Regenerative Braking Background
Hybrid and electric vehicles typically have a system known as regenerative (regen) braking which helps to improve efficiency. Essentially, regen braking turns the motors into generators, thereby converting the cars kinetic energy into electrical energy and later into chemical energy in the battery pack. Regen braking is only effective at high speeds so for reliability and safety reasons all hybrid and electric vehicles also have traditional mechanical brakes. Another important point about regen braking is that if too much kinetic energy is converted to electrical energy in too short a time, damage can incur (thermal runaway, fire, explosions, earthquakes, viral pandemics, alien invasions, etc.)
The problem
While driving, there are scenarios when it is be more desirable to use more regen braking but also scenarios when it is best to use mostly mechanical braking:
The problem we face is designing the interface and control architecture to allow for these two states as well as some middle ground. Some important criteria include:
- The system must provide a way to engage regenerative braking without engaging the mechanical brakes
- The mechanical brake must be robust and able to provide the necessary maximum braking forces
Possible Solutions
1) Dual position brake pedal
Design the brake pedal mechanism to have two stages: one where the pedal position is measured with a sensor to control regen braking and one where the pedal engages the mechanical brakes.
The above sketch is a gross abstraction but it illustrates the concept. The mechanical brake would not engage until part way through the pedal travel
2) Two brake pedals
Have two brake pedals, one for regen braking and one for mechanical braking
3) Automatic regenerative braking
When neither the brake pedal or the accelerator pedal is pressed the regen brake would automatically engage. The brake pedal would have a single state (both mechanical and regen braking). To permit coasting, the accelerator pedal would have two positions. Think of the button that controls power windows in a car. There is usually a detent half way down where the logic goes from "stop when button released or window reaches bottom" to "stop when window reaches bottom". Similarly, our accelerator pedal could have a detent part way down where the logic goes from the "coasting" state to "accelerate" state. That way we still have two braking states: regen only when no pedal is pressed, regen and mechanical when the brake pedal is pressed. Overall we would have 4 states as shown below:
| Truth Table | ||||||
|---|---|---|---|---|---|---|
| Inputs | Outputs | |||||
| Brake pedal pressed | Accel. pedal pressed | Accel. pedal pressed beyond detent | Regen braking | Mechanical braking | Power Driven to Motors | State Name |
| 0 | 0 | 0 | 1 | 0 | 0 | Regen braking |
| 0 | 0 | 1 | Mechanically impossible state | |||
| 0 | 1 | 0 | 0 | 0 | 0 | Coasting |
| 0 | 1 | 1 | 0 | 0 | 1 | Motors driven |
| 1 | X | X | 1 | 1 | 0 | Regular Braking |
The above diagram is a simplification and obviously the control system would be more complex. For example, the "motors driven" state would be analog and dependant on the accelerator pedal position as well as the car's current velocity. The regen gain would also change relative to the car's current velocity.
This system is similar to the system that Tesla uses:
|
|---|
| (excerpt from the the Model S owners manual - section 4.21) |
4) Decoupled brake pedal
Have the brake pedal mechanically decoupled from the master cylinders. A hydraulic power unit would actuate the mechanical brakes and be electronically controlled.
Renault explains this system well on their website:
