Deep learning (DL) has attained remarkable achievements in diagnosing faults for rotary machineries. Capitalizing on the formidable learning capacity of DL, it has the potential to automate human labor and augment the efficiency of fault diagnosis in rotary machinery. These advantages have engendered escalating interest over the past decade. Although recent reviews of the literature have encapsulated the utilization of DL in diagnosing faults in rotating machinery, they no longer encompass the introduction of novel methodologies and emerging directions as DL methodologies continually evolve. Moreover, in practical application, novel issues and trajectories perpetually manifest, demanding a comprehensive exegesis. To rectify this lacuna, this article amalgamates current research trends and avant-garde methodologies while systematizing the utilization of anterior DL techniques. The evolution and extant status of DL in diagnosing faults for rotary machinery were delineated, with the intent of providing orientation for prospective research. Over the bygone decade, archetypal DL theory has empowered the diagnosis of faults in rotating machinery by directly establishing the nexus between mechanical data and fault conditions. In recent years, meta learning methods aimed at solving small sample scenarios and large model transformers aimed at mining big data features have both received widespread attention and development in the field of fault diagnosis of rotating machinery equipment. Although excellent results have been achieved in these two directions, there is no review and summary article yet, so it is necessary to update the review literature in the field of fault diagnosis of rotating machinery equipment. Lastly, predicated on a survey of the literature and the current developmental landscape, the challenges and prospective orientations of DL in rotary machinery fault diagnosis are presented.

Accurate wind speed prediction can improve the utilization efficiency of wind energy effectively. However, the original time series signals of wind speed present nonlinear or non-stationary characteristics, which make wind speed forecasting very difficult. A new pre-processing method of wind speed signal, named improved empirical wavelet transform ( IEWT), is proposed in this paper, whish is inspired from spectrum segmentation theory of EWT. Then, a novel hybrid model combining IEWT and least square support vector machine (LSSVM) is put forward to predict short-term wind speed. In order to achieve better prediction results, the parameters are jointly optimized through bird swarm algorithm (BSA) in this model. The prediction performance of the established hybrid model is tested with practical wind speed signal of four seasons from a wind farm situated in Zhejiang, China. The experimental results indicate that the presented hybrid forecasting model can effectively follow the change of wind speed, which exhibits more superior predicting performance than other popular models.