港口能源储油罐区可燃气体浓度预测建模  

Modeling of Combustible Gas Concentration Prediction in Port Energy Storage Tank

作　　者：姜建军 胡候林 金晓宸[3] 万勇 钟碧良[1] JIANG Jian-jun;HU Hou-lin;JIN Xiao-chen;WANG Yong;ZHONG Bi-liang(School of Ocean Engineering,Guangzhou Maritime University,Guangzhou Guangdong 510725,China;Guangzhou Institute of Energy Testing,Guangzhou Guangdong 511447,China;School of Maritime Law and Traffic Management,Guangzhou Maritime University,Guangzhou Guangdong 510725,China)

机构地区：[1]广州航海学院海洋装备工程学院,广东广州510725 [2]广州能源检测研究院,广东广州511447 [3]广州航海学院海事法律与交通管理学院,广东广州510725

出　　处：《广州航海学院学报》2024年第3期69-73,共5页Journal of Guangzhou Maritime University

基　　金：广州市市场监督管理局科技项目(2024KJ32)。

摘　　要：原油在储运过程中容易挥发、产生可燃气体,可燃气体浓度过高会有爆炸、中毒、使人窒息的危险,因此,预测和实时监控罐区可燃气体浓度具有重要意义。本文研究基于BP神经网络和遗传算法优化的BP神经网络的可燃气体浓度模型,并进行预测和对比分析,通过LabVIEW软件搭建监控系统,利用MATLAB script节点调用可燃气体浓度最优的模型,实现储油罐区可燃气体浓度的预测和监控预警功能。结果表明,GA-BP神经网络模型拟合度最高,更接近真实值的变化趋势。In the process of storage and transportation,crude oil was easy to volatilize and produce combustible gas.If the concentration of combustible gas was too high,it may cause explosion,poisoning and asphyxiation.Therefore,it was of great significance to realize real-time monitoring of combustible gas concentration in tank farm.This paper studied the combustible gas concentration model of BP neural network optimized by BP neural network and genetic algorithm,and makes prediction and comparative analysis.The monitoring system built by LabVIEW software used MATLAB script node to invoke the optimal model of combustible gas concentration to realize the prediction,monitoring and early warning function of combustible gas concentration in the oil storage tank area.The results show that the GA-BP neural network model has the highest fitting degree and is closer to the trend of the real value.

关 键 词：可燃气体 BP神经网络 遗传算法 港口能源 LABVIEW 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]