Early diagnosis of gear transmission has been a significant challenge, because gear faults occur primarily at microstructure or even material level but their effects can only be observed indirectly at a system level. The performance of a gear fault diagnosis system depends significantly on the features extracted and the classifier subsequently applied. Traditionally, fault-related features are extracted and identified based on domain expertise through data preprocessing which are system-specific and may not be easily generalized. On the other hand, although recently the deep neural networks based approaches featuring adaptive feature extractions and inherent classifications have attracted attention, they usually require a substantial set of training data. Aiming at tackling these issues, this paper presents a deep convolutional neural network-based transfer learning approach. The proposed transfer learning architecture consists of two parts; the first part is constructed with a pre-trained deep neural network that serves to extract the features automatically from the input, and the second part is a fully connected stage to classify the features that needs to be trained using gear fault experimental data. Case analyses using experimental data from a benchmark gear system indicate that the proposed approach not only entertains preprocessing free adaptive feature extractions, but also requires only a small set of training data.

This letter reports designs of adaptive metasurfaces capable of modulating incoming wave fronts of elastic waves through electromechanical-tuning of their cells. The proposed elastic metasurfaces are composed of arrayed piezoelectric units with individually connected negative capacitance elements that are online tunable. By adjusting the negative capacitances properly, accurately formed, discontinuous phase profiles along the elastic metasurfaces can be achieved. Subsequently, anomalous refraction with various angles can be realized on the transmitted lowest asymmetric mode Lamb wave. Moreover, designs to facilitate planar focal lenses and source illusion devices can also be accomplished. The proposed flexible and versatile strategy to manipulate elastic waves has potential applications ranging from structural fault detection to vibration/noise control.

Piezoelectric cantilevers are widely used in vibration energy harvesting. Simple cantilever-based harvesters are mostly unidirectional. In this article, we develop a cantilever-pendulum system that can harvest vibratory energy of excitations from an arbitrary direction. The new design consists of a traditional piezoelectric cantilever with a pendulum attached to the tip. It is shown analytically and experimentally that with proper parametric combination this system can induce modal energy interchange between beam vibration and pendulum motions due to 1:2 internal resonance, which ultimately yields multi-directional energy harvesting by a single cantilever. The underlying mechanisms of this design are analyzed in detail.

This letter reports a tunable acoustic prism featuring continuous beam steering for transverse waves at a single frequency. The proposed prism is comprised of arrayed piezoelectric unit-cells with individually connected inductive shunt circuits. Taking advantage of wave velocity shifting in the vicinity of the local resonant frequency of unit-cells, we can steer the harmonic transverse wave by adjusting the inductive loads, i.e., tuning the inductances. This beam steering mechanism is facilitated by add-on piezoelectric circuitry through electro-mechanical coupling, whereas the host medium is not altered. Our analysis shows that the propagation direction of an acoustic wave has a tuning range of more than 30°. This tunable prism can be used as an acoustic metamaterial for various devices operating at broadband frequencies.