Xu Guoqing

We are seeing a pattern with GenAI and LLMs, agents, etc. Performance of experts will be amplified but total beginners will not get super productive results. So, learn fundamentals whether it's CFD or CAE or programming. Then use AI to build while checking everything with your own knowledge. I think this is the way to build/do things from now on.

32

7 Kommentare

📌 PhD milestone! Our paper “𝘚𝘦𝘮𝘢𝘯𝘵𝘪𝘤𝘢𝘭𝘭𝘺 𝘦𝘯𝘳𝘪𝘤𝘩𝘦𝘥 𝘊𝘈𝘋 𝘮𝘰𝘥𝘦𝘭𝘴 𝘧𝘰𝘳 𝘥𝘪𝘨𝘪𝘵𝘢𝘭 𝘮𝘢𝘯𝘶𝘧𝘢𝘤𝘵𝘶𝘳𝘪𝘯𝘨: 𝘢 𝘴𝘺𝘴𝘵𝘦𝘮𝘢𝘵𝘪𝘤 𝘳𝘦𝘷𝘪𝘦𝘸 𝘰𝘧 𝘮𝘰𝘥𝘦𝘭-𝘣𝘢𝘴𝘦𝘥 𝘥𝘦𝘧𝘪𝘯𝘪𝘵𝘪𝘰𝘯” is now published in the 𝐉𝐨𝐮𝐫𝐧𝐚𝐥 𝐨𝐟 𝐈𝐧𝐭𝐞𝐥𝐥𝐢𝐠𝐞𝐧𝐭 𝐌𝐚𝐧𝐮𝐟𝐚𝐜𝐭𝐮𝐫𝐢𝐧𝐠. 🔗 DOI: https://lnkd.in/dyNEQ5rV 📄 Full-text (open access): https://rdcu.be/e5OIu In this systematic review, we synthesize 20 years of research on 𝐌𝐨𝐝𝐞𝐥-𝐁𝐚𝐬𝐞𝐝 𝐃𝐞𝐟𝐢𝐧𝐢𝐭𝐢𝐨𝐧 (#MBD) and discuss how embedding #semantics into 3D #CAD models (beyond geometry) can strengthen the #digital #thread and improve information consistency across the product lifecycle. Key contributions: • a structured state-of-the-art synthesis • practical, implementation-oriented guidelines and adoption enablers • a roadmap for research, standardization, and industrial deployment • evidence of MBD’s potential for end-to-end #automation in digital manufacturing Grateful to my mentors and co-authors Michele Lanzetta and Francesco Lupi for their guidance and collaboration. Ready for the next step toward MBD-enabled automation.

28

1 Kommentar

During the early days of REINFORCE3D, I quickly became convinced of one thing: 𝐂𝐅𝐈𝐏 𝐜𝐚𝐧 𝐛𝐫𝐢𝐝𝐠𝐞 𝐦𝐚𝐧𝐲 𝐨𝐟 𝐭𝐡𝐞 𝐠𝐚𝐩𝐬 𝐭𝐡𝐚𝐭 𝐡𝐚𝐯𝐞 𝐥𝐢𝐦𝐢𝐭𝐞𝐝 𝐚𝐝𝐝𝐢𝐭𝐢𝐯𝐞 𝐦𝐚𝐧𝐮𝐟𝐚𝐜𝐭𝐮𝐫𝐢𝐧𝐠. - Strength. - Large-format parts. - Multimaterial integration. - Cost. Last week, in the lab, we proved one more: 𝐓𝐄𝐌𝐏𝐄𝐑𝐀𝐓𝐔𝐑𝐄🔥 We tested five rods under the same conditions: same geometry, same load, same oven. The results: → PA12 MJF failed at ~80°C → PETG FDM collapsed at ~90°C → ULTEM™ FDM bent at ~170°C → PAHT CF FDM + CFIP stood beyond 200°C → PA12 MJF + CFIP stood beyond 200°C Same part. Same load. Completely different outcome. 𝐓𝐡𝐞 𝐝𝐢𝐟𝐟𝐞𝐫𝐞𝐧𝐜𝐞 𝐢𝐬 𝐰𝐡𝐚𝐭 𝐫𝐮𝐧𝐬 𝐢𝐧𝐬𝐢𝐝𝐞. CFIP places continuous carbon fiber exactly where the load travels, allowing the structure to keep carrying load even when the polymer softens. Another gap in additive manufacturing - reinforced. Watch the test: https://lnkd.in/euAPk95y #Reinforce3D #CFIP #CarbonFiber #Composites #AdditiveManufacturing #AdvancedManufacturing #HighTemperature #MJF #FDM

42

2 Kommentare

From Microns to Megawatts: The Mathematics of Battery Modeling: In battery research, one question always comes up: How much physics do we really need in our model? This review — “A Continuum of Physics-Based Lithium-Ion Battery Models Reviewed” (F. Brosa Planella et al., Progress in Energy, 2022) — provides one of the most elegant answers I’ve seen. It walks through the entire hierarchy of lithium-ion battery models — from the detailed microscale model, through the Doyle–Fuller–Newman (DFN) model, all the way to the simplified Single Particle Model (SPM) — showing how each level can be systematically derived from the one above. 🔬 Key takeaway: Model fidelity isn’t about complexity — it’s about fit for purpose. In battery design, high-fidelity microscale or DFN models help capture fine transport physics. For real-time control and battery management systems (BMS), an equivalent circuit model (ECM) or reduced-order SPMs provide actionable accuracy with low computational cost. Each step down the hierarchy trades detail for speed — but all share a unified physical foundation. The paper also explores thermal coupling, degradation, and mechanical extensions, offering a roadmap toward next-generation multi-physics models for design, optimization, and digital twins. As #simulation increasingly drives #innovation, such frameworks help bridge the gap between fundamental science and real-world engineering. 📖 Reference: F. Brosa Planella et al., A Continuum of Physics-Based Lithium-Ion Battery Models Reviewed, Prog. Energy 4 (2022) 042003. DOI: 10.1088/2516-1083/ac7d31 #BatteryModeling #Electrochemistry #EnergyStorage #LithiumIon #DFNModel #PhysicsBasedModeling #ThermalModeling #BatteryResearch #ModelReduction #EnergyInnovation

38

Exciting news! 📢 My latest paper on numerical simulations of two-stroke marine engines is now live in Mechanism and Machine Theory. We’ve developed a computationally efficient framework using the nodal-based Floating Frame of Reference Formulation to capture the global dynamics of these massive engines. We validated the simulations against full-scale experimental data, acquired from the Everllence Research Center in Copenhagen. A big thank you to my colleague and co-author Anders Vølund and supervisor Ilmar Santos from the Technical University of Denmark for their support and collaboration. The paper is open-access! Check it out here: 👇 https://lnkd.in/eCVZRdRi

66

2 Kommentare

🎙️ Meet Elly - Host of Episode 3 of Simulation Fridays! This Friday, join Ely Vlajos Gomez , Application Engineer at Simtec, as she explores how to make electric motors quieter and more efficient in: “Silent Power: Reducing Noise & Vibration in Electric Motors with Ansys.” Elly is passionate about using simulation to solve real engineering challenges, especially those that make products quieter, smarter, and more reliable. With expertise across multiple physics domains, she’ll walk us through the complete NVH (Noise, Vibration & Harshness) simulation workflow - from electromagnetic excitation to structural dynamics and acoustic performance. 🌟 Key Takeaway: Discover how connecting multiple physics domains early in design can help engineers solve NVH challenges, reduce costly prototypes, and deliver quieter, more reliable products. 📅 Friday, October 17 at 14:00 CET! Don’t miss it, Register here: https://shorturl.at/pmUPa #SimulationFridays #Ansys #NVH #ElectricMotors #EngineeringInnovation #Simtec

72

1 Kommentar

I recently read an interesting paper that I strongly recommend the #CFD community to check. This paper thoroughly addresses the role of round-off error in DNS computation. We usually don't pay too much attention because we all agree that truncation error swamps round off, well, this is what we all see. However, they used the Fourier pseudo-spectral-RK4 method, which has been proven to be accurate and delivers a high-fidelity solution, and found that a quadruple-precision calculation showed not only a later transition but also a solution that differed drastically. Not surprisingly, the error induced by double precision grows much faster; in other words, the error becomes a force, so it seems that round-off error is indeed triggering the transition rather than the actual physics. Why is this important? In the era of #GPUs, where we claim solving #DNS at massive speedups, we only do so by using single-precision on GPUs. In my past experience, I did not see much speed up using double precision when porting to GPUs. Food for thought! I'd love to see this for 3D, where transition mechanisms differ, of course, and although the authors have studied the scheme using the GSA, I wonder if this requires a revision based on this new evidence. If you want the paper, reach out to one of the authors: Tapan Sengupta Paper:  https://lnkd.in/eEEWnaNA

68

6 Kommentare

#Article: "A Review of Machine Learning-Based Thermal Error Modeling Methods for CNC Machine Tools" Authors: Sen Mu, Chunping Yu, Kunlong Lin, Caijiang Lu, Xi Wang, Tao Wang and Guoqiang Fu 👉 https://lnkd.in/g2ESmQ9d Southwest Jiaotong University #thermalerrormodeling #CNCmachinetools #machinelearning MDPI

2

New research out in a great journal! I'm proud to share that our paper has been published in Structural and Multidisciplinary Optimization (Qualis A1, Impact Factor 4.0). Together with my colleagues Daniel Candeloro Cunha, Heitor Nigro Lopes, and my advisor Renato Pavanello, we present: "Imposing a connection constraint on discrete variable topology optimization of pressure-actuated compliant mechanism." Our study explores using the BESO method combined with the Virtual Flux Method (VFM) to design novel and reliable pressure-actuated compliant mechanisms; a key step for advanced soft robots and fluid-structure systems. Key takeaways: 🔹 BESO is great for generating creative topologies. 🔹 The VFM constraint is vital for stability, especially in multi-pressure scenarios. 🔹 This approach provides a solid foundation for future complex analyses. DOI: https://lnkd.in/dDixcem5 #TopologyOptimization #CompliantMechanisms #SoftRobotics #StructuralAndMultidisciplinaryOptimization #JournalPublication #BESO #EngineeringResearch

9