Sebastian Ciceo, Technical University of Cluj-Napoca
Evaluating two different topologies of low-cost reluctance machine-based electric drives used in low-powered electric vehicle (EV) applications applying a real-time co-simulation approach with minimal development time.
Developing the control algorithm and simulating the e-drive in real time using FPGA modules, and providing test stimuli through an LMS Imagine.Lab Amesim EV model running in VeriStand on PXI hardware communicating with the simulated e-drive through a CAN bus interface.
Sebastian Ciceo - Technical University of Cluj-Napoca
Hunor Nagy - Technical University of Cluj-Napoca
The Department of Electrical Machines and Drives at the Technical University of Cluj-Napoca focuses on fundamental and applied research of electromechanical technologies in areas like sustainable transportation.
Electric vehicles (EVs) are a promising alternative for internal combustion engine-equipped vehicles in the automotive industry. The electric vehicles are constantly gaining share of the automotive market. Most of the low-powered EVs (such as four-wheel scooters) come with brushed or brushless DC motors. The brushed versions have low power density and rely on brushes that must be periodically maintained. The brushless versions have higher manufacturing costs due to the reliance on permanent magnets, so low-priced alternatives need to be developed. For this project, we used machines working on the principle of magnetic reluctance, the synchronous reluctance motor (SynRM) and the switched reluctance motor (SRM). The advantage of reluctance machines is their design simplicity and lack of permanent magnets or rotor windings, which provides a cheap and viable alternative for the classical electric drives.
Our challenge is designing new electric drive concepts with their respective control in a time-efficient, reliable, and safe manner. Also, we must determine the interaction between the electric drive and the EV system before building the physical prototype. An electric drive goes through multiple repeated design, simulation, and testing phases before entering the production stage. By means of real-time simulation and testing, we intend to reduce the development time and cost of new electric drives used for low-powered EV applications.
By decoupling the control and analytical representation of the electric drives from the system model of the small EV in a model-based system approach through real-time simulation, we can:
The presence of the salient poles in the stator and the rotor of the SRM determine highly nonlinear magnetic characteristics. We used electromagnetic FEM computations to map the mentioned characteristics and stored them in the form of look-up tables in the block memory of the FPGA for fast access because the data must be retrieved at every time-step. Also, since the drive operation is based on magnetic reluctance and hysteresis current control, we needed high sampling frequency of above 100 kHz. We achieved the fast computation times by programming with the LabVIEW FPGA Module. Also, by employing visual programing for the FPGA, we can reduce the control development time in contrast to programming in hardware descriptive languages. Prior to the LabVIEW FPGA deployment, we tested the drive models using the LabVIEW Control Design and Simulation Module, which further reduced development time.
The main drawback of the SRM drive is the high torque ripple. Therefore, we compared it with the SynRM, which due to the distributed windings in the stator and a non-salient rotor structure, has a smoother mechanical operation.
The real-time controller on the PXI handles time non-critical tasks such as parameter identification and CAN bus communication through the NI-XNET API. We use the FGPA module for the control algorithm deployment and for modelling of behaviour of the motor and power electronics (designed as ideal switches).
We created the EV physical-based model in LMS Imagine.Lab Amesim, which is an integrated simulation platform for multi-domain mechatronic systems simulation. We integrated it into VeriStand software (running on the second PXI) as a third-party simulation model. The forward-facing EV model sends torque setpoints coming from the vehicle control unit to the electric drive under test and imposes the speed to the traction motor defined by a specific drive cycle. The electric drive responds to the test stimuli with the computed torque produced by the traction motor at the given speed setpoint, sending it back to the EV model. As mentioned, the communication is done using CAN frames.
We can evaluate the interaction of the electric drive (component-level) with the EV (system-level) to determine different performance characteristics of the EV, such as the energy consumption of the simulated vehicle. This provides good insight into energy storage (batteries) sizing. Also, we can evaluate the effects that the torque ripple produced, or the electric drive control strategy has on the driver comfort, in a real-time simulation using a planar vehicle model. Because the simulated model runs on the PXI platform, we can achieve high system dynamics by employing a high target rate in VeriStand.
By means of real-time simulation, we can introduce artificial faults in the electric drive and assess the system behaviour without causing any physical harm to the tested components.
We intend to replace the analytical model of the power converter and electric machine with the real components while keeping the same control and EV system model to perform mechanical-level HIL validation testing.
Technical University of Cluj-Napoca
Technical University of Cluj-Napoca Observatorului Street Nr. 2.