SIMBA, Power Electronics Simulation Software

🚀 SIMBA 26.09 is here! Our latest release focuses on what matters most to engineers: making everyday modeling, simulation, and analysis workflows smoother, faster, and more reliable. Among the highlights: ✅ New Three-Phase PLL (SRF & DDSRF) with automatic controller tuning ✅ 2D Ld/Lq machine modeling for PMSM, BLDC and SynRM ✅ Expanded Python API coverage and refreshed documentation ✅ Improved JMAG co-simulation robustness ✅ GPT-5.6 powered SIMBA Assistant ✅ Numerous stability, usability and charting improvements One of my favorite additions is the new Sampling Time Highlighting feature shown below. Simply hold "T" in the design view and SIMBA instantly highlights devices according to their sampling times. Different colors make it much easier to identify control loops operating at different rates and spot potential configuration issues in large models. When developing power electronics systems, small usability improvements like this can save significant debugging and validation time. SIMBA 26.09 also introduces dozens of workflow enhancements across: 🔹 Modeling 🔹 Control Design 🔹 Python Automation 🔹 Electrothermal Simulation 🔹 JMAG Co-Simulation 🔹 Documentation & Support Download the latest version and discover everything that's new. #PowerElectronics #Simulation #DigitalTwin #MotorDrive #ControlSystems #Python #JMAG #EngineeringSoftware #SIMBA Guillaume Fontes Jérôme Cornau Ryoko Imamura Lisa Pintori Olivier TOURY

🚀 What if motor-inverter co-simulation could run 4× faster without sacrificing accuracy? Our latest paper, accepted at IEEE ECCE 2026, demonstrates a new multi-rate co-simulation strategy based on the well-known small-ripple approximation. Instead of running expensive electromagnetic FEA calculations at every inverter time step, the method executes FEA around switching events and reconstructs the intermediate behavior through interpolation. For an 80-kW PMSM drive system, the results were striking: ✅ Simulation time reduced from 40 min to <10 min ✅ 75% reduction in computational cost ✅ Only 0.32% average error ✅ 2.7% maximum error An interesting observation was that conventional fixed-interval subcycling can miss current ripple peaks, while synchronizing FEA calculations with switching events preserves accuracy even under aggressive subcycling conditions. Faster co-simulation means more opportunities for: • Motor-inverter co-design • Design-space exploration • Optimization studies • AI-driven engineering workflows Looking forward to presenting this work at IEEE ECCE 2026 (Session 77: Simulation, Surrogate Models, and AI in Motor Design) in Vancouver. See you there! #ECCE2026 #MotorDrives #ElectricMachines #PowerElectronics #FEA #CoSimulation #EMobility #Engineering Guillaume Fontes Ryoko Imamura Jérôme Cornau Olivier TOURY Lisa Pintori

🚀 Can AI design a 98% efficient power converter in just 30 simulations ? Researchers from National Taiwan University of Science and Technology (NTUST) demonstrated how Bayesian Optimization, combined with SIMBA, can automatically explore thousands of real-world design options, including MOSFETs, magnetic components, switching frequencies, and winding configurations. 📊 Out of 1,920 possible combinations, the optimizer identified a design exceeding 98% efficiency after only ~30 simulations. This research highlights the potential of AI to drastically reduce design cycles while working directly with commercial components and real datasheet parameters. The future of power electronics is not just faster simulation. It's AI-driven design exploration. #PowerElectronics #AI #BayesianOptimization #SIMBA #Engineering #Optimization #MachineLearning #PowerConverter NGUYEN TAN TUNG Olivier TOURY Ryoko Imamura Jérôme Cornau Guillaume Fontes

A converter can still regulate perfectly while already failing its lifetime requirements. In a recent SIMBA study, we looked at a simple buck converter between Beginning of Life (BOL) and End of Life (EOL). Only one parameter was changed: • output capacitor: 100 µF → 40 µF What surprised us was not what changed. It was what didn't. • output voltage remained close to 25 V • inductor current ripple remained nearly unchanged • the converter still operated normally Yet output-voltage ripple increased from 665 mVpp to 1573 mVpp. Depending on the specification, the design can therefore move from compliant to non-compliant while still appearing perfectly healthy from a top-level perspective. That is why BOL/EOL analysis matters. The objective is not simply to check whether the converter still runs. It is to identify which design margins disappear first as components age. In the article, we also compare the SIMBA results with the classical approximation: ΔV ≈ ΔIL / (8fₛC) and show very close agreement. Full article: link in first comment #PowerElectronics #SIMBA #BuckConverter #CapacitorAging #LifetimeAnalysis #DesignSpaceExploration Olivier TOURY Guillaume Fontes Jérôme Cornau Lisa Pintori

⚡ Accelerating DAB converter design with Bayesian Optimization, FEA, and circuit simulation How can engineers efficiently optimize a critical design parameter such as the series inductance of a Dual Active Bridge (DAB) converter when every evaluation requires both circuit simulation and electromagnetic FEA? Researchers from the Advanced Power Electronics Lab at NTUST addressed this challenge by combining: ✅ SIMBA for power electronics simulation ✅ ANSYS Maxwell for electromagnetic FEA ✅ Python automation ✅ Bayesian Optimization using an Infinite-Width Bayesian Neural Network surrogate model The workflow integrates converter-level and magnetic-component evaluation into a single automated optimization loop, enabling efficient exploration of the design space while accounting for both electrical and magnetic losses. 📊 Key results: • Optimal design identified after only a few optimization iterations • 97.735% simulated efficiency • 98.38% peak efficiency measured on a 5 kW hardware prototype This work is a great example of how simulation-driven design optimization can reduce engineering effort while maintaining strong physical accuracy through the combination of: 🔹 Circuit simulation (SIMBA) 🔹 Electromagnetic FEA (ANSYS Maxwell) 🔹 Bayesian Optimization (BoTorch/PyTorch) 🔹 Automated Python workflows We are proud to see SIMBA supporting advanced research in design automation and power electronics optimization. Congratulations to the NTUST research team for this excellent work. 📄 Read the full publication on SIMBA: see full article in first comment. #PowerElectronics #DAB #BayesianOptimization #DesignAutomation #Simulation #Electromagnetics #ANSYSMaxwell #SIMBA #EVCharging #PowerConverterDesign Olivier TOURY Guillaume Fontes Jérôme Cornau Lisa Pintori Bryan M.H. Pong

🚀 Before launching a large optimization campaign, do you know which parameters actually matter? 📊 That is where sensitivity analysis becomes useful. Using real SIMBA transient simulations, we converted simple one-parameter sweeps into a practical sensitivity ranking. 🔍 A few results stood out: • ⚡ Switching frequency dominated output-voltage ripple more strongly than expected • 🧲 Inductance remained the primary driver of inductor-current ripple • 🎯 Load resistance was almost invisible for the selected metrics at this operating point ✅ This is exactly why sensitivity analysis should come first. 📈 Before Monte Carlo analysis. 📈 Before design-space exploration. 📈 Before optimization. 💡 First identify the variables that move the design. Then focus simulation effort where it creates value. 📖 Article: Sensitivity Analysis Using SIMBA: A Practical Workflow Before Optimization #PowerElectronics #SIMBA #SensitivityAnalysis #DesignOptimization #MonteCarlo #Python #DCDC Olivier TOURY Jérôme Cornau Guillaume Fontes Lisa Pintori Ryoko Imamura

Same topology. Same 600 V output. Yet current distortion varies from 18.7% to 5.4%. Using the official SIMBA Totem-Pole PFC example, I swept the input inductor and DC-link capacitor across six transient simulations. Results: • Current distortion: 18.7% → 5.4% • DC-bus ripple: 50.7 Vpp → 12.6 Vpp • Power factor: 0.975 → 0.998 The lesson is simple: A single simulation can tell you whether a design works. A design-space exploration tells you whether it is the best design. Before hardware locks in a compromise, it is worth understanding how the operating point moves across the design space. The same workflow can then be extended to: • CCM vs CRM operation • Switching-frequency optimization • SiC MOSFET selection #PowerElectronics #PFC #TotemPolePFC #Simulation #DesignSpaceExploration #SIMBA Ryoko Imamura Guillaume Fontes Jérôme Cornau Olivier TOURY Lisa Pintori

⚡ A PSFB that works at nominal conditions may not work everywhere A Phase-Shifted Full Bridge converter can meet specifications at its nominal operating point. But across the full input-voltage range, the same design can move from an acceptable operating region into transition, and eventually into degraded performance where ripple targets are no longer met. This map was built from real SIMBA transient simulations by sweeping input voltage and output inductance. Instead of validating a single operating point, engineers can identify where the converter continues to meet specifications and where operating margins begin to disappear. #SIMBA #PowerElectronics #PSFB #DCDC #DesignSpaceExploration Guillaume Fontes Jérôme Cornau Lisa Pintori Ryoko Imamura Olivier TOURY

❤️ Can a power electronics simulator draw a heart? Absolutely. This heart was not created in a graphics tool. Every point comes from a real SIMBA simulation. Using the SIMBA Python API, we automated a parameter sweep of inductance and switching frequency, then mapped the resulting operating points onto a heart-shaped trajectory. A fun experiment, but it illustrates something much more valuable: ✅ Full Python control of SIMBA ✅ Automated parameter sweeps ✅ Custom optimization workflows ✅ Large-scale design-space exploration ✅ Seamless integration with data science toolchains These are the same capabilities engineers use every day to optimize converter designs, evaluate trade-offs and accelerate development. The heart is just a reminder that once a simulator becomes programmable, the limit is no longer the GUI. It's your imagination. #PowerElectronics #Python #Automation #Simulation #Optimization #Engineering #SIMBA Guillaume Fontes Ryoko Imamura Jérôme Cornau Olivier TOURY Lisa Pintori

⚡ Same drive. Same operating point. Different thermal story. Most PWM comparisons focus on efficiency. But efficiency alone does not tell you where the heat goes. In our latest SIMBA article, we compare three PWM strategies on the same PMSM drive: ✅ SPWM ✅ SVPWM ✅ DPWM At 4000 rpm and 20 Nm, SPWM and SVPWM deliver similar total losses. What changes is the loss distribution. The same drive can shift its thermal burden between: 🔹 the inverter 🔹 motor copper losses 🔹 motor iron losses Depending on the operating point: • High torque → copper losses dominate • High speed / light load → inverter and iron losses become more significant This is why PWM strategies should not be selected based on a single efficiency number. They should be evaluated against: ✔️ the mission profile ✔️ thermal constraints ✔️ cooling limitations ✔️ the subsystem closest to its thermal limit With SIMBA, these trade-offs become visible within the same model and simulation workflow. Because the best PWM strategy is not always the one with the lowest losses. It is the one that puts the heat where your system can handle it. #SIMBA #PowerElectronics #MotorDrives #PWM #SVPWM #DPWM #ThermalAnalysis #ElectricMachines #JMAG Olivier TOURY Jérôme Cornau Ryoko Imamura Lisa Pintori Guillaume Fontes