Study On Multi-objective Optimization Method Of Gas Detectors Placement Considering Reliability

Oil refining enterprises prone to dangerous gas leakage accident,and caused great damage to the staff and the surrounding environment.The optimal placement of gas detectors plays an important role in ensuring the safety of refinery staff and reducing the loss of accidents,but at present,the study on optimization of dangerous gas leak detection in China is still weak.The relevant standards introduced only specify the overall requirements for the gas detector placement,but failed to carry out more scientific and meticulous research from the perspective of mathematical planning theory.This study depends on the basic scientific research business central university funds special funds to finance the project(13CX02079A)“Study on Optimization Method of Gas Detection Alarming Devices Layout for Chemical Equipment” and Shandong Province Natural Science Fund Project(2016ZRE28163)“Study on Optimization of Gas Leak Detection with Multi-random Variables for Refinery”.And the study on Optimization Model and Algorithm of Gas Detector placement,providing theoretical and technical support for the improvement of production safety and risk management in refining device.The specific research contents are as follows:1.Identification of Potential Leakage Sources of Hydrogen Sulfide Gas Based on Gray Relational AnalysisAccording to the nature and process flow of diesel hydrogenation unit,the potential leakage source location of hydrogen sulfide gas is determined.Using the gray relational analysis method in gray system theory,the possibility of occurrence of hydrogen sulfide leakage accident,the frequency of exposure to dangerous environment and the consequences of accidents severity as the main identification index of the potential leakage source,calculate the correlation degree between the potential leakage source and the optimal identification index,so as to sort the risk of each potential leakage source,lay the foundation for the subsequent optimization of gas detector placement.2.Leakage Scenario Reduction Based on K-means Clustering AlgorithmThe set of the wind field and the leakage source are used as elements to construct the hydrogen sulfide gas leak scene of the diesel hydrogenation unit.The K-means clusteringalgorithm is used to cluster the differences of the factors that make up the leakage scene,and the representative scenes in each class after clustering are selected.These representative scenes are formed into a new set of leakage scenes with limited discrete scenes.To effectively quantify all the scenes,so as to achieve the purpose of reducing the leakage scene and improving the efficiency of the leakage scene,and to provide a strong basis for improving the efficiency of the gas detector layout optimization.3.Construction of Optimization Mathematical Model for Gas Detector placement Considering Reliability and RobustnessThe evaluation system is established with the timeliness and robustness as the core to take into account the reliability of the gas detector.Probability-weighted detection time expectations reflect the timeliness to meet the standard additional detection change due to the preferred detector failure,and the number of values reflect the robustness,and the objective vector composed of two objective functions is used to describe the target vector of the location decision.The mathematical model of the multi-target location optimization of the gas detector under the failure scenario is established.4.Propose the NSGA-II algorithm and the multi-objective particle swarm optimization algorithm based on simulated annealingThe optimization model of the layout optimization mathematical model is solved by NSGA-II algorithm and multi-objective particle swarm optimization algorithm based on simulated annealing.The optimization scheme of gas detector placement is obtained.The scheme is sorted by ideal point approximation method(TOPSIS).From all the schemes,a number of options are preferred,and the final arrangement is determined from the preferred scheme according to the different needs of the decision maker.Finally,the two algorithms are compared and the validity of the leakage scene reduction method is verified.