Zhong-Yang Li

A Time-Frequency Based Method for the Detection and Tracking of Multiple Non-Linearly Modulated Components With Births and Deaths

The estimation of the components which contain the characteristics of a signal attracts great attention in many real world applications. In this paper, we address the problem of the tracking of multiple signal components over discrete time series. We propose an algorithm to first detect the components from a given time-frequency distribution and then to track them automatically. In the first place, the peaks corresponding to the signal components are detected using the statistical properties of the spectral estimator. Then, an original classifier is proposed to automatically track the detected peaks in order to build components over time. This classifier is based on a total divergence matrix computed from a peak-component divergence matrix that takes account of both amplitude and frequency information. The peak-component pairs are matched automatically from this divergence matrix. We propose a stochastic discrimination rule to decide upon the acceptance of the peak-component pairs. In this way, the algorithm can estimate the number, the amplitude and frequency modulation functions, and the births and the deaths of the components without any limitation on the number of components. The performance of the proposed method, a post-processing of a time-frequency distribution is validated on simulated signals under different parameter sets. The method is also applied to four real-world signals as a proof of its applicability.

Parameter estimation of short-time multi-component signals using damped-amplitude & polynomial-frequency model

This paper concerns the parameter estimation of multi-component damped oscillations having non-linear frequency. In this paper, the instantaneous frequency is approximated by polynomials while the amplitude is characterized by damped exponentials to connect directly to its physical interpretations. A maximum likelihood procedure is developed via an adaptive simulated annealing technique which helps to speed up the convergence. Results on simulated signals show that the proposed algorithm is more efficient than the algorithm based on polynomial amplitude models, and allows the estimation of damping coefficients over a very short time duration. Finally, the proposed algorithm is applied for characterizing the ambient vibrations of a building.

AStrion strategy: from acquisition to diagnosis - Application to wind turbine monitoring

This paper proposes an automatic procedure for condition monitoring. It presents a valuable tool for the maintenance of expensive and spread systems, such as wind turbine farms. Thanks to data-driven signal processing algorithms, the proposed solution is fully automatic for the user. The paper briefly describes all the steps of the processing, from preprocessing of the acquired signal to interpretation of the generated results. It starts with an angular resampling method with speed measurement correction. Then comes a data validation step, in both the time/angular and frequency/order domains. After the preprocessing, the spectral components of the analysed signal are identified and classified into several classes, from sine wave to narrowband components. This spectral peak detection and classification allows the harmonic and side-band series to be extracted, which may be part of the signal spectral content. Moreover, the detected spectral patterns are associated with the characteristic frequencies of the investigated system. Based on the detected side-band series, the full-band demodulation is performed. At each step, the diagnosis features are computed and dynamically tracked, signal by signal. Finally, system health indicators are proposed to conclude the condition of the investigated system. All the steps mentioned create a self-sufficient tool for a robust diagnosis of mechanical faults. The paper presents the performance of the proposed method on real-world signals from a wind turbine drivetrain.