Small Grants for Exploratory Research (SGER): Smart Rotating Machinery

作者：Sherif Noah，亚历山大G. Parlos，Suhada Jayasuriya，Texas A M University，Department of Mechanical Engineering，学院站，Texas 77843提案编号：0100238提案标题：Smart Rotating MachineProject摘要：这个探索性的研究项目解决了智能旋转机械的发展。计划外的机械停机和机械性能差对工业生产力和安全性产生了负面影响，每年的影响高达1万亿美元。因此，开发智能旋转机械是及时的，它高度适应不确定的动态环境，同时保持高水平的性能。这种机器必须包含植根于新信息技术的新的和创新的突破，使它们能够表现出记忆，从经验中学习，并利用这种学习能力来提高它们的适应性，同时以最佳的方式工作。通过基于物理的非线性转子动力学模型和通过真实的-时间传感器数据。闭环早期初期故障诊断是通过使用计算智能工具，如神经网络，模糊逻辑，遗传算法，和其他先进的信号处理方法，如小波分析。为了使智能旋转机械成为现实，将探索一种方法，使智能行为的某些元素嵌入到旋转机械的初始框架。拟议的有限努力被认为是高风险的，因为它构成了智能系统的开创性研究，在开发有效的方法方面存在明显的困难。此外，这项研究将导致早期诊断算法的受控旋转机械，一个主题，尚未在文献中得到解决的实验演示。这些算法将控制和减轻关键旋转机械即将发生的故障，降低计划外停机、紧急停机和灾难性事故的可能性。

AbstractPIs: Sherif Noah, Alexander G. Parlos, Suhada Jayasuriya, Texas A&M University, Department of Mechanical Engineering, College Station, Texas 77843Proposal Number: 0100238Proposal Title: Smart Rotating MachineryProject Abstract:This exploratory research project addresses the development of smart rotating machinery. Unplanned machinery downtime and poor machinery performance impact negatively both industrial productivity and safety at an annual level of $1 trillion. It is therefore timely to develop smart rotating machinery that are highly adaptive to uncertain dynamic environments while maintaining high level of performance. Such machinery must incorporate new and innovative breakthroughs rooted in new information technologies that enable them to exhibit memory, learn from experience and use this learning ability to improve their adaptability while performing in an optimal manner.The technical approach of the proposed research relies on health monitoring, condition assessment and early fault diagnosis through a combination of physics-based nonlinear rotordynamics models and empirical models developed through real-time sensor data. Closed-loop early incipient fault diagnosis is achieved through the use of computational intelligence tools, e.g. neural networks, fuzzy logic, and genetic algorithms, and other advanced signal processing methods, such as wavelet analysis. Towards making smart rotating machinery a reality, an initial framework will be explored for a methodology that will enable embedding certain elements of intelligent behavior into rotating machinery. The proposed limited effort is considered high-risk because it constitutes a pioneering study in smart systems with the perceived difficulties in developing an effective methodology. Furthermore, this research will lead to the experimental demonstrations of early diagnosis algorithms for controlled rotating machinery, a subject that has yet to be addressed in the literature. Such algorithms will control and mitigate impending failures of critical rotating machinery, reducing the probability of unplanned downtime, emergency shutdowns and catastrophic accidents.