Digital Lab Simulation Tools

Computational software While I was a PhD student and later a postdoc, I had to adapt to the lab culture—and to the computational tools preferred by the PI. Once I started my own lab, I realized something important: it’s not about specific codes, but about covering methods. Over time, I converged to a core toolkit—essentially four pillars—that I believe every computational materials science lab should have: 1. Plane-wave DFT (periodic systems) For accurate and reliable simulations of solids, surfaces, and interfaces. My choice is VASP for its efficiency, but alternatives like Quantum ESPRESSO or CASTEP work equally well. 2. Gaussian-type orbital methods (finite systems) Ideal for nanoparticles, clusters, and molecules, including spectroscopy and excited states. I use Turbomole for its speed and convergence, but ORCA and Gaussian are also excellent. 3. Numerical atomic orbital (NAO) DFT Extremely efficient scaling for large systems within pure DFT. Commercial options exist (e.g., DMol³), but SIESTA provides a powerful free alternative, including transport via TranSIESTA. 4. Classical force-field simulations For large-scale dynamics and reactive simulations (e.g., ReaxFF). LAMMPS is, in many ways, a method in itself. ⸻ If you combine: Quantum ESPRESSO + ORCA + SIESTA + LAMMPS you can build a fully functional computational lab at essentially zero cost. As a bonus, this setup naturally teaches: • periodic vs. finite systems • plane waves vs. localized orbitals • quantum vs. classical mechanics • and, inevitably… Linux, scripting, and compiling 🙂

Here is a comprehensive follow-up post building on our previous updates. This time, a 15-minute video is provided, showing the entire workflow to go from nothing to having a HiL-based Driver in the Loop (DiL) setup. This video covers each step in detail, providing a thorough guide for anyone interested in creating their own simulation environment. #Engineering #DrivingSimulator #Tutorial The video is attached to this post for your convenience. You can also download the models and all files used to generate this example from here: https://lnkd.in/eU6bfYpX The fact that these tools are all (apart from Unreal & Google Earth) part of the MathWorks ecosystem meant that this comprehensive simulator could be made in my free time by utilizing the easy workflows and tool integrations between these tools (#MATLAB, #Simulink, #Simscape, RoadRunner, Speedgoat). As always, stay tuned for more updates, and feel free to reach out if you have any questions or feedback! Video Content Breakdown: - 0:00 - Intro - 0:27 - Google Earth: Used to define the inner and outer limits of the circuit. - 1:05 - MATLAB Live Script: Used to obtain elevation data from online sources, export RoadRunner-compatible files, calculate the minimized curvature optimal racing line, make a lap-time estimation, and export the optimal racing line back to Google Earth. - 3:53 - RoadRunner: Import the generated elevation and road files into RoadRunner, add grass and scenery, trees and buildings, and export to Unreal Datasmith. - 5:15 - Unreal: Open in Unreal, add HUD and minimap, any other blueprints needed, and compile to EXE. - 7:29 - Simscape Vehicle Model & Simulink ECU Model & Unreal-Simulink Integration: Quick walkthrough. - 9:08 - MATLAB Live Script: Used to keep track of high scores and changing settings on the car model, as well as what type of simulation (HiL/MiL) and map to run. - 11:14 - Unreal EXE: Example DiL run on Kyalami. - 12:46 - MATLAB Live Script: Post-processing results and performance. - 13:47 - MATLAB & Unreal: Example of changing vehicle size and tire parameters.

Digital twins are well-established in manufacturing, but can they help us in the research lab? Back when digital twins were all the buzz a few years ago, I sat through a wave of vendor pitches. Most promised impressive outcomes: predictive maintenance, plant optimization, operational monitoring, but not a single example truly applied to research. That’s why I was excited to see a recent paper by Jin Qian, Ethan Crumlin, et al. in Nature Computational Science deliver a concrete example of a digital twin tailored for chemical research. The paper begins by clarifying the objective of a research-oriented digital twin: to help scientists test hypotheses and understand mechanisms. To demonstrate this, the authors introduce a system called Digital Twin for Chemical Science (DTCS), a platform that: 🔹Simulates how a proposed reaction mechanism would behave under experimental conditions using physics-based models and reaction networks 🔹Compares those simulations with real experimental data to assess whether the mechanism explains what’s observed 🔹Uses AI to iteratively refine the mechanism or suggest new experiments when discrepancies arise This setup creates a bidirectional feedback loop between theory and experiment. Researchers can simulate expected outcomes before running an experiment, update their understanding during data collection, and analyze mechanisms afterward. It also structures the discovery process, making it more systematic, adaptive, and transparent. But wait—how is this different from a self-driving lab (SDL)? While both rely on closed-loop automation, they serve distinct roles. SDLs are built to explore large parameter spaces and optimize outcomes. DTCS, on the other hand, is focused on mechanistic insight: understanding why a system behaves as it does. In fact, these two approaches could be highly complementary. A SDL can rapidly explore large design spaces to find unexpected or optimal outcomes. Then DTCS can help uncover why it works, providing mechanistic insight to enable generalization and guide theory-driven design. I hope this paper helps others better visualize what a research digital twin can actually look like, and how it might fit into your scientific workflows. 📄 Digital Twin for Chemical Science: a case study on water interactions on the Ag(111) surface, Nature Computational Science, August 27, 2025 🔗 https://lnkd.in/evTfsErH

Digital process twin is how pharma R&D stops re-running the same experiments and starts scaling with confidence. If you’re guiding the pipeline and the process, this is the tool that turns lab learning into reliable production.     I see three pressures collide in tech transfer: fragmented datasets, teams working in silos, and processes that behave differently at each scale. That’s why a digital thread matters. It keeps recipes, models, and sensor data connected so you can test, learn, and update in one place rather than across ten spreadsheets and a month of meetings.     What does this look like in practice? Build a model that mixes physics and process logic. Use CFD to understand fluid dynamics, heat, and mass transfer at lab and plant scales. Add DEM when particle behavior and segregation threaten yield. Create reduced order models so your team gets instant predictions for temperature fields, concentrations, and stress profiles under any boundary condition. Then make it executable by tying in online sensor data and a simple feedback loop for adaptive control.     Heads of R&D will recognize the impact in the handoffs. The material in the attached brief shows how pharma teams use process simulation to identify critical process parameters and connect them to critical quality attributes, run in-silico disturbance tests, and reduce dependency on repeated physical trials. It also outlines how an executable twin brings real-time visualization for operators, supports training, and helps cut waste during vaccine and drug manufacturing. The thread is the same: simulate, simplify, and scale.     Practical takeaway for this week: pick one high-variance unit operation and stand up a minimal digital process twin for it. Connect the data, run a sensitivity study, and align on the few parameters that move quality. When that twin is stable, extend it to the next step in the process. 

I recently received a question about the tools used for the attached simulation. I previously highlighted that I´m using a full open-source workflow, but I didn't actually list the tools.   Some time ago, I regularly posted about open-source simulation tools, but I missed writing a summary for this CFD simulation. Here is the full list of tools used:   Salome Platform – Salome is a toolbox that includes geometry and mesh modules and can act as a GUI for some solvers. I have used Salome to generate a mesh from the input geometry and export a .MED file that can be read by code_saturne Code Saturne – is a CFD FVM solver that can handle several flow types and includes a variety of turbulence modules. Large simulations can be parallelized. BVTKNodes and Blender – Blender is a 3D modelling, animation and rendering tool. With the BVTKNodes plugin, it can also be used to visualize VTK solver outputs for stylized renderings. Paraview can be used for this purpose too, providing a more intuitive way to navigate the visual toolkit's filters and manipulators.   #simulation #visualization #engineering

I’ve compiled a list of simulation rich websites that I regularly use in my teaching. Around 15 of them I use actively, while the others I explore occasionally, depending on the #topic and #learning objectives. The choice of platform always depends on the #concept I want students to explore and the type of simulation that best supports understanding. I’ve also indicated which subjects each platform is best suited for. I began exploring simulations when I was unable to conduct laboratory experiments due to limitations in the lab facilities at the schools where I worked. I found these tools incredibly useful and began designing activities around them to support student learning. If you know any other effective simulation tools, please feel free to share, I’m always happy to explore new resources. 1. PhET Interactive Simulations, University of Colorado Boulder 📚 Subjects: Physics, Chemistry, Biology, Mathematics, Earth Science 2. LabXchange 📚 Subjects: Biology, Biotechnology, Life Sciences 3. JavaLab 📚 Subjects: Physics, Chemistry, Mathematics 4. Labster 📚 Subjects: Biology, Chemistry, Physics, Health Sciences 5. GeoGebra 📚 Subjects: Mathematics 6. Falstad Math & Physics 📚 Subjects: Physics, Mathematics 7. Gizmos (ExploreLearning) 📚 Subjects: Science (Biology, Chemistry, Physics), Mathematics 8. ChemCollective 📚 Subjects: Chemistry 9. HHMI BioInteractive 📚 Subjects: Biology 10. Desmos 📚 Subjects: Mathematics 11. Wolfram Demonstrations Project 📚 Subjects: Mathematics, Physics, Engineering, Computer Science 12. Learn Genetics (University of Utah) 📚 Subjects: Biology, Genetics 13. TryEngineering 📚 Subjects: Engineering, STEM 14. CK-12 Foundation Simulations 📚 Subjects: Physics, Chemistry, Biology, Mathematics, Earth Science 15. EduMedia 📚 Subjects: Physics, Chemistry, Biology, Earth Science, Mathematics 16. oPhysics 📚 Subjects: Physics 17. UCAR Center for Science Education (SciEd) 📚 Subjects: Earth Science, Climate Science, Meteorology 18. BioDigital 📚 Subjects: Biology, Human Anatomy, Health Science 19. Merge 📚 Subjects: STEM (Science, Technology, Engineering, Math) 20. Explerify.com 📚 Subjects: General Science #Teacher #ScienceTeacher #BiologyTeacher #EdTech #EdTechSpecialist #InstructionalCoach #Simulations #Education #STEMeducation #ScienceEducation #EducationalTechnology #DigitalLearning #InteractiveLearning #VirtualLabs #InquiryBasedLearning #TeachingTools #TeacherResources #Teaching #21stCenturySkills LinkedIn #Singapore