一种基于遗传算法的红外无损检测缺陷的新方法  

A Novel Defect Evaluation Method in Infrared NDT Based on Genetic Algorithm

作　　者：梅林[1] 吴立德[1] 王裕文[2] 

机构地区：[1]复旦大学计算机科学与工程系,上海200433 [2]西安交通大学机械工程学院,西安710049

出　　处：《光学学报》2002年第12期1452-1456,共5页Acta Optica Sinica

摘　　要：红外无损检测的本质可归为热传导反问题的求解。利用通过解析法求解热传导正问题得出的结论,提出基于遗传算法优化求解的红外无损检测缺陷信息新方法,该算法采用浮点数编码,并采用多遗传算子并行操作,不仅获得了好的结果,而且维系了种群的多样性。模拟实验验证了算法的鲁棒性和可靠性。The essence of infrared nondestructive testing is to find a solution to inverse problem of heat conduction. A novel solving method to this problem based on genetic algorithm was proposed. It uses floating point coding technique and multi-operators parallel style to improve the precision of the result and to maintain the variety of the group. The simulation experimental results prove that this method behaves a strong robustness and reliability.

关 键 词：缺陷 红外线无损检测 遗传算法 热传导反问题 

分 类 号：TN215[电子电信—物理电子学] O551.3[理学—热学与物质分子运动论]