Power Electronics Simulations Help Speed Converter Prototyping

The design of a power electronics system involves the electrical circuit and the control algorithms. The goals for power density and efficiency improvements have led to the development of new topologies for power conversion. Further, Wide Band Gap (WBG) semiconductors such as silicon carbide (SiC) and gallium nitride (GaN) enable much faster-switching frequencies and slew rates with lower parasitics. Accurate modeling and simulations are important aspects of any new converter design, especially in applications using WBGs where higher switching frequencies and slew rates used for efficiency improvements can require greater attention to transient and thermal performance. In this article, we will explore some of the features of SIMBA, a simulation tool that addresses key requirements in developing power converters.

Accuracy versus Simulation Time Tradeoffs in the Modeling of Power Electronics Devices

A model is a representation of device behavior. The goal is to predict the static, dynamic, transient, thermal, and EMI effects presented by this device when used in a certain circuit, such as a power converter. Important metrics such as power dissipation, overshoot voltages, currents, time to reach stable voltages, etc., can be determined. Manufacturers of semiconductor devices usually offer free models of their devices on their websites. Detailed models that account for the physical behavior of the device across the entire operation range can accurately predict performance but usually with unacceptably long simulation times, especially when used in complex circuits. Shorter simulation times in power electronics systems can be met with simplified device models, such as the piecewise linear ones, where switches are treated as ideal devices across a switching event. Such models are typically available in the component libraries within simulation programs like SIMBA. Control loops can be subsequently simulated using these device models with reasonable accuracy and simulation times.

Accuracy for non-linear and inter-dependent effects (such as thermal resistance and junction temperature) can be improved using predictive models based on measured device performance. Such behavior can be represented sometimes by placing additional components, such as capacitors or resistors, that follow certain behavior, for example, voltage and temperature, in series or parallel with the active switch. The goal is to create piecewise linear approximations of non-linear events without impacting simulation times greatly. Simulation software vendors thus need to give the user the flexibility to enhance the simpler device models that are part of its libraries with device manufacturer or end-user data.

SIMBA Overview

SIMBA, developed by Powersys and released around two years ago, is a predictive time-step solver for power electronics circuits. It finds and uses the optimal time-step finder (OTSF) to simulate a system’s time constants and events without compromising accuracy. The OTSF algorithm is called at the beginning of the transient simulation and after every switching event. The temporal accuracy of a discontinuity such as a switching event is improved through the Next Discontinuity Event Time Estimator (NDETE) that runs parallel to the main solver. Modified Nodal Analysis in its solver with sparse matrices makes the computation time grow linearly with the number of nodes (rather than quadratically as in other simulation tools), resulting in fast simulations even for complex converter topologies.

Figure 1 shows the full suite of simulations needed in a converter design. The SIMBA platform currently simulates the system and control design features.

Figure 1: The full range of simulations and code development needed for a <a href= — Figure 1: The full range of simulations and code development needed for a power converter design, with SIMBA currently meeting the needs for system and control design (Source: Powersys)

In 2023, SIMBA added several new features, such as a multi-time step solver, an online version of its tool, improved motor models, and FFT calculations. Figure 2 shows the comprehensive feature set available with SIMBA.

Figure 2: The feature set available with SIMBA (Source: Powersys)

Scripting with Python

One of the unique features of SIMBA is the ability for users to create their scripting code in Python and use the vast collection of pre-built libraries and modules. Routine tasks can be automated, reducing human error. Tools within Python, such as NumPy and Matplotlib, can be used for numerical computation and data visualization. The aesim.simba SIMBA Python library allows users to run simulations without using the SIMBA Desktop. The addition of the online version gives users three different ways to run simulations of the SIMBA platform (desktop, using aesim.simba, or online). Important features such as parametric and sensitivity analysis can be performed in an automated with the help of Python scripts. Version control and shareability are intrinsic features of aesim.simba, with many examples available for users in the GitHub repository.

Some Examples of SIMBA Simulations

Thermal loss comparisons

Figure 3 shows the schematic of a simple DC-DC buck converter. Transient simulations were performed in SIMBA using the thermal description files provided by the device manufacturer. This allowed instantaneous comparison of loss and junction temperature across the multiple Infineon IGBTs used for the switch in this application.

Figure 3: Loss and Junction temperature comparison for <a href= — Figure 3: Loss and Junction temperature comparison for IGBTs used in a buck converter with SIMBA simulations (Source: Powersys)

Transient simulations

Figure 4(a) shows a schematic of a Flyback converter simulated on SIMBA. The resultant transient response shown in Figure 4(b) was simulated within approximately 100ms and shows stable voltage achieved at around 0.3ms from start-up.

Power Electronics Simulations Help Speed Converter Prototyping

Figure 4: (a) Schematic of the Flyback converter simulated on <a href= — Figure 4: (a) Schematic of the Flyback converter simulated on SIMBA, and (b) the resultant transient response on the output voltage (Source: Powersys)

EV Powertrain efficiency map

The electric vehicle (EV) traction inverter is a prime example of the use of silicon carbide (SiC) switching devices for improved efficiency, power density, and thermal performance, especially at light (cruising) loads. The power conversion efficiency depends on the motor load conditions. Figure 5 shows simulations performed on a 3-phase inverter: a Vbus voltage of 450V, inverter switching frequency of 50 kHz, and the Wolfspeed SiC MOSFETs case temperature used in the example at 80˚C. The motor torque and speed were varied, and an efficiency 2-D map was obtained, as shown in the figure.

Figure 5: Schematic and simulated efficiency for a SiC MOSFET-based traction inverter as a function of motor torque and speed (Source: Powersys). — Figure 5: Schematic and simulated efficiency for a SiC MOSFET-based traction inverter as a function of motor torque and speed (Source: Powersys)

Future Roadmap

The plan is to enhance and improve the suite of simulations offered on future SIMBA releases. These include:

Adding simulation for the magnetics and the EMI performance
Including more detailed model options for active devices such as MOSFETs and IGBTs, including SPICE model compatibility.

Frederique Peyret, Marketing Director at Powersys, commented: “We are excited to offer a power electronics simulation platform that offers Python scripting capability. The SIMBA platform enables fast simulation speed even for complex, parameterized simulations without compromising accuracy.”

Pre：Top 10 5G chips and modules Next：RF-SOI市场现状是这样...未来还将走向3D化