ADVANCED ELECTRIC DRIVES M

Team : Professor:�Francesco Gambelli Prof. Giovanni Franceschini�Alessio Morgante�Gerardo Benedettini�Tommaso Dalmonte�

Electrification of an Alfa Romeo Giulia using first a Wound Stator DC motor and then an Induction Motor

Alfa Romeo Giulia

WOUND STATOR DC

​

Motor Datasheet

Wound stator

Motor parameters �and equations

Motor equations

Equivalent circuit

Global scheme

Speed control

ECU

Varying PID

Current control iq

Battery

Vehicle body

Motor

Current control id

Body Vehicle

Battery Sizing

Samsung INR21700-50E 21700

​

Load inertia estimation

Control strategy useful equations

Current control

Target:

• Include zero frequency inside the bandwidth to deal with constant torques.
• Include high frequencies till 500rad/s.

​

​

Saturation &�Anti-windup strategy

​

​

Electrical and mechanical quantities used in our control strategy, such as currents, voltages and torques, are limited in magnitude. Saturation in our PID regulators is therefore necessary.

​

​

Wind-up is a problem that appears when PID output is limited and the integral component accumulates error.

Clamping is the anti-windup strategy we have chosen.

Speed control

Target:.

• Increase bandwidth till 30 rad/s.
• Make the controlled speed insensible to torque loads.

​

Lead-lag network

It helps to recover the phase margin, by adding a zero before the crossing frequency and then a pole.

Using specific values in a lead-lag network added in the feedback loop, the final closed loop system works as an Ideal First Order System (no overshoot).

Flux weakening�& stator current control

Varying PID

Due to flux weakeing the values of Kt and Ke change proportionaly with the current Id/Id nominal. �In order to take in account this effect a varying PID controller startegy is used.

Varying PID

δ

Rigenerative braking

Our electric machine has an independent excitation circuit, and is reversable, so it can function as a motor or a generator

0 Nm

Blue : WLTP Class3 @ Torque �Red : State of charge (SOC)

T, i

w, V

WLTP Class 3

Performances

​

• 0 – 100 km/h : 4.514 s

​

• Torque plot �(speed limited to 160 km/h)�

A

A

B

B

INDUCTION

INDUCTION MOTOR

​

Motor Datasheet

TESLA MODEL S 2016

Motor parameters�

Motor equation

Control strategy

Flux control

ECU

Flux weakening selector

Flux Obserers

Vehicle body

Slip control

Motor

Speed control

Flux control

Current control

Global scheme

Current

control

The first variable to be controlled is the stator current Is, in particular the two components on the rotor reference frame

• Isd is used to control the rotor flux.
• Isq is used to control the torque�

For both of them we used Kp=1/16 and Ki=100*Kp so that there's no steady-state error, no overshoot and crossing frequency is w_c=4870 rad/s.

​

The optimal value was Kp = 1/16. So doing the system has no overshoot anymore, but there still is a huge steady-state error, that we fixed adding an integral action.

Isd 0-50A step at 0.1s

Isq 0-100A step at 0.3s

The isq request is seen from isd as a disturbance at 0.3 second

Flux control

Flux Observer

Stator equation

Rotor equation

For flux control the value needs to be known, unlike i_s it can't be measure so needs to be estimate with an observer

Actual observer combination of the two

I use stator flux to stimulate the total rotor flux. From it, with the rotor equition, we can calculate the current i_s. But we already know the i_s because we measure it, so we make the difference between the two and that will give us an error signal that we multiply by a complex signal and send it back into the stator loop to compensate it.

�

Speed control

At first, we tested the system without saturation to compute the real time constant and fix the correct Kp, a good result was found with Kp=200.

​

We decided not to use an integrator in the speed control, because the steady-state error of an external loop is compensated by the inner loop.

​

Once the PID design was completed, we added torque saturation, flux weakening and slip control, that is the dynamic saturation on the Simulink scheme.

Slip control

During strong accelerations and decelerations, the rear traction tires of our model cannot stand that strong torque TE_max applied on them and start slipping, resulting in a loss of control.

​

For that reason, we implemented a slip control which reduced saturation values of the electrical torque when the slip increases. It is implemented on Simulink using a dynamic saturation block and two very simple PID regulators.

Vehicle dynamic : Mechanical brake and ABS

ABS control

During the braking the slip of the front wheels must be controlled in the case of over pressure on the brake or in extreme condition of the road

Braking strategy

Wltp Class 3 & FTP-72

Time [s]

Velocity [km/h]

Acceleration

Step 0-100 km/h

As expected, the final result is that, during a 0-100 km/h step input, most of the transient is completed with the maximum torque (linear growth), and the actual control starts acting only in the final part (decaying shape).

Comparison

INDUCTION

WSDC

• A control more complex.�
• Extremely lightweight motor�
• Fast rotor angular velocity�
• In general, cheap, but die casting for creating brass bars for the rotor is a complex and expensive�processing
• No necessity of rare earth

• Easy control�
• Heavy motor�
• Limited rotor angular velocity�
• Requires maintenance for the slip rings
• Cannot be used in all environments due to the possible sparks of the splip rings contacts.
• No necessity of rare earth

THANKS FOR YOUR ATTENTION

Francesco Gambelli �Alessio Morgante �Tommaso Dalmonte �Gerardo Benedettini