Mechanism-Informed Degradation Assessment for Machinery Incorporating Monitoring and Diagnosis Under Nonstationary Operating Conditions

The reliability and stability of mechatronic machinery significantly affect the sustained industrial service. Degradation assessment for vital components bridges the gap between the operation and the predictive maintenance. However, most of the current available methods, in regard of industrial applications, expose deficiency in the integration of condition monitoring and fault diagnosis. Besides, the adaptability under nonstationary operating conditions remains a tough issue. To this end, a mechanism-informed degradation assessment architecture, which concurrently incorporates the monitoring and diagnosis, is proposed in this article. First, with the dynamic signal demodulation, the mechanism-informed target characteristic components (TCCs) are identified and sequentially extracted. Second, hypothesis test is conducted to verify the randomness of the TCCs, achieving the diagnostic result from the statistic aspect. Afterwards, an amplitude normalization strategy is contrapuntally designed to diminish the nonstationary operating condition induced interference. On this basis, a novel degradation assessment indicator, by integrating the condition monitoring and diagnosis results, is specifically constructed for mechatronic machinery. Experiment tests orienting industrial environment at a mechatronic gearbox platform and an offshore wind turbine transmission system demonstrate that the proposed method can achieve more credible diagnosis-affirmed degradation assessment result under nonstationary operating conditions by comparing with the conventional approaches.