Development of artificial neural networks for the prediction of the pressure field along a horizontal pipe conveying high-viscosity two-phase flow

Ajbar, W.; Torres, L.; Guzmán, J.E.V.; Hernández-García, J.; Palacio-Pérez, A.

doi:10.1016/j.flowmeasinst.2024.102541

Cited by 6 publications

(1 citation statement)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This noise can adversely affect subsequent wind turbine fault diagnosis. Common transfer functions include the log-sigmoid (LOGSIG), TANSIG, and purelin functions [ 29 ]. Among these, the TANSIG function is less sensitive to small fluctuations in the input, resulting in greater noise resistance [ 30 ].…”

Section: Introductionmentioning

confidence: 99%

Fault Diagnosis of Wind Turbine Gearbox Based on Modified Hierarchical Fluctuation Dispersion Entropy of Tan-Sigmoid Mapping

Wang,

2024

Entropy

View full text Add to dashboard Cite

Vibration monitoring and analysis are important methods in wind turbine gearbox fault diagnosis, and determining how to extract fault characteristics from the vibration signal is of primary importance. This paper presents a fault diagnosis approach based on modified hierarchical fluctuation dispersion entropy of tan-sigmoid mapping (MHFDE_TANSIG) and northern goshawk optimization–support vector machine (NGO–SVM) for wind turbine gearboxes. The tan-sigmoid (TANSIG) mapping function replaces the normal cumulative distribution function (NCDF) of the hierarchical fluctuation dispersion entropy (HFDE) method. Additionally, the hierarchical decomposition of the HFDE method is improved, resulting in the proposed MHFDE_TANSIG method. The vibration signals of wind turbine gearboxes are analyzed using the MHFDE_TANSIG method to extract fault features. The constructed fault feature set is used to intelligently recognize and classify the fault type of the gearboxes with the NGO–SVM classifier. The fault diagnosis methods based on MHFDE_TANSIG and NGO–SVM are applied to the experimental data analysis of gearboxes with different operating conditions. The results show that the fault diagnosis model proposed in this paper has the best performance with an average accuracy rate of 97.25%.

show abstract

Section: Introductionmentioning

confidence: 99%