No, Title
|
SCD 4.1: AI controlled redundant powertrain
|
|||
Leader
|
ZF
|
|||
Contributing Partners
|
BUT, TUDO, MBAG, IFAG, IFAT, IFI
|
|||
Description
|
Fail-operational 800V powertrain based on a three-level Active Neutral Point (ANPC) topology which can operate as a two-level inverter in case of a fault, and allows the use of 650V GaN power transistors in an 800V system.
Fault detection is achieved with several hardware and software (SW) components. Driver and FPGA boards perform data acquisition for controlling transistor deadtime and short-circuit faults. An add-on board monitors for electrical arcing in the HV system. These boards interface with SW-based cognitive diagnostic system which also monitors the health-state of the electric motor and defines system reactions to prevent further damage in case of faults.
The diagnostic system is executed on a new generation AURIX with a parallel processing unit (PPU), speeding up the calculations of AI systems. Further components are a new generation power management integrated circuit (PMIC) and smart rotor position sensor, both of which are capable of performing self-diagnosis.
|
|||
Deployment/utilization
|
The demonstrator has been specified based on standard series requirements for 800V automotive powertrain applications. The power core, i.e. power board with GaN transistors, DC-link capacitor board, and gate-drivers, has been tested with double-pulse experiments up to voltages and currents of 920V and 180A respectively, where fast and low-oscillation switching has been confirmed. The full inverter including closed-loop control has been tested together with an electric motor on a testbench from MBAG, where functionality in six-phase and three-phase operation has been demonstrated.
|
|||
Pictures/visuals with titles
|
![]()
|