence related feature into a 32-dimensional feature vector, the space feature branch is a GraphConv-diagram convolution network, converts the space related feature into a 32-dimensional feature vector, and splices three branches to output to obtain a 128-dimensional fusion feature; The hierarchical attention fusion layer comprises feature type attention and feature internal attention, feature type weights are learned through a single-layer perceptron, a feature internal local weight matrix is calculated through a self-attention mechanism, and 128-dimensional fusion features are subjected to weighted fusion based on the weight matrix to obtain a cross-modal fusion feature vector; The dynamic risk calculation module calculates a static risk score, a dynamic risk score and a diffusion association correction score based on the cross-modal fusion feature vector, and calculates a preliminary risk value based on the three components; The dynamic threshold calibration module calculates an environment calibration coefficient and a dynamic threshold interval based on the environment parameters, and calculates a final risk assessment value based on the preliminary risk value and the dynamic threshold interval; And inputting the standardized multi-mode training characteristics into the risk assessment model for training, constructing a total loss function, adopting AdamW optimizers, carrying out model training by combining with L2 regularization and early-stop strategies, and obtaining the risk assessment model when the training result meets the requirements.
8. The toxicant pre-warning monitoring management system of claim 1, further comprising a protection module configured to query a protection policy database based on the risk level, determine a target protection policy, and perform security protection based on the target protection policy.
9. The toxicant pre-warning monitoring management system of claim 8, wherein the method for constructing the protection policy database comprises: The method comprises the steps of carrying out hierarchical coding on toxin agent characteristics by adopting a level coding rule based on four dimensions of the chemical structure, the toxicity level, the volatility characteristic and the diffusion characteristic of the toxin agent, and constructing a scene attribute set by interval division and combination rules from three dimensions of a space range, crowd distribution and environmental conditions, wherein the toxin agent characteristics are uniquely associated with the basic attribute set by adopting the level coding rule; Acquiring four types of dynamic rule dimensions of time attenuation, environment change, resource state and crowd movement, constructing a main rule-sub rule-trigger threshold three-layer nested structure, associating each basic strategy with a dynamic rule group, binding the dynamic rule group with the basic strategy in a three-dimensional strategy base through a strategy ID, and establishing a strategy-rule association table, wherein the association table is stored according to a rule ID-strategy ID-main rule condition-sub rule logic-trigger threshold format, constructing a secondary index taking a strategy ID+rule type as a core, and completing the construction of a dynamic rule engine; Defining three conflict types of resource conflict, operation conflict and effect conflict, and constructing a policy combination-conflict type-conflict threshold-coordination priority-coordination rule five-dimensional conflict coordination matrix, wherein the matrix only pre-calculates conflict for policies with risk level more than or equal to level III; when a plurality of strategies are called simultaneously, traversing the strategy combination corresponding to the strategy coordination matrix matching and the conflict type, judging as conflict when a conflict threshold is met, executing conflict solution according to the coordination priority and the coordination rule in the matrix, storing the coordinated strategy combination into a dynamic adaptation library, and recording conflict coordination records to complete the construction of the conflict coordination matrix; And integrating based on the three-dimensional strategy base, the dynamic rule engine and the conflict coordination matrix to obtain a protection strategy database.

Description

Toxin agent early warning monitoring management system Technical Field The invention relates to the technical field of toxin agent monitoring, in particular to a toxin agent early warning monitoring management system. Background In the fields of chemical production, dangerous chemical storage and transportation, emergency rescue and the like, leakage or artificial release of a toxic agent can cause serious safety accidents, and serious threat is caused to personnel life health, ecological environment and social public order. Therefore, the real-time and accurate monitoring and the timely early warning of the toxic agent are key links for preventing the related risks of chemical decomposition and guaranteeing public safety. In the prior art, most monitoring equipment is in a single-point independent working mode, the data acquisition range is limited and the compatibility is poor, synchronous acquisition of multi-type and multi-point poison monitoring data is difficult to realize, meanwhile, part of equipment has low acquisition frequency, instantaneous change of the concentration of the poison cannot be captured, so that the integrity and timeliness of real-time monitoring data are insufficient, hidden dangers are buried for subsequent risk judgment, secondly, the existing monitoring system is used for judging the risk level in a simple threshold comparison mode, an accurate risk assessment model is built without combining multi-dimensional information such as the poison type, concentration change trend, environmental factors (such as temperature, humidity and wind speed), and the like, so that the risk level judgment subjectivity is strong, the accuracy is low, false early warning or leakage early warning phenomenon is easy to occur, and the actual risk degree cannot be accurately reflected. Therefore, a toxic agent early warning monitoring management system is needed to solve the above problems. Disclosure of Invention The invention aims to provide a toxic agent early warning monitoring management system, which is used for analyzing toxic agent monitoring data based on a risk assessment model, accurately determining the risk level of a toxic agent and providing scientific basis for subsequent decision making. In order to achieve the above purpose, the present invention provides the following technical solutions: a toxicant pre-warning monitoring management system, comprising: The data acquisition module is used for acquiring real-time toxin agent monitoring data output by the toxin agent monitoring equipment; The data processing module is used for preprocessing the real-time toxin agent monitoring data, classifying and integrating the preprocessed data and transmitting the preprocessed data to the data analysis module based on a wireless protocol mode; The data analysis module is used for inputting the received preprocessed data into a pre-trained risk assessment model for analysis and determining a risk level; and the early warning module is used for sending out early warning prompt when the risk level is determined to be greater than or equal to a preset risk level threshold value. Preferably, the data processing module includes: the pretreatment sub-module is used for comparing, warehousing, digitizing and integrating analysis processing on the real-time toxin agent monitoring data, wherein the integrating analysis processing comprises integrity analysis, validity analysis, trend analysis and abnormality analysis; And the data transmission sub-module is used for transmitting the preprocessed data to the data analysis module based on a wireless protocol mode. Preferably, the preprocessing sub-module includes: the data verification unit is used for carrying out analysis, statistics, query and visualization operation on the real-time toxin agent monitoring data, and eliminating interference abnormal data through verifying the real-time toxin agent monitoring data; The marking unit is used for acquiring real-time toxin agent monitoring data after interference abnormality is eliminated, dividing the real-time toxin agent monitoring data according to the attribute of the real-time toxin agent monitoring data, and obtaining a real-time toxin agent monitoring data distribution diagram; the instruction determining unit is used for determining a plurality of different data integration rules for the real-time toxin agent monitoring data distribution diagram according to a plurality of data integration requirements, and establishing a dynamic data integration instruction based on the different data integration rules; and the data integration unit is used for dynamically integrating the verified real-time toxin agent monitoring data based on the dynamic data integration instruction to obtain preprocessed data. Preferably, the instruction determination unit includes: The acquisition subunit is used for acquiring a multisource data integration requirement set and a real-time toxin agent monitoring data distribution d