The comprehensive spectral diagnostic method for crane slewing mechanism fault diagnosis—innovatively developed by WTAU—approaches core technical research from the perspective of multi-dimensional integration and analysis of test data, with the aim of ensuring the operational safety of cranes. Specifically addressing the analysis of vibration data generated during variable-speed crane operation, the company has developed corresponding multi-dimensional analysis algorithms, generalized dynamic data analysis algorithms, and fault feature extraction algorithms.
This suite of algorithmic tools—developed specifically to address the complexity of comprehensive fault spectra—integrates various time-frequency domain analysis methods to support a robust data analysis platform. It enables broad-spectrum, low-frequency data analysis; dynamic data analysis; and the extraction of nonlinear and non-stationary fault features, as well as fault diagnosis and prediction. The suite is broadly applicable to all types of cranes.
Leveraging innovative fault diagnosis techniques grounded in equipment failure mechanisms, the system performs diagnostics regarding fault type, location, severity, and trend. In addition to supporting automated early-warning diagnostics, it features data augmentation algorithms that enable comprehensive analysis of various crane faults, while maintaining compatibility with a diverse range of algorithms to facilitate deep data mining of crane-related big data.
The WTAU Online Multi-dimensional Fault Diagnosis and Analysis System for Crane Slewing Mechanisms encompasses three primary functional categories:
1. Classic Diagnostics: Time-domain analysis and frequency-domain analysis. These classic signal analysis methods—which historically played a pivotal role in the fault analysis of rotating machinery—are based on Fourier transforms and are applicable to stationary signals.
2. Precision Diagnosis: Utilizing time-frequency analysis methods—specifically modern spectral analysis techniques such as filtering and demodulation based on wavelet analysis (linear) and Wigner distribution (nonlinear), adaptive high-frequency demodulation, intensity spectra, and composite spectra—to identify specific signals (e.g., short-duration sinusoids and linear frequency-modulated pulses) and effectively extract fault features.
3. Intelligent Diagnostics: An independently developed order spectrum method based on the smoothed pseudo-Wigner distribution, capable of adaptively adjusting order resolution. Features include a self-learning function for variable-load alarm thresholds and intelligent early warning capabilities.
