论文部分内容阅读
为了解决利用声发射在线检测技术对储罐底板腐蚀状态进行评价时,主要依赖检测人员经验的问题,使该项技术能更好地推广和应用,利用储罐底板在线检测的声发射信息和外观检查信息,并根据相关标准及专家经验,确定与储罐底板腐蚀状态相关的表征因素。采用遗传算法(GA)改进贝叶斯网络(BN)搜索方法,建立基于GA的BN智能评价方法。针对声发射在线检测信息和外观检查信息,分别建立基于标准用声发射因素、基于声发射因素和综合考虑声发射因素和外观检查因素的基于在线检测信息的储罐底板腐蚀状态评价模型。通过对测试样本的评价,对比声发射检测专家评价结果,其中基于在线检测信息的储罐底板腐蚀状态评价模型的准确率为96%,该模型能够对储罐底板腐蚀状态进行可靠的智能评价。
In order to solve the problem of using the on-line acoustic emission detection technology to evaluate the corrosion status of the tank floor, the main reliance on the experience of the inspectors makes the technology better popularized and applied. The acoustic emission information and appearance Check the information and, based on relevant standards and expert experience, determine the characterization factors associated with tank floor corrosion status. The genetic algorithm (GA) is used to improve the Bayesian Network (BN) search method, and a GA-based BN intelligent evaluation method is established. According to the on-line detection information of acoustic emission and the visual inspection information, the evaluation models of corrosion status of tank bottom based on the on-line detection information based on the standard acoustic emission factors, the acoustic emission factors and the comprehensive consideration of the acoustic emission factors and the appearance inspection factors are respectively established. Through the evaluation of test samples, the results of expert evaluation of acoustic emission were contrasted. The accuracy of tank corrosion evaluation model was 96% based on the on-line detection information. The model was able to evaluate the corrosion status of tank bottom reliably and intelligently.