Moving heavy objects with overhead cranes requires the operator to fasten the object to a hook with ropes or chains. This is a time-consuming process, which could be avoided by using universal grippers that can lift objects of any shape. This study was conducted to find if a universal gripper, based on granular jamming, can be used for crane scale applications. Maximum lifting capacity of granular jamming grippers was analytically evaluated and experimentally tested with various material combinations. Objects with different shapes, sizes and weights were successfully lifted with selected gripper configurations. The results showed that grain size and grain compressibility both affect the performance of the gripper. It was demonstrated that in order to efficiently lift heavy objects with granular jamming, the granular material has to be compressed sufficiently. Pressure difference between environment and the sealed pouch, filled with granular material, has to be correct. With this setup, gripper based on granular jamming was able to lift objects with various shapes; and weights up to 120Â kg.
6 September 2023 in Mechanical Systems and Signal Processing
Accurate system health state prediction through deep learning requires extensive and varied data. However, real-world data scarcity poses a challenge for developing robust fault diagnosis models. This study introduces two extensive datasets, Aalto Shim Dataset and Aalto Gear Fault Dataset, collected under controlled laboratory conditions, aimed at advancing deep learning-based fault diagnosis. The datasets encompass a wide range of gear faults, including synthetic and realistic failure modes, replicated on a downsized azimuth thruster testbench equipped with multiple sensors. The data features various fault types and severities under different operating conditions. The comprehensive data collected, along with the methodologies for creating synthetic faults and replicating common gear failures, provide valuable resources for developing and testing intelligent fault diagnosis models, enhancing their generalization and robustness across diverse scenarios.
The support structures of a rotating machine affect its overall dynamic behavior. The stiffness of the support structures often differs at the actual sites compared to the test rigs, which leads to uncertain dynamics. In this research, a novel method is developed to identify the support stiffness for an in-situ machine using a simulation-data-driven, deep learning algorithm. In this transfer learning approach, a deep learning algorithm is trained with a simulation model and then tested with measured vibration of a rotor-bearing-support system. To validate the algorithm for multiple cases, an experimental test rig with variable horizontal support stiffness is used. The results from the deep learning algorithm are compared with Linear regression (LR), Artificial Neural Network (ANN), and Support vector regression (SVR) for benchmarking. The models are trained with filtered frequency domain response, and challenges due to measurement uncertainty are analyzed. With the proposed pre-processing steps of the frequency domain response and outlier elimination, the deep learning-based virtual sensor can predict the support stiffness with reasonable accuracy, where the limiting factor is the data quality and lack of excitation at critical speed frequencies.
Intelligent fault diagnosis (IFD) models have the potential to increase the level of automation and the diagnosis accuracy of machine condition monitoring systems. Many of the latest IFD models rely on convolutional layers for feature extraction from vibration data. The majority of these models employ batch normalisation (BN) for centring and scaling the input for each neuron. This study includes a novel examination of a competitive approach for layer input normalisation in the scope of fault diagnosis. Network deconvolution (ND) is a technique that further decorrelates the layer inputs reducing redundancy among the learned features. Both normalisation techniques are implemented on three common 1D-CNN-based fault diagnosis models. The models with ND mostly outperform the baseline models with BN in three experiments concerning fault datasets from two different rotor systems. Furthermore, the models with ND significantly outperform the baseline models with BN in the common CWRU bearing fault tests with load domain shifts, if the data from drive-end and fan-end sensors are employed. The results show that whitened features can improve the performance of CNN-based fault diagnosis models.