ABSTRACT
The assembled camshaft is a novel manufacturing product which connects the cam and the mandrel by tube hydroforming (THF) technology after they are processed separately. However, in the process of THF, the structure of the cam-bores has a crucial influence on the connection strength of the assembled camshafts. Therefore, three kinds of cam-bores with circular structure, isometric-trilateral profile and logarithmic spiral profile are selected for hydroforming with a hollow mandrel (tube) in this study. The finite-element-analysis is carried out by ABAQUS software, the variations of (residual) contact pressure and contact area under different structures are obtained, and the torsional angle variations after assembly are measured. Further, the connection strength of the assembled camshaft under three structures is discussed. The results show that the evaluation of connection strength of the assembled camshaft is affected by many factors, including contact pressure, maximum residual contact pressure, axial and circular residual contact pressure, contact area and its rate, residual contact area percentage and torsional angle. Through the comprehensive analysis of various factors, the torsional angle of the camshaft with circular structure is the largest, i.e. poor connection strength. By contrast, the torsional strength of the camshaft with isometric-trilateral profile is the largest, namely, the best connection strength.
ABSTRACT
Intelligent fault diagnosis of roller bearings is facing two important problems, one is that train and test datasets have the same distribution, and the other is the installation positions of accelerometer sensors are limited in industrial environments, and the collected signals are often polluted by background noise. In the recent years, the discrepancy between train and test datasets is decreased by introducing the idea of transfer learning to solve the first issue. In addition, the non-contact sensors will replace the contact sensors. In this paper, a domain adaption residual neural network (DA-ResNet) model using maximum mean discrepancy (MMD) and a residual connection is constructed for cross-domain diagnosis of roller bearings based on acoustic and vibration data. MMD is used to minimize the distribution discrepancy between the source and target domains, thereby improving the transferability of the learned features. Acoustic and vibration signals from three directions are simultaneously sampled to provide more complete bearing information. Two experimental cases are conducted to test the ideas presented. The first is to verify the necessity of multi-source data, and the second is to demonstrate that transfer operation can improve recognition accuracy in fault diagnosis.
ABSTRACT
Time-frequency ridge not only exhibits the variable process of non-stationary signal with time changing but also provides the information of signal synchronous or non-synchronous components for subsequent detection research. Consequently, the key is to decrease the error between real and estimated ridge in the time-frequency domain for accurate detection. In this article, an adaptive weighted smooth model is presented as a post-processing tool to refine the time-frequency ridge which is based on the coarse estimated time-frequency ridge using newly emerging time-frequency methods. Firstly, the coarse ridge is estimated by using multi-synchrosqueezing transform for vibration signal under variable speed conditions. Secondly, an adaptive weighted method is applied to enhance the large time-frequency energy value location of the estimated ridge. Then, the reasonable smooth regularization parameter associated with the vibration signal is constructed. Thirdly, the majorization-minimization method is developed for solving the adaptive weighted smooth model. Finally, the refined time-frequency characteristic is obtained by utilizing the stop criterion of the optimization model. Simulation and experimental signals are given to validate the performance of the proposed method by average absolute errors. Compared with other methods, the proposed method has the highest performance in refinement accuracy.
