基于多传感器信息融合技术的电梯智能诊断系统的研究

发布时间：2018-06-01 20:59

本文选题：多传感器信息融合 + BP神经网络　；参考：《新疆大学》2013年硕士论文

【摘要】：随着我国经济持续增长、城镇化建设的加速和房地产行业的进一步发展，电梯市场需求量因多方面需求得到迅速增长，目前，全国的电梯拥有量超过了100万台，并以每年20%的速度递增。虽然近来年电梯的安全性、可靠性大大增加，，但随着人们对生活质量的日益重视，人们对电梯运行的安全性、舒适性以及故障的诊断和预测都提出了更高的要求。由于电梯的使用数量和范围的扩大，电梯故障、事故也时常发生，电梯运行的安全问题受到了社会的广泛关注。如何提高电梯的运行效率、避免过多故障的发生。提高电梯运行的稳定性、安全性和舒适性已成为电梯行业面临的研究课题。本文结合电梯的结构特点和运行状态特点，从乘坐电梯的舒适度出发，分析了产生轿厢振动的原因及振动机理，设计了基于CAN总线的分布式检测系统，可以实现对电梯轿厢垂直加速度、水平加速度、噪声等参数的检测，为智能诊断提供数据基础；讨论了多传感器信息融合术的原理、结构形式；研究多传感器信息融合不同层次的特征，比较多传感器信息融合各种算法的优缺点；结合电梯轿厢振动的特点，设计了基于多传感器信息融合的电梯诊断模型，首先给出了基于神经网络信息融合的诊断方法，分析了BP神经网络的特点及神经网络用于电梯诊断的可能性和有效性。其次，针对故障的不确定性，提出了基于D-S论证理论的决策融合。最后给出了基于神经网络和D-S论证理论相结合的信息融合电梯诊断法方法。并设计了基于VC++与Matlab联合编程的电梯智能诊断软件。多传感器信息融合提高了诊断的可靠性，克服传统的单一传感器检测所带来的局限性与片面性，更为准确地反映电梯性能。
[Abstract]:With the sustained growth of our economy, the acceleration of urbanization and the further development of the real estate industry, the market demand for elevators has increased rapidly because of various demands. At present, the number of elevators in China has exceeded 1 million. And at an annual rate of 20%. Although the safety and reliability of elevators have greatly increased in recent years, with the increasing attention to the quality of life, people put forward higher requirements for the safety, comfort, fault diagnosis and prediction of elevator operation. Due to the expansion of the number and scope of elevator use, elevator failures and accidents often occur, the safety of elevator operation has been widely concerned by the society. How to improve the efficiency of elevator operation and avoid too many failures. Improving the stability, safety and comfort of elevator operation has become a research topic facing the elevator industry. Based on the structure and running state of elevator, this paper analyzes the cause and mechanism of car vibration, and designs a distributed detection system based on CAN bus. The vertical acceleration, horizontal acceleration, noise and other parameters of elevator car can be detected, which provides the data basis for intelligent diagnosis, and discusses the principle and structure of multi-sensor information fusion. The characteristics of different levels of multi-sensor information fusion are studied, and the advantages and disadvantages of various algorithms of multi-sensor information fusion are compared. Combined with the characteristics of elevator car vibration, the elevator diagnosis model based on multi-sensor information fusion is designed. Firstly, the diagnosis method based on neural network information fusion is presented, and the characteristics of BP neural network and the possibility and effectiveness of BP neural network in elevator diagnosis are analyzed. Secondly, the decision fusion based on D-S theory is proposed for the uncertainty of fault. Finally, a method of information fusion elevator diagnosis based on neural network and D-S theory is presented. The elevator intelligent diagnosis software based on VC and Matlab is designed. Multi-sensor information fusion improves the reliability of diagnosis, overcomes the limitation and one-sidedness of traditional single sensor detection, and more accurately reflects the elevator performance.
【学位授予单位】：新疆大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TU857;TP277

【相似文献】