Online Condition Monitoring of Electric Machines

电机（EM）（即，电动机和发电机）通常用于各种家用和工业应用，例如电动车辆、机床和风力涡轮机。提高新兴市场的绩效和效率对各行各业都具有极其重要的意义。尽管几十年来已经进行了巨大的努力来开发用于自动EM诊断的技术和专家系统，但不幸的是，由于可靠性差，这些系统中的大多数不能适当地用于工业监测应用，并且会发生错过的警报（即，未被识别的现有故障）和错误警报（即，由于除真实的故障以外的原因而触发的警报）。该研究计划的目标是开发新的技术和工具，用于EM的智能诊断和故障诊断（IDP）。其目标是在最早阶段识别EM缺陷的发生，以防止EM性能下降，故障，甚至相关设施的灾难性故障。当出现潜在问题时，IDP监控器可以精确定位故障组件并估计故障EM的剩余使用寿命。除了提高生产率外，预测信息还有助于降低运营成本，因为可以适当安排维护，以避免意外的设备停机。由于最常用的EM是感应电动机（IM），每年也消耗世界上超过50%的电能，因此拟议的研究将集中在IM上。第一个研究主题是开发新的信号处理技术，以实现更有效的去噪和IM故障检测。第二个目标是全面评估所有可用的IM故障检测技术对应于不同的操作和电机条件的鲁棒性。第三个目标是开发一种新的调节预测器，用于高维系统的多步预测。这项创新研究计划的另一个目标是开发一个IDP平台，整合诊断和预后信息，以便真实的对IM状况进行更积极的评估。新的系统训练策略将被提出，以提高IDP的适应能力，以适应不同的IM条件。这个多学科的研究计划将提供独特的和领先的机会，在这些相关领域的培训HQP。此外，开发的技术和智能工具将有利于加拿大公司寻求提高其在全球市场上的竞争力，提高其生产率和质量，降低成本。

Electric machines (EMs) (i.e., motors and generators) are commonly used in various domestic and industrial applications such as electric vehicles, machine tools, and wind turbines. Improving the performance and efficiency of EMs is of utmost significance to a wide array of industries. Although enormous effort has been undertaken over the decades to develop techniques and expert systems for automatic EM diagnosis, unfortunately, most of these systems cannot be used properly for industrial monitoring application due to poor reliability, with the occurrence of missed alarms (i.e., existing faults that are not identified) and false alarms (i.e., alarms triggered for reasons other than real faults). The objective of this research program is to develop new technologies and tools for intelligent diagnostics and prognostics (IDP) of EMs. The goal is to recognize the occurrence of an EM defect at its earliest stage, so as to prevent EM performance degradation, malfunction, or even catastrophic failure of the related facilities. When a potential problem arises, the IDP monitor can pinpoint the faulty components and estimate the remaining useful life of the faulty EM. In addition to improving production rates, the prognostic information can help to reduce operational costs because maintenance can be scheduled properly to avoid unexpected equipment shutdowns. Since the most commonly used EMs are induction motors (IMs) that also consume more than 50% of the electrical energy in the world each year, the proposed research will focus on IMs. The first research theme is to develop new signal processing techniques for more efficient denoising and IM fault detection. The second objective is to comprehensively assess the robustness of all available IM fault detection techniques corresponding to different operating and motor conditions. The third objective is to develop a new regulated predictor for multi-step-ahead forecasting of high-dimensional systems. Another objective of this innovative research program is to develop an IDP platform to integrate both diagnostic and prognostic information for a more positive assessment of IM conditions in real time. New strategies for system training will be proposed to improve the adaptive capability of the IDP to accommodate different IM conditions. This multidisciplinary research program will provide unique and leading-edge opportunities to train HQP in these related areas. In addition, the developed technologies and intelligent tools will benefit Canadian companies seeking to enhance their competitiveness in the global market by improving their production rates and quality and reducing costs.