The University of Osaka · Engineering
Professor Zhongyuan Feng's research lab specializes in advanced welding materials and thermal-mechanical modeling, focusing on low transformation temperature (LTT) weld metals to enhance residual stress control and fatigue performance in welded structures. The lab develops innovative LTT welding materials with tailored martensite start temperatures and solidification behaviors to prevent cracking and improve mechanical properties in various welding positions. Through integrated experimental and numerical approaches—including finite element analysis, synchrotron XRD, and in-situ phase evolution observation—the lab advances computational efficiency and predictive accuracy in welding process simulation. Their work bridges materials design, microstructure control, and structural integrity, particularly for critical engineering applications.
Figures are computed from collected data and may differ slightly.
We investigated whether low transformation temperature (LTT) welding materials are beneficial to the generation of compressive residual stress around a weld zone, thus enhancing the fatigue performance of the welded joint. An experimental and numerical study were conducted in order to analyze the residual stress in multi-pass T-welded joints using LTT welding wire. It was found that, compared to the conventional welded joint, greater tensile residual stress was induced in the flange plate of the
Abstract Finite element analysis is commonly used to investigate the thermal-mechanical phenomena during welding. To improve the computing efficiency of finite element analysis for welding thermal conduction, a novel Newton – Raphson method (NRM) without the computation of inverse matrix and a hybrid method combing the NRM and conventional implicit method (IMP) were developed. Comparison of computing time between the hybrid method implemented in an in-house software JWRIAN and the IMP used in a
Two types of low-transformation-temperature weld metals were devised, one associated with primary austenite solidification, the other primary ferrite solidification. The martensite start temperature of both low-transformation-temperature weld metals was about 125°C. Experimental results showed that low-transformation-temperature weld microstructure associated with primary austenite solidification was martensite with 8.0% retained austenite, whereas that one related to primary ferrite solidificat
A new low transformation temperature (LTT) welding material 16Cr8Ni was developed to satisfy the optimal characteristics in diluted welds due to all repair welding positions. Its good weldability for all welding positions was tested. The measured Ms temperature of weld metals with different dilutions was between 150°C and 250°C which is the optimal range for compressive residual stress generation and fatigue life extension. The in-situ observation of phase evolution by the synchrotron-based X-ra
Two low-transformation-temperature weld metals, 12Cr6Ni and 8Cr10Ni weld metals, were devised and their martensite start temperatures were 245°C and 225°C, respectively. Despite a small difference of 20°C in martensite start temperature, the martensite finish temperature of the former was 100°C, while the one of the latter was below room temperature. The difference in martensite finish temperature was due to the stacking fault energy. Lower stacking fault energy accelerated martensitic transform
In rational decision-making processes, the information interaction among individual robots is a critical factor influencing system stability. We establish a game-theoretic model based on mutual information to address division of labor decision-making and stability issues arising from differential information interaction among swarm robots. Firstly, a mutual information model is employed to measure the information interaction among robots and analyze its influence on the behavior of individual ro
Transformer model has achieved excellent results in many fields, owing of its huge data volume and high precision requirements, the traditional analog compute-in-memory circuit can no longer meet its needs. To solve this dilemma, this paper proposes a digital compute-in-memory circuit based on the improved Booth algorithm. The 6T SRAM array stores the multiplicand, and the multiplier is encoded by the booth encoder, and then, local computing cell (LCC) read the corresponding value from the array
A Correction to this paper has been published: https://doi.org/10.1007/s00170-020-06437-w
Open papers in the app to read, cite, and organize with AI.
