基于支持向量机起重机载荷谱获取方法的研究
发布时间：2018-08-27 17:58
【摘要】：近年来，随着国家经济的突飞猛进及对基础设施的大力投资，使得重大技术装备行业蓬勃发展，而起重机械在重大技术装备行业中占有很大比重。由于起重机械大部分是重型设备，一旦发生事故，经济损失惨重，极易造成人员伤亡。对此国家对起重机的使用安全提出了很高要求，并把其列为国家特种设备。通过对起重机事故的调查发现，其主要原因是疲劳断裂。为此，各国政府和科研机构开始对起重机械疲劳断裂问题进行大量研究，可望在设备疲劳断裂之前能够报告给用户，从而避免事故的发生。 本文通过对金属疲劳断裂相关理论知识进行深入的研究，得出解决起重机疲劳断裂的先决条件是编制出能模拟起重机金属结构真实使用情况，具有代表性的典型载荷——时间历程，即载荷谱。由于载荷的随机性和不确定性，导致无法将实测结果直接应用于理论分析与工程实践，而需构建一种能本质反映起重机金属结构在各种工况下所受载荷随时间变化的当量载荷谱。 本文以铸造桥式起重机和通用桥式起重机为研究对象，通过现场调研，收集了部分起重机工作状况的数据样本。并首次使用基于统计学习理论的支持向量机非线性回归理论，通过对收集的样本数据进行训练，建立了相应类型桥式起重机工作循环次数与不同起升载荷之间的非线性映射关系，利用此映射关系，可实现对相应类型或未知桥式起重机当量载荷谱的预测。本文利用可视化程序设计语言VC++6.0编制了基于支持向量机起重机载荷谱获取与预测的应用软件，将软件应用于工程实例，并用软件预测结果与实际结果进行比较，表明具有较高的吻合性和实用性。然后，将此方法与最小二乘法和神经网络两种方法进行比较，利用同一样本数据进行起重机载荷谱的获取与预测，证明了支持向量机方法的优越性。而且，本软件操作简单明了，使得操作人员不需要具备支持向量机的相关知识，就可以使用本软件。最重要的是本研究成果为后续起重机疲劳寿命预测软件的开发奠定了基础。
[Abstract]:In recent years, with the rapid development of national economy and the great investment in infrastructure, the major technical equipment industry is booming, and the lifting machinery occupies a large proportion in the major technical equipment industry. Because the lifting machinery is mostly heavy equipment, once the accident occurs, the economic loss is heavy, and it is easy to cause casualties. This country put forward the very high request to the crane safe use, and listed it as the national special equipment. Through the investigation of crane accident, it is found that the main reason is fatigue fracture. Therefore, many governments and scientific research institutions have begun to study the fatigue fracture of lifting machinery, which can be reported to the users before the fatigue fracture of the equipment, thus avoiding the occurrence of accidents. In this paper, through the deep research on the theory of metal fatigue fracture, it is concluded that the precondition to solve the fatigue fracture of crane is to draw up a program to simulate the real use of crane metal structure. Typical load-time history, namely load spectrum. Due to the randomness and uncertainty of load, it is impossible to directly apply the measured results to theoretical analysis and engineering practice. It is necessary to construct an equivalent load spectrum which can essentially reflect the variation of load on crane metal structure with time under various working conditions. In this paper, casting bridge crane and general bridge crane are taken as research objects, and some data samples of crane working condition are collected through field investigation. The support vector machine (SVM) nonlinear regression theory based on statistical learning theory is used for the first time to train the collected sample data. The nonlinear mapping relationship between the working cycle number of the corresponding type bridge crane and different lifting loads is established. Using this mapping relationship, the equivalent load spectrum of the corresponding type or unknown bridge crane can be predicted. In this paper, the application software of load spectrum acquisition and prediction based on support vector machine crane is programmed by using visual programming language VC 6.0.The software is applied to engineering example, and the result of prediction is compared with the actual result. It shows that it has high consistency and practicability. Then, the method is compared with the least square method and the neural network method, and the same sample data is used to obtain and predict the load spectrum of the crane, which proves the superiority of the support vector machine (SVM) method. Moreover, the software is easy to operate, so that the operator can use the software without the knowledge of support vector machine. The most important is that the research results lay a foundation for the development of the following crane fatigue life prediction software.
【学位授予单位】：太原科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH21
[Abstract]:In recent years, with the rapid development of national economy and the great investment in infrastructure, the major technical equipment industry is booming, and the lifting machinery occupies a large proportion in the major technical equipment industry. Because the lifting machinery is mostly heavy equipment, once the accident occurs, the economic loss is heavy, and it is easy to cause casualties. This country put forward the very high request to the crane safe use, and listed it as the national special equipment. Through the investigation of crane accident, it is found that the main reason is fatigue fracture. Therefore, many governments and scientific research institutions have begun to study the fatigue fracture of lifting machinery, which can be reported to the users before the fatigue fracture of the equipment, thus avoiding the occurrence of accidents. In this paper, through the deep research on the theory of metal fatigue fracture, it is concluded that the precondition to solve the fatigue fracture of crane is to draw up a program to simulate the real use of crane metal structure. Typical load-time history, namely load spectrum. Due to the randomness and uncertainty of load, it is impossible to directly apply the measured results to theoretical analysis and engineering practice. It is necessary to construct an equivalent load spectrum which can essentially reflect the variation of load on crane metal structure with time under various working conditions. In this paper, casting bridge crane and general bridge crane are taken as research objects, and some data samples of crane working condition are collected through field investigation. The support vector machine (SVM) nonlinear regression theory based on statistical learning theory is used for the first time to train the collected sample data. The nonlinear mapping relationship between the working cycle number of the corresponding type bridge crane and different lifting loads is established. Using this mapping relationship, the equivalent load spectrum of the corresponding type or unknown bridge crane can be predicted. In this paper, the application software of load spectrum acquisition and prediction based on support vector machine crane is programmed by using visual programming language VC 6.0.The software is applied to engineering example, and the result of prediction is compared with the actual result. It shows that it has high consistency and practicability. Then, the method is compared with the least square method and the neural network method, and the same sample data is used to obtain and predict the load spectrum of the crane, which proves the superiority of the support vector machine (SVM) method. Moreover, the software is easy to operate, so that the operator can use the software without the knowledge of support vector machine. The most important is that the research results lay a foundation for the development of the following crane fatigue life prediction software.
【学位授予单位】：太原科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH21
【参考文献】
相关期刊论文 前7条
1 左治兴;朱必勇;孙学森;易斌;;桥式起重机典型事故分析及安全管理[J];工业安全与环保;2006年10期
2 李鹏;;桥式起重机主梁变幅疲劳寿命试验载荷谱[J];机械强度;1991年03期
3 王德俊;平安;徐灏;;疲劳载荷谱编制准则[J];机械强度;1993年04期
4 肖涵,侯澍e，
本文编号：2208011
本文链接：https://www.wllwen.com/kejilunwen/jixiegongcheng/2208011.html
