A system-level approach has allowed the University of Bath to deliver a simple, cost-effective solution to torque ripple in Permanent Magnet Synchronous Motors, allowing the University’s client to quickly expand its market into high-quality, light-weight electric vehicles.
Electric golf buggies may be simple compared to road-going passenger cars, but the drivers and occupants are typically very demanding. Electrification offers the potential for near-silent, zero emissions operation, but only if every aspect of driveline refinement can meet the high standards expected by the people who use them. Permanent Magnet Synchronous Motors, a popular choice for some of the world’s most highly-regarded electric passenger cars, have a lot of potential in the light off-highway vehicle market, offering a compact, high-efficiency traction solution with high efficiency and high peak power, but only if their limitations can be cost-effectively overcome.
This is the challenge that was brought to the University of Bath by an established manufacturer of electric motors used across a wide range of industrial and light-vehicle applications.
A light electric vehicle generally offers only modest damping of powertrain vibration, allowing the torque ripple inherent with Interior Permanent Magnet Synchronous Motors (IPMSM) to create a torsional vibration when running at low speed. The low frequency of this vibration can cause significant discomfort to vehicle occupants. The approach applied by the University of Bath began with a rigorous characterisation of the problem and its possible causes, allowing fast, efficient development of a mitigation strategy.
Starting at a whole-vehicle level, the golf buggy was equipped with accelerometers to characterise the powertrain-induced vibration profiles that affect driver and passenger comfort. Working with the departments specialists in Noise, Vibration and Harshness (NVH), this allowed identification of the speeds and frequencies that caused most discomfort and which should therefore be the focus of the mitigation.
Physical testing characterised a vibration that was mainly excited by the 24th harmonic torque ripple of the IPMSM and which was more severe when the torque ripple frequency is close to the natural frequency of the drivetrain, which had been identified as around 10 Hz. It was found that the amplitude of the vibration increased when the motor runs at around 25 RPM (0.41 Hz).
Building on this data, a software model was developed to test and refine various possible solutions. A multidisciplinary approach, calling on specialists in driveline, electric motors, motor contol and power electronics, ensured that each stage of investigation encompassed all relevant aspects of the system and their critical interactions.
The powertrain has a nonlinear dynamic response with a natural frequency that is mainly determined by the half shaft, rotor inertia and vehicle inertia. It is therefore necessary to take into account both the powertrain dynamic response and the frequency of the machine’s torque ripple. While it is undoubtedly possible to reduce the effect of torque ripple through improvements in machine design, and several design revisions were recommended, it was decided that managing the cause of the ripple would be a more elegant and more cost-effective solution.
Mitigating cogging torque
The University of Bath concluded that the optimum solution was to reduce the torque ripple at source by changing the control strategy. When the speed increases to above 30 RPM, the 24th harmonic speed ripple can be mitigated by additional calibration of the existing Proportional Integral (PI) controller. However, the vehicle-level study had revealed that because the vehicle inertia would also reduce the controller authority, the conventional PI controller did not offer sufficient bandwidth to sufficiently address the problem at lower speeds where the effect of torque ripple is felt most by the vehicle occupants.
The solution developed by the Bath team employed a novel application of proven Resonant Controller techniques. A new Resonant Controller (RC) was developed to operate in parallel with the conventional PI controller to increase the gain at the speed ripple frequency. This novel application of a proven technique allows very fast modulation of the control signal, requesting more or less torque in real-time to cancel the ripple. The result of this new approach is that the 24th harmonic speed ripple was reduced by around 90 percent.
New power electronics
To deliver the new control strategy, the University’s multidisciplinary team developed a new set of control algorithms and a new inverter that brought a range of other benefits. Sensing was improved and the sampling frequency was increased to improve the linearity of response, providing further improvements in vehicle refinement and driveability.
By solving a major barrier to customer acceptance, the University of Bath helped its client introduce a highly competitive new product into a fast-growing market. The Permanent Magnet motor can now directly replace an asynchronous induction machine, bringing benefits in durability and efficiency (which means fewer charges and increased range) without compromising refinement or controllability.
Low speed torque ripple has proved to be a barrier to the adoption of high-efficiency PM motors across a wide range of applications. From lifts to fork-lift trucks and light electric vehicles, frequencies below around five Hz can cause significant discomfort. The novel application of proven theories has placed the University of Bath at the forefront of resonance control for Permanent Magnet motors, giving the team a unique ability to quickly and affordably mitigate refinement issues and accelerate the adoption of these high-efficiency motors across a wide range of applications.