Optimized driving cycle simulation for automotive applications

Binapart Electronics GroupAbout BinapartBinapart Electronics GroupAbout BinapartBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart Electronics GroupBinapart

Home
News
Power management
Optimized driving cycle simulation for automotive applications

February 18 , 2023 0

Silicon carbide (SiC) has transformed power transmission in many industries, especially in the electric vehicle (EV) charging and on-board power conversion sectors. Due to its excellent thermal characteristics, low loss and high power density, SiC offers higher efficiency and reliability than more traditional technologies such as Si and IGBT. These topologies, systems, and applications consisting of SiC must be simulated for maximum system efficiency and accurate predictive performance.

Wolfspeed EAB450M12XM3 is a vehicle gauge grade SiC power module that has been optimized for the traction inverter market. The power module has many advantages, but the designers need to simulate as many systems as possible to take full advantage of these advantages. When simulating traction inverters for electric vehicle applications, designers must strike a balance between complexity, accuracy, and simulation time. This paper will explore how to develop a model suitable for simulating inverters in automotive driving cycles where the advantages of SiC can be exploited.

（1）Device level and module level performance

There are many ways to simulate SiC traction inverter, but we divide it into four levels:

1. Emulating the semiconductor component itself (device level) is central to successfully predicting system performance.

2. The assembly or packaging of simulated devices (module level) can predict electrical, thermal and mechanical properties.

3. Simulate each subsystem (system level) to map the entire driving cycle to electrical requirements.

4. Emulating real-world application systems (application level) can provide insight into the performance of specific use cases (such as driving cycles).

The EAB450M12XM3 (shown in Figure 1) offers a number of device - and mode-level features that reduce switching losses while providing temperature feedback and voltage sensing/overcurrent detection. The device also features an easy-to-integrate size, optimized thermal management, low parasitic inductance (6.7nH) and low inductance busbar interconnection, and higher power density (60% reduction in volume and 55% reduction in size) compared to industry-standard packages.

Figure 1: Wolfspeed EAB450M12XM3 gauge class 1200V, 450 A full SiC passon optimized half-bridge module

The EAB450M12XM3 not only provides high performance, Wolfspeed scale components, but also optimized and validated for automotive applications. In addition, there are some other advantages to choosing Wolfspeed devices for designers:

1. We provide comprehensive electronic device simulation models.

2. Ability to create average models using thermal analysis.

3. Able to complete driving cycle analysis in PLECS.

4. It can simulate the impact of inverter performance on the driving range and battery cost of electric vehicles.

Lucid Air's XM3 traction drive (shown in Figure 2 with EAB450M12XM3) has achieved 74 kg of weight, more than 670 hp of power, and more than 9 hp/kg of power density. These specifications earned the Lucid Air the MotorTrend Car of the Year award (2022).

Finally, the 600 kW XM3 dual inverter reference platform (Figure 2) helps designers prototype drivetrains quickly and with ease. The reference platform provides major power density and efficiency across the entire power range (200-600 kW).

Figure 2: Lucid Air's XM3 traction drive (left) and Wolfspeed's XM3 dual inverter (right)

(2)System-level simulation and tools

Wolfspeed uses three main tools to emulate power modules: PLECS, SPICE, and FEA.

Piecewise Linear Electrical Circuit Simulation (PLECS) allows the modeling and simulation of complete power electronics systems, including power supply, conversion, and load. In addition to the electronic domain, it helps model physical domains such as thermal, magnetic, and mechanical, providing a comprehensive system-level assessment. It AIDS in device selection and, although it has many practical functions, may oversimplify some aspects, such as the switching state of the device. For inverter applications, PLECS can be modeled using parameters in the data sheet to determine on-off and on-off losses, while also providing a thermal model of the thermal resistance of the crust (see Figure 3).

Figure 3: Sample PLECS model for loss and thermal simulation

Simulation Program with Integrated Circuit Emphasis (SPICE) is a commonly used open source circuit simulator that can be used to evaluate electrical properties in analog circuits. By simulating the dynamic current between parallel devices, it helps to optimize the current imbalance, which is helpful for the application of inverter. In addition, SPICE supports Monte Carlo simulation and EMC analysis. Although it is a standard platform with many possible use cases, it can be influenced by circuits (such as grid drive characteristics and parasitic effects) and lead designers down the wrong design path with misleading parameters.

Figure 4: FEA simulation example, EAB450M12XM3 thermal analysis

Finite Element Analysis (FEA) is used to break down a shape or structure into many smaller elements that are subject to calculations related to real world forces such as heat, vibration, stress/strain, and other environmental and physical influences. At the module level, it enables designers to estimate RTH, parasitic inductance, current-carrying capacity, and determine chip interactions (see Figure 4). However, it is usually expensive (cost-wise) and requires very detailed inputs and CAD models. It also has the longest emulation time compared to PLECS and SPICE.

Using any or all of these simulation platforms can help optimize efficiency, loss, and heat management. For example, the simulation of the switching characteristics of a SiC half-bridge module helps to predict the switching voltage waveform and determine the ripple. You can use 2D and 3D lookup tables (using parameters such as VDC, IAM1, TJ, and RG) to calculate power loss and on/off energy. PLECS can also plug these simulated power losses into the Cauer model with some other parameters (such as RTH and TFLUID) to model the thermal characteristics of the half-bridge. After the load and controller are modeled, the designer can proceed to complete system design to help optimize device selection, topology, thermal management, efficiency, and control methods. Figure 5 shows the general flow of a complete system-level simulation. For driving cycle simulation, load model is the key part. In this simulation, the load is a current source following the driving cycle curve.

Figure 5: Full system simulation, including electrical, thermal, load, and controller

In order to model such a system, it is necessary to understand the characteristics of the power module through tests or data tables. Here Wolfspeed provides a downloadable PLECS model for its entire portfolio of power modules.

Once electrical, loss, and thermal models have been determined, they can be used for complete vehicle simulations, including Global Uniform Light Vehicle Test Procedure (WLTP) driving cycle tests under appropriate electrical operating conditions.

(3)Simulation of inverter performance and driving cycle

At this point, we have discussed three of the four simulation architectures: device, module, and system level. These are necessary to establish a core understanding and expectation of system-level functionality in the driving cycle. While electrical operating points, thermal/electrical characteristics, loss calculations, and models can be handled in Wolfspeed, the Global Uniform Light Vehicle Test Cycle (WLTC) (the sample diagram shown in Figure 6) will specify torque, speed, acceleration, and the operating points for these parameters.

Figure 6: WLTC diagram

Let's analyze WLTP to see how it affects the underlying electronics. Under average WLTP conditions, usually 95% of the energy loss is switching loss. This typically pushes the range of the vehicle up and uses only 10% (or less) of the available thermal performance.

Maximum IOUT or WLTP peak conditions will produce maximum operating temperatures, but may still only use 30% of the available thermal performance. Normal operation usually does not encounter 30 seconds close to the peak current condition, but it may lead to the need for heat dissipation processing in the SiC region, thus increasing the inverter cost, volume and reduce efficiency.

Understanding WLTP and its requirements for system current can help designers calculate losses and optimize system configuration. Figure 7 shows how the lower current in the module drives the switching loss, while the higher current is dominated by the conduction loss. Another diagram in the figure describes how the junction temperature of the module usually increases with the current.

Figure 7: Current and loss of a module (left) and current and junction temperature (right)

Figure 8 shows the relationship between the WLTP driving cycle speed and the inverter phase current

If we take the figure above and correlate the current shown with the conversion power loss, we can see that the switching loss is much higher than the conduction loss (while the diode loss is negligible) during the WLTP driving cycle. Participate in Figure 9 below.

Figure 9. Inverter power loss in the WLTP cycle

Taking the same WLTP driving cycle and converting it to junction temperature allows designers to understand heat dissipation requirements and even predict reliability and product life.

From the simulation of the WLTP driving cycle, the comparison of switching energy loss with time due to the selection of external grid resistance can be obtained. Low RG(EXT) values (coupled with high loop parasitic inductance) may result in voltage overshoot and high dv/dt and di/dt values. This can cause module failure and may even stratify the motor windings. However, choosing a lower RG(EXT) reduces total energy consumption throughout the driving cycle, resulting in cost savings for customers and delayed battery life. This is another example of how understanding the driving cycle and simulating it in the early stages helps guide designers in selecting devices/components and optimizing inverter design while keeping the customer's interests in mind.

(4) Summary

In summary, it is important to emulate not only the underlying subsystem and system level configuration, but also the application (driving cycle), as this helps to establish a core understanding of the actual use of the system. Optimizing modules (such as EAB450M12XM3) and inverters to take advantage of SIC-based advantages such as efficiency, switching speed/loss and thermal characteristics throughout the driving cycle can extend the life of the product. Wolfspeed's dual inverter application and evaluation suite helps designers get to market faster and with confidence.