摘要： 选定一种舰用环氧涂料为研究对象,采用MATLAB线性最小二乘法、非线性最小二乘法和遗传算法建立乙醛和甲苯的气体释放模型,并对其拟合效果进行比较。结果表明:遗传算法可以将拟合的误差精度控制在-7.87%～7.69%,是最为准确的拟合方法。研究得到的数学模型可以预测任意时间内有害气体的释放浓度,节省检测时间,也可作为评价非金属材料毒性和进行非金属材料筛选的辅助手段。

关键词: 遗传算法, 涂料, 气体, 乙醛, 甲苯

Abstract: An epoxy coating used in vessels was chosen-as the research object. The vapour emission models of aldehyde and toluene were established by - use of MATLAB linear least square method, non-linear least square method and genetic algorithm, respectively. The fitted results of three models were compared. The results showed that the error precision of genetic algorithm was -7. 87%~7.69% and it was the most precised fitting method. The mathematical model could be used to calculate the emmitted vapour concentration at any periods of time. And it could save the testing time, as well as be used as a supplementary means to assess the toxicity of non-metal materials and selection of non-metal materials.

Key words: genetic algorithm, coating, vapour, aldehyde, toluene