Nowadays, chemical safety has attracted considerable attention, and chemical gas leakage monitoring and source term estimation(STE) have become hot spots. However, few studies have focused on sensor layouts in scenari...Nowadays, chemical safety has attracted considerable attention, and chemical gas leakage monitoring and source term estimation(STE) have become hot spots. However, few studies have focused on sensor layouts in scenarios with multiple potential leakage sources and wind conditions, and studies on the risk information(RI) detection and prioritization order of sensors have not been performed. In this work, the monitoring area of a chemical factory is divided into multiple rectangles with a uniform mesh. The RI value of each grid node is calculated on the basis of the occurrence probability and normalized concentrations of each leakage scenario. A high RI value indicates that a sensor at a grid node has a high chance of detecting gas concentrations in different leakage scenarios. This situation is beneficial for leakage monitoring and STE. The methods of similarity redundancy detection and the maximization of sensor RI detection are applied to determine the sequence of sensor locations. This study reveals that the RI detection of the optimal sensor layout with eight sensors exceeds that of the typical layout with 12 sensors. In addition, STE with the optimized placement sequence of the sensor layout is numerically simulated. The statistical results of each scenario with various numbers of sensors reveal that STE is affected by sensor number and scenarios(leakage locations and winds). In most scenarios, appropriate STE results can be retained under the optimal sensor layout even with four sensors. Eight or more sensors are advised to improve the performance of STE in all scenarios. Moreover, the reliability of the STE results in each scenario can be known in advance with a specific number of sensors. Such information thus provides a reference for emergency rescue.展开更多
The leakage of hazardous gases poses a significant threat to public security and causes environmental damage.The effective and accurate source term estimation(STE)is necessary when a leakage accident occurs.However,mo...The leakage of hazardous gases poses a significant threat to public security and causes environmental damage.The effective and accurate source term estimation(STE)is necessary when a leakage accident occurs.However,most research generally assumes that no obstacles exist near the leak source,which is inappropriate in practical applications.To solve this problem,we propose two different frameworks to emphasize STE with obstacles based on artificial neural network(ANN)and convolutional neural network(CNN).Firstly,we build a CFD model to simulate the gas diffusion in obstacle scenarios and construct a benchmark dataset.Secondly,we define the structure of ANN by searching,then predict the concentration distribution of gas using the searched model,and optimize source term parameters by particle swarm optimization(PSO)with well-performed cost functions.Thirdly,we propose a one-step STE method based on CNN,which establishes a link between the concentration distribution and the location of obstacles.Finally,we propose a novel data processing method to process sensor data,which maps the concentration information into feature channels.The comprehensive experiments illustrate the performance and efficiency of the proposed methods.展开更多
基金supported by National Natural Science Foundation of China (61988101)National Natural Science Fund for Distinguished Young Scholars (61725301)Fundamental Research Funds for the Central Universities。
文摘Nowadays, chemical safety has attracted considerable attention, and chemical gas leakage monitoring and source term estimation(STE) have become hot spots. However, few studies have focused on sensor layouts in scenarios with multiple potential leakage sources and wind conditions, and studies on the risk information(RI) detection and prioritization order of sensors have not been performed. In this work, the monitoring area of a chemical factory is divided into multiple rectangles with a uniform mesh. The RI value of each grid node is calculated on the basis of the occurrence probability and normalized concentrations of each leakage scenario. A high RI value indicates that a sensor at a grid node has a high chance of detecting gas concentrations in different leakage scenarios. This situation is beneficial for leakage monitoring and STE. The methods of similarity redundancy detection and the maximization of sensor RI detection are applied to determine the sequence of sensor locations. This study reveals that the RI detection of the optimal sensor layout with eight sensors exceeds that of the typical layout with 12 sensors. In addition, STE with the optimized placement sequence of the sensor layout is numerically simulated. The statistical results of each scenario with various numbers of sensors reveal that STE is affected by sensor number and scenarios(leakage locations and winds). In most scenarios, appropriate STE results can be retained under the optimal sensor layout even with four sensors. Eight or more sensors are advised to improve the performance of STE in all scenarios. Moreover, the reliability of the STE results in each scenario can be known in advance with a specific number of sensors. Such information thus provides a reference for emergency rescue.
基金The work was supported by the National Natural Science Foundation of China(Basic Science Center Program:6198810121706069),Natural Science Foundation of Shanghai(17ZR1406800)National Science Fund for Distinguished Young Scholars(61725301).
文摘The leakage of hazardous gases poses a significant threat to public security and causes environmental damage.The effective and accurate source term estimation(STE)is necessary when a leakage accident occurs.However,most research generally assumes that no obstacles exist near the leak source,which is inappropriate in practical applications.To solve this problem,we propose two different frameworks to emphasize STE with obstacles based on artificial neural network(ANN)and convolutional neural network(CNN).Firstly,we build a CFD model to simulate the gas diffusion in obstacle scenarios and construct a benchmark dataset.Secondly,we define the structure of ANN by searching,then predict the concentration distribution of gas using the searched model,and optimize source term parameters by particle swarm optimization(PSO)with well-performed cost functions.Thirdly,we propose a one-step STE method based on CNN,which establishes a link between the concentration distribution and the location of obstacles.Finally,we propose a novel data processing method to process sensor data,which maps the concentration information into feature channels.The comprehensive experiments illustrate the performance and efficiency of the proposed methods.