Artificial intelligent aided design and manufacturing have been recognized as one kind of robust data-driven and data-intensive technologies in the integrated computational material engineering (ICME) era. Motivated by the dramatical developments of the services of China Railway High-speed series for more than a decade, it is essential to reveal the foundations of lifecycle management of those trains under environmental conditions. Here, the smart design and manufacturing of welded Q350 steel frames of CR200J series are introduced, presenting the capability and opportunity of ICME in weight reduction and lifecycle management at a cost-effective approach. In order to address the required fatigue life time enduring more than 9×10⁶ km, the response of optimized frames to the static and the dynamic loads are comprehensively investigated. It is highlighted that the maximum residual stress of the optimized welded frame is reduced to 69 MPa from 477 MPa of previous existing one. Based on the measured stress and acceleration from the railways, the fatigue life of modified frame under various loading modes could fulfil the requirements of the lifecycle management. Moreover, our recent developed intelligent quality control strategy of welding process mediated by machine learning is also introduced, envisioning its application in the intelligent welding.

Drug target relationship (DTR) prediction is a rapidly evolving area of research in computational drug discovery. Despite recent advances in computational solutions that have overcome the challenges of in vitro and in vivo experiments, most computational methods still focus on binary classification. They ignore the importance of binding affinity, which correctly distinguishes between on-targets and off-targets. In this study, we propose a deep learning model based on the microstructure of compounds and proteins to predict drug-target binding affinity (DTA), which utilizes topological structure information of drug molecules and sequence semantic information of proteins. In this model, graph attention network (GAT) is used to capture the deep features of the compound molecular graph, and bidirectional long short-term memory (BiLSTM) network is used to extract the protein sequence features, and the pharmacological context of DTA is obtained by combining the two. The results show that the proposed model has achieved superior performance in both correctly predicting the value of interaction strength and correctly discriminating the ranking of binding strength compared to the state-of-the-art baselines. A case study experiment on COVID-19 confirms that the proposed DTA model can be used as an effective pre-screening tool in drug discovery.

This study aims to explore a method suitable for welding 7A52 high-strength aluminum alloy plates with continuously varying thicknesses and the causes of microscopic defects in welds in order to improve welding quality. Comparative tests were conducted to analyze weld defects and deformation when welding the aluminum alloy plates with varying thicknesses at constant laser power. The laser power required for melting welds at varying-thickness positions was estimated. Weld defects and deformation when welding aluminum alloy plates with varying thicknesses at continuous variable laser power were detected. The causes of microscopic weld defects during constant-power welding were analyzed. The welding defects and deformation and the welding quality were improved by welding aluminum alloy plates at continuous variable power.

Integrated computational materials engineering (ICME) has emerged to be one of the most powerful materials genome engineering (MGE) approaches in designing new materials and manufacturing processes in recent years. It has successfully deployed many new products for the electronic, automotive, and aerospace industries. This paper reviews the current status of research on first principles in the design of high-strength Mg alloys, discusses the application of crystal plasticity finite element models to the microscale slip, twinning, microstructure morphology, texture evolution, and macroscopic forming of Mg alloys, and introduces the research progress of crystal plasticity finite element models and phase field models, meta cellular automata models and first principles coupled models respectively, around the need for multi-scale coupled simulations of Mg alloys. The key technology obstacles of integrating the first principles, crystal plasticity finite element, and microstructure models for Mg alloys have been solved. This paper can provide a reference for the design of new Mg alloy compositions and the development of high-performance Mg alloys.

Integrated computational materials engineering (ICME) is to integrate multi-scale computational simulations and key experimental methods such as macroscopic, mesoscopic, and microscopic into the whole process of Al alloys design and development, which enables the design and development of Al alloys to upgrade from traditional empirical to the integration of composition-process-structure-mechanical property, thus greatly accelerating its development speed and reducing its development cost. This study combines calculation of phase diagram (CALPHAD), Finite element calculations, first principle calculations, and microstructure characterization methods to predict and regulate the formation and structure of composite precipitates from the design of high-modulus Al alloy compositions and optimize the casting process parameters to inhibit the formation of micropore defects in the casting process, and the final tensile strength of Al alloys reaches 420 MPa and Young’s modulus reaches more than 88 GPa, which achieves the design goal of the high strength and modulus Al alloys, and establishes a new mode of the design and development of the strength/modulus Al alloys.

A solid sustained-release energetic material sample, an eruption device and a complete test system were prepared further to analyse the combustion characteristics of solid sustained-release energetic materials. The high-temperature heat flux generated by the combustion of the samples from the eruption device was used to penetrate the Q235 target plate. In addition, the meaning and calculation formula of energy density characterising the all-around performance of heat flux were proposed. The numerical simulation of the combustion effect of samples was carried out. According to the data comparison, the numerical simulation results agreed with the experimental results, and the maximum deviation between the two was less than 8.9%. In addition, the structure of the combustion wave and high-temperature jet was proposed and analysed. Based on theoretical analysis, experimental research and numerical simulation, the theoretical burning rate formula of the sample was established. The maximum error between the theoretically calculated mass burning rate and the experimental results was less than 9.8%. Therefore, using the gas-phase steady-state combustion model to study the combustion characteristics of solid sustained-release energetic materials was reasonable. The theoretical burning rate formula also had high accuracy. Therefore, the model could provide scientific and academic guidance for the theoretical research, system design and practical application of solid sustained-release energetic materials in related fields.

The temperature response calculation of thermal protection materials, especially ablative thermal protection materials, usually adopts the ablation model, which is complicated in process and requires a large amount of calculation. Especially in the process of optimization calculation and parameter identification, the ablation model needs to be called many times, so it is necessary to construct an ablation surrogate model to improve the computational efficiency under the premise of ensuring the accuracy. In this paper, the Gaussian process model method is used to construct a thermal protection material ablation surrogate model, and the prediction accuracy of the surrogate model is improved through optimization.

X-ray computed tomography (XCT) has recently emerged as a powerful tool for characterizing the evolution of microstructure during phase transformation in three dimensional (3D) such as dendritic solidification of alloys. This paper briefly reviews the recent advances in the in-situ observation of aluminium alloys, magnesium alloys and nickel-based superalloys during solidification using laboratory XCT and synchrotron X-ray sources. The focus is on the growth kinetics of dendrites, porosity and secondary phases. In addition, in-situ characterization during the loading and corrosion process is also discussed.

In order to provide more insights into the damage propagation composite wind turbine blades (blade) under cyclic fatigue loading, a stiffness degradation model for blade is proposed based on the full-scale fatigue testing of a blade. A novel non-linear fatigue damage accumulation model is proposed using the damage assessment theories of composite laminates for the first time. Then, a stiffness degradation model is established based on the correlation of fatigue damage and residual stiffness of the composite laminates. Finally, a stiffness degradation model for the blade is presented based on the full-scale fatigue testing. The scientific rationale of the proposed stiffness model of blade is verified by using full-scale fatigue test data of blade with a total length of 52.5 m. The results indicate that the proposed stiffness degradation model of the blade agrees well with the fatigue testing results of this blade. This work provides a basis for evaluating the fatigue damage and lifetime of blade under cyclic fatigue loading.

To improve the operation situation of difficulty and low efficiency in the extraction of fermented grains (FG), a high-load and large-workspace reclaiming robot for ceramic cylinder fermentation is designed, and a reclaiming effector is designed according to the operating characteristics. Firstly, the kinematics and singularity of the mechanism are analyzed. A multi-domain polar coordinate search method is proposed to obtain the workspace and the volume of the mechanism. Secondly, the dynamic modeling is completed and the example simulation is carried out. Thirdly, the motion-force transmission index of the mechanism is established. And based on the global transmissibility and the good-transmission workspace, the dimensional synthesis of the driving mechanism is completed by using the performance atlas-based method. Finally, aiming at the regular workspace size, stiffness and loading capacity, the Pareto optimal solution set of the executive mechanism dimension is obtained by using the multi-objective particle swarm optimization (MOPSO) algorithm. This paper can provide a theoretical basis for the optimal design and control of FG reclaiming robot.