A Boussinesq-Type Model for Nonlinear Wave-Heaving Cylinder Interaction

In the present work, a Boussinesq-type numerical model is developed for the simulation of nonlinear wave-heaving cylinder interaction. The wave model is able to describe the propagation of fully dispersive and weakly nonlinear waves over any finite water depth. The wave-cylinder interaction is taken into account by solving simultaneously an elliptic equation that determines the pressure exerted by the fluid on the floating body. The heave motion for the partially immersed floating cylinder under the action of waves is obtained by solving numerically the body’s equation of motion in the z direction based on Newton’s law. The developed model is applied for the case of a fixed and a free-floating circular cylinder under the action of regular waves, as well as for a free-floating cylinder undergoing a forced motion in heave. Results (heave and surge exciting forces, heave motions, and wave elevation) are compared with those obtained using a frequency domain numerical model, which is based on the boundary integral equation method.

An Optimized Fuzzy Controlled Charging System for Lithium-Ion Batteries Using a Genetic Algorithm

Fast charging is an attractive way of charging batteries; however, it may result in an undesired degradation of battery performance and lifetime because of the increase in battery temperature during fast charge. In this paper we propose a simple optimized fuzzy controller that is responsible for the regulation of the charging current of a battery charging system. The basis of the method is a simple dynamic equivalent circuit type model of the Li-ion battery that takes into account the temperature dependency of the model parameters, too. Since there is a tradeoff between the charging speed determined by the value of the charging current and the increase in temperature of the battery, the proposed fuzzy controller is applied for controlling the charging current as a function of the temperature. The controller is optimized using a genetic algorithm to ensure a jointly minimal charging time and battery temperature increase during the charging. The control method is adaptive in the sense that we use parameter estimation of an underlying dynamic battery model to adapt to the actual status of the battery after each charging. The performance and properties of the proposed optimized charging control system are evaluated using a simulation case study. The evaluation was performed in terms of the charge profiles, using the fitness values of the individuals, and in terms of the charge performance on the actual battery. The proposed method has been evaluated compared to the conventional contant current-constant voltage methods. We have found that the proposed GA-fuzzy controller gives a slightly better performance in charging time while significantly decreasing the temperature increase.

Modelling the Impedance Response of Graded LiFePO4 Cathodes for Li-Ion Batteries

Graded electrodes for Li-ion batteries aim to exploit controlled variations in local electrode microstructure to improve overall battery performance, including reduced degradation rates and increased capacity at high discharge rates. However, the mechanisms by which grading might deliver performance benefit, and under what conditions, are not yet fully understood. A Li-ion battery electrochemical model (a modified Doyle-Fuller-Newman type model capable of generating impedance functions) is developed in which local microstructural changes are captured in order to understand why and when graded electrodes can offer performance benefits. Model predictions are evaluated against experimental electrochemical impedance data obtained from electrodes with micro-scale, controlled variations in microstructure. A region locally enriched with carbon at the electrode/current collector interface is shown to significantly reduce the overpotential distribution across the thickness of a LiFePO$_4$-based Li-ion battery cathode, resulting in a lower charge transfer resistance and impedance. The insights gained from the LiFePO$_4$-based electrodes are generalised to wider design principles for both uniform and graded Li-ion battery electrodes.

STHarDNet: Swin Transformer with HarDNet for MRI Segmentation

In magnetic resonance imaging (MRI) segmentation, conventional approaches utilize U-Net models with encoder–decoder structures, segmentation models using vision transformers, or models that combine a vision transformer with an encoder–decoder model structure. However, conventional models have large sizes and slow computation speed and, in vision transformer models, the computation amount sharply increases with the image size. To overcome these problems, this paper proposes a model that combines Swin transformer blocks and a lightweight U-Net type model that has an HarDNet blocks-based encoder–decoder structure. To maintain the features of the hierarchical transformer and shifted-windows approach of the Swin transformer model, the Swin transformer is used in the first skip connection layer of the encoder instead of in the encoder–decoder bottleneck. The proposed model, called STHarDNet, was evaluated by separating the anatomical tracings of lesions after stroke (ATLAS) dataset, which comprises 229 T1-weighted MRI images, into training and validation datasets. It achieved Dice, IoU, precision, and recall values of 0.5547, 0.4185, 0.6764, and 0.5286, respectively, which are better than those of the state-of-the-art models U-Net, SegNet, PSPNet, FCHarDNet, TransHarDNet, Swin Transformer, Swin UNet, X-Net, and D-UNet. Thus, STHarDNet improves the accuracy and speed of MRI image-based stroke diagnosis.

Modelling of stress transfer in root-reinforced soils informed by four-dimensional X-ray computed tomography and digital volume correlation data

Vegetation enhances soil shearing resistance through water uptake and root reinforcement. Analytical models for soils reinforced with roots rely on input parameters that are difficult to measure, leading to widely varying predictions of behaviour. The opaque heterogeneous nature of rooted soils results in complex soil–root interaction mechanisms that cannot easily be quantified. The authors measured, for the first time, the shear resistance and deformations of fallow, willow-rooted and gorse-rooted soils during direct shear using X-ray computed tomography and digital volume correlation. Both species caused an increase in shear zone thickness, both initially and as shear progressed. Shear zone thickness peaked at up to 35 mm, often close to the thickest roots and towards the centre of the column. Root extension during shear was 10–30% less than the tri-linear root profile assumed in a Waldron-type model, owing to root curvature. Root analogues used to explore the root–soil interface behaviour suggested that root lateral branches play an important role in anchoring the roots. The Waldron-type model was modified to incorporate non-uniform shear zone thickness and growth, and accurately predicted the observed, up to sevenfold, increase in shear resistance of root-reinforced soil.

Predicting flow stress of Ni steel based on machine learning algorithm

This article builds a stress–strain prediction model based on production data from the steel industry by using machine learning algorithms. Based on the stress–strain data of 9Ni steel hot deformation behavior, the prediction model of flow stress constitutive equation of 9Ni steel is established. Four models, including Arrhenius-type model considering strain compensation, Arrhenius-type model of Stochastic Configuration Networks (SCNs) neural network, Arrhenius-type model of Multi-objective Particle Swarm Optimization (AMPSO) and Support Vector Machine (SVM) model, are adopted in this research. The results show that the Arrhenius-type model considering strain compensation can predict the stress trend under different deformation conditions, but its prediction accuracy has some deviation. The prediction model based on SVM algorithm has the best prediction accuracy. The square of Correlation Coefficient (R2), the Average Absolute Relative Error (AARE), and Mean Square Error (MSE) are 0.99996, 0.002455, and 0.1998, respectively. Based on the data of 9Ni steel hot deformation behavior, the prediction models of machine learning algorithm have good application prospects in steel industry.