More than a century ago, a race between the internal combustion engine vehicles (including petrol and diesel-powered cars) and electric vehicles that helped to shape the future of transportation occurred. The internal combustion engine vehicles emerged victorious in that race, which is why we have used them for so many following decades. Had electric vehicles won that early contest, the automotive landscape might look very different today. But it is heartening to see the renaissance of electric vehicles (EVs) as a sustainable and environmentally responsible alternative.
At the heart of the electric vehicle lies the traction motor. That is why, the efficiency of the motor is critical for maximizing the overall energy efficiency and driving range of EVs. Power losses within the motor, such as copper losses, iron losses, friction losses, etc. adversely affect vehicle performance and energy efficiency. Therefore, effective loss minimization strategies are a must.
One of the main techniques where loss minimization is crucial is regenerative braking (RB). This is a key technique for enhancing EV driving range and performance by recovering kinetic energy during braking.
What is regenerative braking? In conventional brakes, when a vehicle slows down, the kinetic energy is dissipated as heat due to friction between brake pads and the wheel. However, in regenerative braking, instead of wasting this energy, the electric motor works in reverse as a generator when the vehicle decelerates. It converts the vehicle’s kinetic energy into electrical energy and recovers it in the battery.
Consider an example when you are driving a car downhill, and you release the accelerator. The traction motor now switches to generator mode, charging the battery while slowing the car down. This is regenerative braking.
Several approaches have been developed to improve the efficiency of regenerative braking systems (RBS). Control-based approaches aim to optimize existing braking algorithms, often by adjusting the brake force distribution between friction and RBS. Techniques such as fuzzy logic control, artificial neural networks (ANN), and torque (torque is a rotational force that causes an object to rotate about its axis) allocation strategies have been successfully demonstrated to improve energy regeneration efficiency.
One of the main problems with regenerative braking, is its reduced effectiveness at low speeds. For this, the low-speed cutoff point (LSCP) is taken into consideration for disabling RB.
However, this approach often fails to account for varying operating conditions such as motor drive parameters and battery state of charge (SoC). Dynamic approaches for detecting LSCP, such as monitoring dc-link current flow to determine the direction of power flow, have been proposed but face challenges related to sensor accuracy and delays in detection.
Therefore, to address these limitations, in this study, the authors Ms. M. K. Deepa, Prof. Srikanthan Sridharan, and Prof. Shankar C. Subramanian from the Department of Engineering Design, Indian Institute of Technology (IIT) Madras, Chennai, India, have taken an analytical approach to calculate the LSCP value.
The aim of this approach is to extend the RB period by minimizing traction drive losses (a traction drive is the complete system used to produce and control the motion of a vehicle by converting electrical energy into mechanical energy at the wheels) , thereby achieving a lower LSCP. Although some computational effort is involved with this approach, it can be easily done offline with no burdensome computation requirement on-the-go, helping to enhance the vehicle’s overall energy efficiency.
Along with the calculation of LSCP, another aspect is taken into consideration, called the motor flux. The motor flux is the magnetic field produced inside a motor that enables energy conversion from electrical energy to mechanical energy and vice-versa, inside the vehicle.
Model-based controllers are used to determine the appropriate flux level, but they are sensitive to parameters such as stator and rotor resistances and magnetizing inductance (stator refers to the stationary part of the motor, while rotor is its rotating part. The magnetizing inductance is responsible for producing magnetic field in the motor). These parameters can fluctuate under variations in temperature and magnetic saturation. This impacts performance variables such as torque, speed, and traction power.
Therefore, in this study, the authors have also evaluated the robustness of what is known as the loss minimizing algorithm (LMA) through a sensitivity analysis of key parameters, which include stator and rotor resistances and magnetizing inductance.
To conclude, an analytical method has been developed to determine the LSCP, below which RB must be disabled. A model-based LMA extends the LSCP to enhance RB efficiency. Significant efficiency improvements were found in low-torque, high-speed regions.
Experiments conducted showed a 13% reduction in system losses under the Modified Indian Driving Cycle (MIDC), and 7% under the US EPA High way cycle. Implementation on a full-scale EV would offer more insights into system-wide behaviour, including interactions with battery dynamics, and is intended to be part of future scope of study.
Dr. Franco Leonardi, who is the Technical Leader at eMachines Research, Research and Advanced Engineering, Ford Motor Company, Dearborn, USA, acknowledged the importance of the work done by the authors with the following comments: “The authors present a rigorous study on optimizing energy recovery in electric vehicles (EVs) during deceleration. The paper introduces a dual-pronged strategy: a model-based approach to analytically determine dynamic low-speed thresholds for regenerative braking, and a loss-minimization framework utilizing a variable flux approach. A key strength of this work is the achievement of significant energy recovery gains without the computational overhead typically associated with artificial intelligence methodologies. Validated through both simulation and experimental results, the proposed method offers a compelling balance of energy efficiency and real-time performance, making it highly suitable for industrial implementation.”
Article by Akshay Anantharaman
Click here for the original link to the paper
