Politecnico di Milano 1863

School of Industrial and Information Engineering

Master Degree in Mechanical Engineering
Mechanical engineering department - Applied Mechanics section

Reports on the laboratories

Report on the exercises of the course of
Electrical Drives for Industry and Transport Applications
Teachers : Students :
Prof. Davide Tarsitano Renato Desideri - 919980
Dr. Maria Laura Bacci Lorenzo Scocco - 921216

Academic year 2018-2019

Reports on the laboratories R. Desideri, L. Scocco

Contents

2
Reports on the laboratories R. Desideri, L. Scocco

1 Report 1

In this laboratory we are going to perform an energy saving analisys for a pnumatic compressor
comparing a direct drive controlled electrical drive.
An air compressor for industrial application is made by: electric motor, compressiong stage, air
tank, air load.

Data

Symbol Value Units Description

Pn 75 [kW] electric motor power
Jtot 28 [kgm2 ] overall inertia
Nn 1300 [rpm] motor speed
Qlm 1250 [l/min] intake air flow
p2 10 [bar] rated pressure
p1 1 [bar] ambient pressure
t1 20 [°C] ambient temperature
Cls 250 [l] air tank volume
∆p ±1 [bar] pressure tolerance for on/off hysteresis control
ηc 0.6 [-] overall efficiency
R 287 [J/kg/K] constant
ρ1 1.245 [kg/m3 ] air density @ 20 °C 1 bar
ckWh 0.13 [e/KWh] Cost of one kWh

Requests

1. Energy per hour with three different air flows with hysteresis on/off control and fixed speed.

2. Energy per hour with three different air flows with a control that varies the motor speed to
keep pressure and air flow constant,

3. Make a energy consumption comparison and express the annual money saving of solution 2
considering 8 hours/day 48 week/year.

4. Make the calculations 1, 2, 3 for a reduced size compressor with Pn , Jtot , Qlm , Cls of 1/10th .

3
Reports on the laboratories R. Desideri, L. Scocco

Request 1 - Hysteresis on/off control and fixed speed

Figure 1: Overall system

In order to solve request 1, first of all let’s consider the hysteresis control and the relative pressure
profile.

Figure 2: Pressure control with hysteresis.

To compute the work done by compression consider a polytropic process:

pv k = const

p1 v1k
pv k = p1 v1k =⇒ p =
vk

4
Reports on the laboratories R. Desideri, L. Scocco

where p1 is the ambient pressure, v1 is the specific volume and k is the ratio of specific heats at
constant pressure and constant volume. Thus, the work needed to bring a gas from condition 1 to
condition 2 can be calculated as:
Z 2 Z 2
p1 v1k p2 k−1
 
p1 v1 k
W12 = pdv = dv = 1− = 186887.09 [J/kg]
1 1 vk k−1 p1

Notice that, since the pressure p2 is higher than pressure p1 , the work would be negative, but we’ll
consider the absolute value.
Let’s now consider the air tank in Figure ?? describing it with a dynamic model:

Figure 3: Pressure control with hysteresis.

The mass flow rate entering the tank is proportional to the volumetric flow rate in 1, while the
flow rate coming out from the tank is a fraction of the entering one:

G1 = Qlm ρ1

G2 = αG1

Since the volume is constant, the flow rate entering the tank contributes to increase the pressure,
whereas the one coming out decreases it. The dynamic equation describing the mass flow rate can
be expressed as:
dp
G1 − G2 = C
dt
Cls
C= = 2.97145 · 10−6 [ms2 ]
RT1

where C is a constant related to the capacity of the tank and to the state of the gas as:

∆pon
G1 (1 − α) = C
∆ton

∆pon
∆ton = C
G1 (1 − α)

By computing the finite difference equation, we can distinguish two different conditions for the
charge and the discharge phase.

Charge condition In charge condition the motor is turned on and the on-time interval can be
calculated:
kg
 
∆pon
G1 (1 − α) = C = 0.0259375
∆ton s

∆pon
∆ton = C
G1 (1 − α)

5
Reports on the laboratories R. Desideri, L. Scocco

Discharge condition In discharge condition the motor is off and the mass flow rate entering
the tank is equal to zero:

∆poff
−αG1 = C
∆toff

∆poff
∆toff = C
−αG1

Notice that, increasing the α value the ∆ton time interval is increasing, whereas the ∆toff time
interval is decreasing.

The power needed for compression is:

W12 G1
Pc = = 8078.973341 [W]
ηc

and the torque follows:

Pc
Tc = = 59.3450 [Nm]
ωc

Then, the dynamic equation of motion:

Tm − Tr = Jtot ω̇

Where Tm = P ωn = 550.9210 [Nm] (nominal power and nominal speed of the motor) and Tr = Tc .
n

At the beginning of the compression, the motor needs a certain time ∆tstart to reach the speed ωc :

ωc
∆tstart = Jtot = 7.7542 [s]
Tm − Tr

Notice that, at the end of ∆toff , pressure reaches a value lower than 9 bar, since the motor needs a
start time to turn on again. However, this transient is small, thus it can be considered negligible.
In Figure ?? the pressure, speed, torque and power are shown as a function of time. It can be
noticed that at the starting time, the torque is equal to the nominal one, then, when all the inertia
are overcome, the torque reaches the Tc value. Thus, the power increases to its nominal value, then
it decreases to the value needed for the compression.

6
Reports on the laboratories R. Desideri, L. Scocco

Figure 4: Pressure, speed, torque and power in a fixed speed hysteresis on/off control.

The energy consumption in one cycle is the work done during the start and the on intervals:

7
Reports on the laboratories R. Desideri, L. Scocco

1
Wcycle = Wstart + Won = Pn ∆tstart + Pc ∆ton
2
While the energy consumption in an hour is:
1h
Wh = Wcycle
∆tcycle

where
∆tcycle = ∆tstart + ∆ton + ∆toff

Request 2 - Controlled speed electrical drive

Using a controlled speed electrical drive, the air tank is no more needed: the motor compressor
reaches a constant power value after its first start and it provides the demanded pressure for the
entire day. Thus, the power needed for compression can be expressed as:
W12 G1 W12 αQlm ρ1
Pc = =
ηc ηc

The energy needed for an hour can be calculated as the weighted mean on a day of two contributions:
one is the contribution of energy absorbed by the DC motor during the start-up transient and the
other is the energy consumed in the steady state condition at constant power.

s
 
1 1   1
Wh = Wstart + Won = · Pc · 8 h · 3600 − ∆tstart + Pn ∆tstart
8 h h 2

Request 3 - Energy consumption comparison

By considering 8 h/day, 5 days/week and 48 weeks/year, we can compute the annual energy
consumption for this configuration:

h day week
Wyear = Wh · 8 ·5 · 48
day week year

And finally, the annual cost of this solution is:

Cyear = Wyear · ckW h

8
Reports on the laboratories R. Desideri, L. Scocco

Results

For the larger motor, the final results are displayed below:
Value
Symbol Units Description
α = 25% α = 50% α = 75%
Hysteresis control
k 1.4 [-] Polytropic index
v 0.8032 [m3 /kg] Specific volume
W12 186887.0942 [J/kg] Work
G1 0.0259375 [kg/s] Mass flow in
G2 0.006484 0.01297 0.01945 [kg/s] Mass flow out
C 2.9715·10−6 [ms2 ] Calculation constant
∆ton 30.5498 45.8247 91.6494 [s] On-phase duration
∆toff 91.6495 45.8247 30.5498 [s] Off-phase duration
Pc 8078.9733 [W] Compressor power at
steady state
Tc 59.3450 [N·m] Resistive torque
Tm 550.9209 [N·m] Motor torque
∆tstart 7.7542 [s] Start up time
Ecycle 537595,3739 661001,0146 1031217,937 [J] Energy required for a cy-
cle
Ehour 14892575.22 23938773.96 28567006,78 [J] Energy per hour of opera-
tion
Speed control
G1 = G2 0.006484 0.01297 0.01945 [kg/s] Mass flow
Pc 2019.7433 4039.4867 6059.2300 [W] Compressor power at
steady state
7271076.007 14542152.01 21813228.02 [J] Energy per hour, neglect-
Ehour
ing start-ups
7305466.321 14574584.63 21843702.94 [J] Energy per hour, consid-
ering start-ups
Cost analysis
hyear 1920 [h] hours of service per year
Costhyst-ctrl 1032.552 1659.755 1980.646 [e] Cost for hysteresis control
tstart/year 1861.0182 [s] Start ups time in one year
Estart/year 69788182.22 [J] Energy for start ups in a
year
504.128 1008.256 1512.384 [e] Speed control cost (con-
Costspeed-ctrl
sidering start-ups)
506.512 1010.504 1514.497 [e] Speed control cost (ne-
glecting start-ups)
526.039 - 51% 649.250 - 39% 466.149 - 24% [e] Saving (considering the
Saving
start-ups)
528.424 - 51% 651.499 - 39% 468.262 - 24% [e] Saving (neglecting the
start-ups)

9
Reports on the laboratories R. Desideri, L. Scocco

Equivalent calculations were performed for the smaller motor, whose results are reported below:

Request 4 - Compressor with reduced size

Value
Symbol Units Description
α = 25% α = 50% α = 75%
Hysteresis control
k 1.4 [-] Polytropic index
v 0.8032 [m3 /kg] Specific volume
W12 186887.0942 [J/kg] Work
G1 0.002594 [kg/s] Mass flow in
G2 0.0006484 0.001297 0.001945 [kg/s] Mass flow out
C 2.9715·10−7 [m/s2 ] Calculation constant
∆ton 30.5498 45.8247 91.6494 [s] On-phase duration
∆toff 91.6495 45.8247 30.5498 [s] Off-phase duration
Pc 807.8973 [W] Compressor power at
steady state
Tc 5.9345 [N·m] Resistive torque
Tm 55.0921 [N·m] Motor torque
∆tstart 7.754 [s] Start up time
Ecycle 53759.5374 66100.1015 103121.7937 [J] Energy required for a cy-
cle
Ehour 1489257.522 2393877.396 2856700.678 [J] Energy per hour of opera-
tion
Speed control
G1 = G2 0.0006484 0.001297 0.001945 [kg/s] Mass flow
Pc 201.9743 403.9487 605.9230 [W] Compressor power at
steady state
727107.601 1454215.201 2181322.802 [J] Energy per hour, neglect-
Ehour
ing start-up
730546.6321 1457458.463 2184370.294 [J] Energy per hour, consid-
ering start-up
Cost analysis
hyear 1920 [h] hours of service per year
Costhyst-ctrl 103.2551882 165.9754994 198.0645803 [e] Cost for hysteresis control
tstart/year 1861.0182 [s] Start ups time in one year
Estart/year 6978818.222 [J] Energy for start ups in a
year
50.413 100.826 151.238 [e] Speed control cost (with
Costspeed-ctrl
the start-ups)
50.651 101.050 151.450 [e] Speed control cost (with-
out the start-ups)
52.604 - 51% 64.925 - 39% 46.615 - 24% [e] Saving (considering the
Saving
start-ups)
52.842 - 51% 65.150 - 39% 46.826 - 24% [e] Saving (neglecting the
start-ups)

The following chart is referred to the standard motor and represents the work per hour in case of
hysteresis and speed control and the energy saving by exploiting the speed control, in terms of cost
difference and percentage. The results are obtained by changing the percentage output air flow α
from 0 to 1.
The maximum energy savings are 651.7866 e/year (percentage energy saving of 40.7933% ), cor-
responding to α = 0.468.

10
Reports on the laboratories R. Desideri, L. Scocco

On the other hand, the second one is referred to the motor with reduced size. Notice that the work
per hour of the 2 speed profiles and the energy savings per year by exploiting the speed control
are scaled of a factor 0.1, but the percentage energy savings are exactly the same of that ones
referred to the standard motor, as α changes. The maximum energy savings are 65.1787 e/year
(percentage energy saving of 40.7933% ), corresponding to α = 0.468.

Figure 5: Work per hour and energy savings per year for hysteresis and speed control of the
standard motor.

11
Reports on the laboratories R. Desideri, L. Scocco

Figure 6: Work per hour and energy savings per year for hysteresis and speed control of the motor
with reduced size.

12
Reports on the laboratories R. Desideri, L. Scocco

2 Report 2
An electrical drive controls via a worm gear and a reduction stage a linear guide to move a item
in a production industry. The duty cycle of the linear motion is defined as follows:
1. Stage A: an item of 300 kg mass and is moved of 15 cm in 0.4 s against a resistance force of
2000 N,
2. Stage B: the system stops for 0.2 s,
3. Stage C: the linear drive goes back in 0.4 s (15 cm) with no load,
4. Stage D: the system stops for 0.2 s.

Data
Symbol Value Units Description
m 300 [kg] Mass to transport
tA 0.4 [s] Duration Stage A
tB 0.2 [s] Duration Stage B
tC 0.4 [s] Duration Stage C
tD 0.2 [s] Duration Stage D
s 15 [cm] Distance
θamb 65 [°C] Ambient temperature
Jtransm 250·10−6 [kgm2 ] Inertia transmission gear
For the data regarding the motor, refer to the datasheet in Figure ??

Requests
a) Under the hypothesis of a symmetric triangular speed profile size the electrical motor suited
for the described cycle using as reference the motor data sheet annexed.
b) Make a thermal analysis in order to verify the chosen motor.
c) Under the hypothesis of an intermittent cycle (30% on over a 10 minutes period) calculate
with the motor sized at point 1 the maximum transportable mass.
d) Under the hypothesis of a limited time cycle of 30 minutes calculate with the motor sized at
point 1 the maximum transportable mass.
e) Make the analysis for the following different speed profiles:
1. Trapezoidal
2. Pseudo-sinusoidal with a constant speed interval
For each profile make the points “b” to “d” analyses.

13
Reports on the laboratories R. Desideri, L. Scocco

Figure 7: Catalog of the available motors

14
Reports on the laboratories R. Desideri, L. Scocco

Request 1 - Electromechanical sizing

Kinematic analysis The average speed during the outward stage of duration ton is:
Z ton
1 s
vavg = v(t)dt = = 0.375 [m/s]
ton 0 ton

In the triangular speed profile, the average speed is the ratio between the triangle area (the integer
of speed in time) and time, that corresponds to the triangle base. Thus, the maximum speed, that
is the height of the triangle, is two times the average speed:

vmax = 2 · vavg = 0.75 [m/s]

Instead with the other two speed profiles, the maximum speed is calculated imposing that the
distance traveled in the time ton must be equal to s:
Z ton
v(t)dt = s
0

So, for the trapezoidal profile, keeping fixed vavg = 0.375 [m/s], we can calculate the maximum
speed as:
 
tA1 tA3
s = vmax · tA2 + vmax +
2 2
vavg (tA1 + tA2 + tA3 ) hmi
vmax = = 0.6
tA1
+ tA2 +
tA3 s
2 2
Where tA1 is the accelerating period in tA , tA2 is the constant speed tract and tA3 is the decelerating
section Finally, also for the pseudosinusoidal profile, keeping fixed vavg = 0.375 [m/s], we can
calculate the maximum speed as:
 Z tA1    
1 vmax 2π
vavg = vmax · tA2 + 2 · 1 − cos dt
tA 0 2 tA1

vmax = 0.6 [m/s]

The accelerations’ calculation is obtained by differentiating speed with respect to time:

dv(t)
a(t) =
dt

Dynamic analysis of the motor shaft

Through the dynamic equation of the mechanical system, the profile of the torque demanded to
the motor in different duty cycles can be obtained:

dω(t)
Tm − Tres = Jtot ·
dt
where Tres is the torque needed to win the resistant force Fres .
The moment of inertia Jtot at the motor shaft has to be known. To do this, the gear ratio k of
the reducer has to be found. This ratio can be obtained through the maximum rotational speed at
the motor shaft and the maximum traslation speed of the carriage, that depends on the particular
speed profile in use (for simplicity the reducer has unitary efficiency). Once k is known, all the
inertia and resistant forces can be carried to the shaft. The motor data are available on datasheet
from which we can find the maximum angular speed of motors in exam:

rad
 
ωmax = 700
s

15
Reports on the laboratories R. Desideri, L. Scocco

The overall gear ratio is:  

ωmax rad
k= =
vmax m
To obtain the moment of inertia referred to the mass, consider the law of conservation of energy:
1
E= mv 2 kinetic energy of translation
2
1
E = Jtot ω 2 kinetic energy of rotation
2
Assuming that the gear ratio has no losses:
1 1
mv 2 = Jtot ω 2
2 2
thus:
v2
Jtot = m ·
ω2
And the moment of inertia referred to the mass at the motor shaft is:
m
Jmass =
k2
Now, we can calculate the overall moment of inertia after making some hypothesis on the moment
of inertia of the motor, in order to obtain a 1st attempt value. If the moment of inertia of the
motor is similar to the one of the load (1:1 ratio), then we’ll have a very small acceleration time
from 0 to the nominal speed and very small time response for the speed regulator: this is typical
of robotic applications and servomechanisms. Conversely, if the moment of inertia is 20 times the
one of the load, we’ll have long acceleration time and very low dynamic: this is the case of electric
traction applications. Since this is a positioner, we want a good dynamics, thus we can consider a
moment of inertia of the motor 20% of that one of the load as 1st attempt value. Thus:

Jtot = 1.2 · (Jmass + Jtransm )

With Jtransm = 250 · 10−6 kgm2 as stated in the data. In the third stage, there is no more mass to
transport, so:

J3 = Jmot + Jtransm
where Jtransm = 250 · 10 kgm
−6 2

To calculate the resistant torque, let’s consider the resistant force and apply the law of conservation
of power:

P =F ·v power in linear motion

P =T ·ω power in rotational motion
So we obtain that:
v
T =F·
ω
and finally we can write the resistant torque by introducing the gear ratio:
Fres
Tres =
k
Let’s now define the angular acceleration from the linear one. Since the angular and linear speed
are proportional via the gear ratio

ωavg = k · vavg
And since the acceleration is the time derivative of speed
dω
=k·a
dt

16
Reports on the laboratories R. Desideri, L. Scocco

The angular acceleration is:

ω̇ = k · a

From the dynamic equation of motion, we can calculate the needed torque in all the different stages,
with the formula:

dω(t)
Tm = Jtot · + Tres
dt
Considering that we have Tres and the mass m only in stage A, we can distinghish different phases
with respect to motor torque, in case of triangular speed profile:

dω
Tm1 = Jtot · + Tres
dt

dω
Tm2 = −Jtot · + Tres
dt
dω
Tm3 = −J3 ·
dt
dω
Tm4 = J3 ·
dt
On the other hand, for the other 2 speed profiles, torque can be generalized as follows:
(
Tres + Jtot · ω̇(t) t < ttot /2
Tm =
J3 · ω̇(t) t ≥ ttot /2

Now we have different values of torque during the duty cycle because of different phases. But we
would like to have a torque quantity that takes into account not only the amount of torque needed
for a single phase but one that considers also the amount of time for which each torque must
be produced. This value is called RMS torque and it represents a torque value that, if produced
continuously by the motor, would yield the same heating and motor sollecitation as all the various
torques. It can be calculated as:
s
Z ttot
1
Trms = Tm (t)2 dt
ttot 0

At this point, we are able to choose the proper motor for this application, for each speed profile of
the mass that has to be carried, remembering that some checks have to be satisfied. The procedure
consists in keeping fixed one of the speed profiles, start from the first motor of the catalogue
(503402, the smallest one) and verify whether all the following checks are satisfied. In case one of
them is not verified, the following motor of the catalogue is chosen.
1. Trms : the motor must have a nominal torque that is higher than the RMS torque found for
the fixed speed profile.
2. Tmax : the maximum torque needed for fixed speed profile has to be lower than the maximum
torque provided by the selected motor.
3. ωavg : the average angular speed has to be less than or equal to the nominal angular speed of
the selected motor.
4. ωmax : the maximum angular speed found from calculation has to be lower then the maximum
allowed by the selected motor.
5. ω̇max : the maximum angular acceleration has to be lower than the maximum one allowed by
the selected motor, in order not to exceed centrifugal forces on the rotor.
6. Jmotor : the motor inertia has to be less than or equal to the motor inertia calculated before
as 1st attempt.
Jmotor ≤ 0.2 · (Jmass + Jtransm )

17
Reports on the laboratories R. Desideri, L. Scocco

Triangular profile

Mass speed

0.75
v [m/s]

-0.75

0 0.2 0.4 0.6 0.8 1

t [s]
Mass acceleration
3.75
a [m/s2 ]

-3.75
0 0.2 0.4 0.6 0.8 1
t [s]

Angular speed

700

-700

0 0.2 0.4 0.6 0.8 1

Angular acceleration
3500

-3500
0 0.2 0.4 0.6 0.8 1

18
Reports on the laboratories R. Desideri, L. Scocco

Motor torque

4.5732
T [Nm]

1.225

0
-0.2875

-1.225

0 0.2 0.4 0.6 0.8 1

t [s]

19
Reports on the laboratories R. Desideri, L. Scocco

Trapezoidal profile

Mass speed
0.6
v [m/s]

-0.6
0.15 0.25 0.4 0.6 0.75 0.85 1
t [s]
Mass acceleration
4
a [m/s2 ]

-4

0.15 0.25 0.4 0.6 0.75 0.85 1

t [s]

Angular speed
560

-560
0.15 0.25 0.4 0.6 0.75 0.85 1

Angular acceleration
3733.3333

-3733.3333
0.15 0.25 0.4 0.6 0.75 0.85 1

20
Reports on the laboratories R. Desideri, L. Scocco

Motor torque
4.7352

2.1429
T [Nm]

1.3067

0
-0.4495

-1.3067

0.15 0.25 0.4 0.6 0.75 0.85 1

t [s]

21
Reports on the laboratories R. Desideri, L. Scocco

Pseudosinusoidal profile

Mass speed
0.6
v [m/s]

-0.6
0.15 0.25 0.4 0.6 0.75 0.85 1
t [s]
Mass acceleration
6.2832
a [m/s2 ]

-6.2832
0.15 0.25 0.4 0.6 0.75 0.85 1
t [s]

Angular speed
560

-560
0.15 0.25 0.4 0.6 0.75 0.85 1

Angular acceleration
5864.3062

-5864.3062
0.15 0.25 0.4 0.6 0.75 0.85 1

22
Reports on the laboratories R. Desideri, L. Scocco

Motor torque

6.2150

X 0.925
Y 2.0525
T [Nm]

2.1429

X 0.325
Y -1.9292
-2.0525

0.15 0.25 0.4 0.6 0.75 0.85 1

t [s]

Request 2 - Thermal analysis

With the thermal analysis, we want to verify that the motor, while working, does not exceed its
thermal limits, that is, does not exceed the maximum admittable temperature. The isolation class
of electric machine in the annexed catalogue is Class F. This means that these motors can reach
a maximum admittable temperature of 155 °C. However, to guarantee a further safety margin, we
are going to consider a tolerance of 20%, that is, the maximum admittable temperature is set to
130 °C. The thermal equivalent circuit is shown in Figure ??

Ptot Cth Rth θ

Figure 8: Equivalent thermal circuit

The 1st order dynamic equation of this model (neglecting any kind of mechanical losses) is:
dθ θ
Ptot = Cth · +
dt Rth
And at steady state equation is:
θ
Ptot =
Rth
where:
• Ptot is the overall power losses (iron and copper losses)

23
Reports on the laboratories R. Desideri, L. Scocco

• Ptot = Pcu + Pf e
• Cth is the thermal capacitance [J/°C]
• Rth is the thermal impedance [°C/W]
• θ is the difference between motor temperature and ambient temperature [°C]
• θ = θmot − θamb ; we assume θamb = 65°C
The motor has a forced cooling; therefore, it does not depend on the angular speed of the shaft.
The absolute temperature of the motor can be expressed as:

θmot = θ + θamb

.
To find steady state temperatures we have to define power losses, that is, iron and copper losses.
Thus, we need the following data available on datasheet:
ωn = rad
  Nm  J K
s Kt = A R w0 = [Ω] τ = [s] Cth = K Rth = W

Iron Losses
Iron losses depend linearly on angular speed of the motor. We have to calculate iron losses at
our operating condition from the iron losses in nominal condition. We can do this by defining a
proportionality constant (losses/speed) that can be computed at nominal operation point. Then,
by computing the average angular speed in a cycle, we can find iron losses.

Pf e
kf e =
ωn
Then, it’s possible to get the iron losses by exploiting the calculation of the mean integral of the
absolute value of the speed of the selected profile.
Z ttot
1 Pf e
Pf e = |ω(t)| dt
ttot 0 ωn

Copper Losses
To calculate copper losses, we have to know the RMS value of current and the temperature of
windings. The RMS value of corrent can be easly calculated from the RMS value of torque through
the torque constant Kt , available on datasheet, that is the ratio between the torque provided to
the shaft and the current flowing through a single motor phase. 1

Trms
Irms =
Kt
Then, the copper losses are:

2
Pcu = 3 · R(θmot ) · Irms

Steady State Temperature

Since the winding resistance is temperature dependent and since the resistance value on datasheet
is expressed for a temperature different from the one that will be reached during the duty cycle
(that we don’t know yet), we have to follow an iterative procedure to find the correct result. The
procedure is:
1. Calculate Pcu losses under the hypothesis that the working temperature is the one expressed
in datasheet at the nominal resistance value;
2. Calculate the overall power losses as the sum of iron and copper losses;
3. Calculate the steady state temperature;
4. Update the value of winding resistance as a function of the new temperature value obtained
in point 3;
5. Calculate the copper losses with the new resistance value obtained in point 4;
6. Repeat the iterative procedure from point 2 until temperature converges to a value
1 Pay attention, Kt = n · 3 · Ψm , where n is the number of pole pairs and Ψm is the magnetic flux.

24
Reports on the laboratories R. Desideri, L. Scocco

The procedure expressed by formulas is:

Ptot = Pcu + Pf e
θmot = Rth · Ptot + θamb
Once the temperature θmot is known, the real value of winding resistance can be found:
 
234.5 + θmot Rw0
R(θmot ) = ·
234.5 + θ0 2
with θ0 = 20°C, reference temperature. The resistance in datasheet is a phase to phase resistance,
to find the phase resistance, since the motor is a three-phase motor, we have to divide its value by
two. Let’s calculate copper losses:
2
Pcu = 3 · R(θmot ) · Irms
Now, we can update the overall power losses Ptot and then calculate the temperature. Finding
that it is different from the one supposing at the beginning of iteration, we can correct the first
value and repeat the procedure until reaching that the temperature converges to a value. The
resultant temperature has to be lower than the maximum admittable temperature of 130 °C. Thus,
the selected motor is able to follow this duty also from a thermal point of view.

Request 3 - Intermittent cycle analysis

In this paragraph, we want to find the maximum admittable torque that the motor can provide
during an intermittent duty, without exceeding thermal limits. We are going to assume that the
increasing of total losses is due to the increasing in maximum torque needed to transport a bigger
mass. The characteristics of the following intermittent periodic duty are:
1. Time period: tcycle = 10min
2. On phase time: ton = 3min
3. Off phase time: toff = 7min
Decomposing the duty cycle into on and off phases, two equations can be found:

dθ θ
Ptot = Cth · + 0 < t ≤ ton


dθ θ dt Rth
Ptot = Cth · + =⇒
dt Rth dθ θ
0 = Cth · ton < t ≤ tcycle

 +
dt Rth
The solution is: ( t
θon = Aon · e− τ + Ptot · Rth 0 < t ≤ ton
θ(t) = t−ton
θoff = Aoff · e− τ ton < t ≤ tcycle
To calculate the maximum torque, we assume that the maximum admittable temperature is reached
at the end of on phase. The overtemperature during off phase (no losses) is:
h t−ton
i h t−ton
i
θ(t) = θmax e− τ = (130 − θamb ) · e− τ with ton ≤ t ≤ tcycle
thus, the temperature at the end of the cycle, assuming t = 10 min = 600s, is:
h tcycle −ton i
θmin = θmax e− τ

The overtemperature during on phase is:

t
θ(t) = Aon · e− τ + Rth · Ptot with 0 ≤ t ≤ ton
To calculate the Aon e Ptot constants, we have to impose the following initial conditions:
(
θon (0) = θ(tcycle ) = θmin
θon (ton ) = θof f (ton ) = θmax
It is possible to obtain the absolute temperature profile during the intermittent cycle
θmot (t) = θ(t) + θamb

25
Reports on the laboratories R. Desideri, L. Scocco

Intermittent cycle temperature profile

130

125

120

115

110

105
0 100 200 300 400 500 600
t [s]

Figure 9: Temperature profile.

Notice that the temperature profile is the same indipendently of the speed profile because the
former is imposed during the calculation and depending on the latter a different mass is being
transported during the cycle.
From the overall power losses, the copper losses can be derived. Copper losses will lead us to
calculate rms current and torque.

Pcu = Ptot − Pf e
This is possible since the iron losses do not depend on duty type but only on speed. To find rms
current from copper losses, the average temperature over the cycle has to be derived in order to
compute the resistance value at that temperature. The average temperature of the cycle is:
tonZ+tof f
1
θmot,avg = θmot (t)dt
ton + tof f
0

Thus, the resistance values at that temperature is:

 
234.5 + θmot,avg Rw0
R(θmot,avg ) = ·
234.5 + θ0 2
The rms current is: s
Pcu
Irms =
3 · R(θmot,avg )
And the rms torque:
Trms = Irms · Kt
The rms value of the torque will exceed the nominal torque of the motor, but this is the torque
value during on phase, the rms value during the overall cycle will be lower than the machine limit.

26
Reports on the laboratories R. Desideri, L. Scocco

From the rms value of torque we can calculate the total moment of inertia from which the new
mass and new torque can be derived. To find it we have to solve the non-linear equation where the
new total moment of inertia is unknown.
r
1
Trms = · Tm (Jtot,new ), t)2
T
When considering the triangular speed profile, the equation can be solved in a closed form, otherwise
a calculation program (such as Matlab) is needed. From new inertia value, the mass inertia can
be derived, then the mass and finally the new value of maximum torque.

Jmass,new = Jtot,new − Jtransm − Jmot

mnew = Jmass,new · k 2
Tmax = Tres + Jtot,new · ω̇
By using an intermittent duty and working with the same speed profile, a heavier mass can be
transported.

Request 4 - Limited time cycle analysis

As done before, we are looking for the maximum admittable torque, without exceeding thermal
limits and using a machine with a limited time duty of ti = 30 minutes. Assuming that at the end
of the cycle the overtemperature will reach the maximum temperature θmax . The method is the
similar, but now we have only one phase:
t
θ(t) = A · e− τ + Rth · Ptot
Considering the following initial conditions: θ(0) = 0 and θ(ti ) = θmax
0
θ(0) = A · e− τ + Rth · Ptot = 0
ti
θ(ti ) = A · e− τ + Rth · Ptot = θmax
Let’s calculate as done before, iron losses and average temperature:

Pcu = Ptot − Pf e
ton
Z+toff
1
θmot,avg = θmot (t)dt
ton + toff
0

Now, as for intermittent duty, we can obtain new values of transportable mass, rms current and
torque and maximum torque.

Results

Symbol Value Units

Triangular profile Trapeziodal profile Semisinusoidal profile
Electromechanical sizing
Motor 503402 503402 503402
vavg 0.375 0.375 0.375 [m/s]
vmax 0.75 0.6 0.6 [m/s]
amax 3.75 4 6.283 [m/s2 ]
ωmax 700 560 560 [rad/s]
K 933.3 933.3 933.3 [rad/m]
ωavg 350 350 350 [rad/s]
ω̇max 3500 3733 5864 [rad/s2 ]
Jmass 0.0003444 0.0003444 0.0003444 [kgm2 ]

27
Reports on the laboratories R. Desideri, L. Scocco

Jtransm 0.00025 0.00025 0.00025 [kgm2 ]

Jmot 0.0001 0.0001 0.0001 [kgm2 ]
Jtot 0.0006944 0.0006944 0.0006944 [kgm2 ]
Tmax 4.573 4.735 6.215 [Nm]
Trms 2 1.907 2.032 [Nm]
Thermal analysis
Kiron 0.09547 0.09547 0.09547 [Nm]
Piron 22.28 22.28 22.28 [W]
θamb 65 65 65 [° C]
θmax 130 130 130 [° C]
Irms 2.778 2.649 2.823 [A]
Rw (20° C) 2.2 2.2 2.2 [Ω]
Rw (θmot ) 1.396 1.39 1.398 [Ω]
Pcopper 32.3 29.26 33.4 [W]
Ploss 54.57 51.53 55.67 [W]
θmot 88.47 87.16 88.94 [° C]
Intermittent cycle
tcycle 600 600 600 [s]
ton 180 180 180 [s]
toff 420 420 420 [s]
Rth 0.43 0.43 0.43 [° C/W]
Aon -286.1 -286.1 -286.1 [° C]
Aoff 130 130 130 [° C]
τa 2001 2001 2001 [s]
Ploss 910.4 910.4 910.4 [W]
Pcopper 888.2 888.2 888.2 [W]
θavg 117.4 117.4 117.4 [° C]
Rw (θavg ) 1.521 1.521 1.521 [Ohm]
Irms 13.95 13.95 13.95 [A]
Trms 10.04 10.04 10.04 [Nm]
Jtot 0.004921 0.005329 0.004795 [kgm2 ]
Jmass 0.004571 0.004979 0.004445 [kgm2 ]
m 3981 4337 3872 [kg]
Limited cycle
tcycle 1800 1800 1800 [s]
Rth 0.43 0.43 0.43 [° C/W]
A -109.6 -109.6 -109.6 [° C]
τa 2001 2001 2001 [s]
Ploss 406 406 406 [W]
Pcopper 383.7 383.7 383.7 [W]
θavg 102.3 102.3 102.3 [° C]
Rw (θavg ) 1.456 1.456 1.456 [Ω]
Irms 9.373 9.373 9.373 [A]
Trms 6.749 6.749 6.749 [Nm]
Jtot 0.003265 0.003537 0.003181 [kgm2 ]
Jmass 0.002915 0.003187 0.002831 [kgm2 ]
m 2539 2776 2466 [kg]

28
Reports on the laboratories R. Desideri, L. Scocco

3 Report 3

An electric motor moves a mechanical system using the pseudo-sinusoidal speed reference here
represented.

Angular speed
504

-504
0.15 0.25 0.4 0.6 0.75 0.85 1

Angular acceleration
5277.8756

-5277.8756
0.15 0.25 0.4 0.6 0.75 0.85 1

Figure 10: Angular speed and acceleration of the motor.

Using the motor chosen in the previous exercise n°2, size the electrical drive that better fits the
duty cycle described using, as reference, the datasheet annexed. The power supply line provided
is the industrial low-voltage three-phase Vn = 230Vac ± 5%Vac with fn =50 Hz.
The following calculations are requested:
1. Check if the voltage and the frequency of the power supply line are consistent with the ones
of the electrical drive.
2. Check if the maximum voltage and frequency that the drive can provide are consistent with
the maximum values required by the work cycle.
3. Check if the nominal power of the drive is greater than the nominal power of the motor.
4. Check if the nominal current of the drive is greater than the RMS current during the work
cycle.
5. Check if the maximum current of the drive is greater than the maximum current requested
during the work cycle.
6. Thermal check of the electrical drive.
7. Select the braking resistor (datasheet annexed).

29
Reports on the laboratories R. Desideri, L. Scocco

30
Reports on the laboratories R. Desideri, L. Scocco

31
Reports on the laboratories R. Desideri, L. Scocco

Figure 11: Inverter Datasheed.

From the calculation done while choosing the electric motor, the following data are known:

32
Reports on the laboratories R. Desideri, L. Scocco

Semisinusoidal profile
Symbol Value Units
Electromechanical sizing
Motor 503402
vavg 0.375 [m/s]
vmax 0.6 [m/s]
amax 6.283 [m/s2 ]
ωmax 504 [rad/s]
K 840 [rad/m]
ωavg 315 [rad/s]
ω̇max 5278 [rad/s2 ]
Jmass 0.0004252 [kgm2 ]
Jtransm 0.00025 [kgm2 ]
Jmot 0.0001 [kgm2 ]
Jtot 0.0007752 [kgm2 ]
Tmax 6.472 [Nm]
Trms 2.1 [Nm]
Thermal analysis
Kiron 0.09547 [Nm]
Piron 20.05 [W]
θamb 65 [° C]
θmax 130 [° C]
Irms 2.916 [A]
Rw (20° C) 2.2 [Ω]
Rw (θmot ) 1.398 [Ω]
Pcopper 35.65 [W]
Ploss 55.7 [W]
θmot 88.95 [° C]

To be noted: the gear ratio was lowered of 10% with respect to the previous exercise in order
to fulfil the check on the voltage of the inverter, so the graphs of angular speed, acceleration and
torque are consequently changed.

Since the electric motor is a three-phase low voltage synchronous motor, the most suitable converter
is a three-phase inverter made by IGBT. To define the size of the converter, the procedure is similar
to the one followed for the sizing of the electric motor; we are going to rely on the current absorbed
by the motor during its working cycle.

Request 1
Preliminary Checks
In order to size the converter, some preliminary checks are made. First of all, the converter output
voltage has to be higher or equal to the motor nominal one and the supply voltage conforms to the
available line. Then, we have to verify that the nominal output power of the converter is higher
than the motor nominal one. The annexed datasheet refers to converters that have three-phase
voltage at fixed frequency as input and a three-phase output voltage at variable magnitude and
frequency. These converters are then made by a bridge rectifier as input stage and an inverter as
output stage.
Considering input and output voltage, all the converters in datasheet are valid, since they are
related to a power supply of 230-460 Vac at 50-60 Hz.
Considering power, the choice has to be done taking into account a converter with an output power
higher than the nominal one of the motor.

33
Reports on the laboratories R. Desideri, L. Scocco

Request 2
Maximum frequency check
Next check concerns the maximum frequency at which the output stage will operate. Once the
maximum speed Ωmax , expressed in rad/s, and p, the number of pole pairs of the motor, are known,
the maximum frequency can be derived:

Ωmax · p/2
fmax =
2π
where the maximum angular speed can be derived from the speed profile shown.
This value has to be lower than the maximum frequency of the output voltage of the converter (see
datasheet).

Request 3
Maximum voltage check
The maximum voltage required by the motor is calculated in a precautionary formulation.
The phase voltage2 V , assuming Id = 0 and Iq = I, is:
√
V = (j · RI + j · nΩmax Ψm − nΩmax Ls I) 3

Considering that Kt = 3nΨm , it can be replaced obtaining the following equation, where the value
of Kt , downgraded in temperature, is called Kt,hot :
q √
Vm = (Rs (θmax ) · ITmax + Ωmax · Kt,hot /3)2 + (nΩmax Ls · ITmax )2 · 3

where:
• Kt,hot is the Kt of the motor at θmax (reduced by the effect of temperature)
 
θmax − 20
Kt,hot = Kt − 0.09
100

• ITmax,hot is the current required at maximum torque in hot condition

Tmax
ITmax,hot =
Kt,hot

In the above equation written through the rational approach, the following contributions are
present:
a) Phase voltage drop on the inductance (at maximum speed and maximum current in hot
condition):
ωpk · Ls · ITmax,hot
b) Phase voltage drop on the resistance in hot condition when the motor provides the maximum
torque:
Rs (θmax ) · ITmax,hot
c) Phase back-emf at maximum speed:

Kt,hot
Ωmax
3
The maximum phase voltage required by the working cycle of the motor has to be lower than the
maximum voltage provided by the converter. The maximum voltage provided by the converter (as
reported in datasheet) is:

Vmax,converter = 0.98 · Ac Input Voltage

2 Notice that the synchronous inductance L is half the winding inductance L
s w that you can find on motor
datasheet.

34
Reports on the laboratories R. Desideri, L. Scocco

Notice that the above formulation is very precautionary, since it is not said that the cycle requires
the maximum torque at maximum speed.
In fact, considering the motor speed and torque profile (known from exercise 2):

Motor torque

6.4722
T [Nm]

X 0.925
2.3810 Y 1.8473

X 0.325
Y -1.7103
-1.8473
0.15 0.25 0.4 0.6 0.75 0.85 1
t [s]

Figure 12: Angular speed, acceleration and torque profile of the motor
Notice that the maximum torque is not required at the maximum speed.

Then, we can evaluate the following:

• Voltage required by the motor at maximum torque and in hot condition (the speed ΩT max
is the speed of the motor when it is providing the maximum torque):
q √
Vm (Tmax ) = (Rs (θmax ) · ITmax + ΩTmax · Kt,hot /3)2 + (nΩTmax Ls · ITmax )2 · 3
• Voltage required by the motor at maximum speed and in hot condition (the current IΩmax
is the current required by the motor at maximum speed):
q √
Vm (Ωmax ) = (Rs (θmax ) · IΩmax + Ωmax · Kt,hot /3)2 + (nΩmax Ls · IΩmax )2 · 3

If both these rational voltage values are lower than the maximum voltage provided by the converter,
then the voltage check is over.

Request 4
Current RMS check From the absorbed current profile, the maximum and RMS current values
of the working cycle, with the same procedure followed for the motor sizing, can be evaluated.
The RMS current value can be derived from the definition:
s Z tcycle
1
Irms = I(t)2 dt
tcycle 0

35
Reports on the laboratories R. Desideri, L. Scocco

or from the following equation:

Trms
Irms =
Kt,hot
The nominal current of the converter has to be at least equal to the RMS current value calculated
above.

Request 5
Peak and average current check
The average current of the working cycle is the following:
Tavg
Iavg =
Kt,hot
It has to be lower than the nominal current of the converter.
Tmax
Imax =
Kt,hot
The peak current of the working cycle has to be lower than the peak current of the converter.

Request 6
Thermal Analysis
The magnitude current profile required by the motor can be drawn starting from the torque profile,
as:
|T (t)|
I(t) =
Kt,hot
The current profile that is obtained from exercise 2 is the following:

Current profile

I(t)
10.9 I n, conv
I rms
9.5839
I [A]

3.5244
3.1091
2.7354

0
0 0.2 0.4 0.6 0.8 1 1.2
t [s]

Figure 13: Current profile compared to the nominal current and the RMS value

36
Reports on the laboratories R. Desideri, L. Scocco

Then, the time constant of the converter can be calculated with the following equation, by using
the nominal and maximum current that the chosen inverter can provide
−tovld, conv
τ= 2 2
!
Iovld, conv − In, conv
ln 2
Iovld, conv

where Iovld, conv is the maximum value of output current for time tovld, conv , while In, conv is the
continuous output current of the inverter (see table of the inverter).
If the inverter time constant is at least 10 times higher than the overload time we can immediately
conclude that the thermal check is over.
If the time constant of the converter can be compared with the overload time, then the overheating
has to be considered through ??. In this case, we should analyze the current waveform and
approximate it with constant current parts and then apply ??.
The above time constant has been derived with the following assumptions. The converter over-
heating is proportional to the square of delivered current:

∆θ = Kth I 2
h i
where Kth is the thermal constant of the converter A°Crms
.
The maximum over temperature will then be at the nominal current In of the inverter:

∆θmax = Kth In2 (1)

Assuming that the thermal circuit of the converter can be represented with a single time constant
model, the over temperature profile can be expressed with the following equation:
t
 
−
∆θ = Kth I 2 1 − e τ  (2)

The maximum current delivered by the converter Imax for a short time tmax will lead the converter
to the maximum working over temperature.
tmax
 
−
2
∆θmax = Kth Imax 1 − e τ  (3)

Where Imax and tmax values can be derived from tables. Combining the equations ?? and ?? the
time constant of the inverter is defined.

Request 7
Choice of braking resistance
First of all, the instantaneous power of the working cycle has to be calculated:

P (t) = T (t) · Ω(t)

37
Reports on the laboratories R. Desideri, L. Scocco

The power profile of the motor over the cycle (from exercise 2) is the following:

Figure 14: Power profile and highlight on the negative power section

Motor power
2283.2171

1200
P [W]

604.7134

-604.7134
0.15 0.25 0.4 0.6 0.75 0.85 1
t [s]
Motor negative power
0
P [W]

-498.8926

-604.7134
0.15 0.25 0.4 0.6 0.75 0.85 1
t [s]

From the power profile, we can notice that the moments in which the energy goes back from the
motor to the grid are the ones in which the instantaneous power is negative. In the above figure,
there are two time intervals in which the power is negative. We have to calculate the extremities
of these intervals and by integration of power, the energy that has to be dissipated on braking
resistance can be derived. The energy to be dissipated on braking resistance can be calculated
with the following equation:
Z tbrake, 1, f

Ebrake, 1 = P (t)dt
tbrake, 1, i

where tbrake, 1, f e tbrake, 1, i are the extremities of interval in which the power is negative. The sum
of the above calculated energies provides the total energy of the working cycle to be dissipated on
the braking resistance. In this example, we are going to calculate two energies and the total one:

Ebrake = Ebrake, 1 + Ebrake, 2

Assuming to choose a given braking resistance, we can derive from datasheet the following data:
• PNBR nominal braking power
• RBR braking resistance value
• EBR maximum energy to be dissipated while braking
The maximum dissipation time at nominal braking power is:

EBR
tN BR =
PN BR

A) if the braking times are lower than time tNBR , the following procedure can be considered:

38
Reports on the laboratories R. Desideri, L. Scocco

A1) the braking energies of the working cycle have to be lower than the maximum energy
EBR to be dissipated on the resistance.
That is: Ebrake, 1 < Ebrake, conv and Ebrake, 2 < Ebrake, conv
A2) The average power of the working cycle has to be lower than the nominal power of the
resistance.
That is:
Ebrake = Ebrake, 1 + Ebrake, 2
total energy of the cycle in this example.
If these two conditions are satisfied, then the braking resistance is suitable for the
working cycle.
B) If the braking times are higher than time tNBR , we have to choose a braking resistance with
a nominal power PNBR higher than the maximum braking power of the working cycle.

In this exercise, the allowed braking time tNBR is much higher of the braking times
tbrake, 1 and tbrake, 2 : so it corresponds to Case A.

39
Reports on the laboratories R. Desideri, L. Scocco

Results

Symbol Value Units Description

Voltage of the power supply
Drive XVy 21020 [-] Model of the converter chosen
Uln 230 [V] Input voltage
Vout, conv 225.4 [V] Output voltage = 0.98Uln
Vn 185 [V] Nominal voltage of the motor
Maximum voltage and frequency
n 4 [-] Pole pairs
fmax, mot 320.9 [Hz] Maximum frequency of the motor
fmax, conv 450 [Hz] Maximum frequency of the converter
Vmax (ω(Tm, max ), Tm, max ) 134.6 [V] Max voltage evaluated at the max torque
Vmax (ωmax , Tm (ωmax ) 209.4 [V] Max voltage evaluated at the max speed
Nominal power
Pn, out 3000 [W] Output power of the converter in continous service
Pn, motor 1097 [W] Nominal power of the motor
Nominal current
In, conv 10.9 [A] Nominal current converter
Irms 3.109 [A] RMS current of the cycle
Iavg 1.175 [A] Average current of the cycle
Maximum current
Iovld, conv 21.8 [A] Overload current of the converter
tovld, conv 1 [s] Maximum overload duration of the converter
Imax 9.584 [A] Maximum current in the cycle
tovld, duration 0 [s] Overload duration in the cycle
Thermal check
τ 3.476 [s] Time constant of the converter
Braking resistor
External braking resistor MRI/T900 68R [-] Model of the braking resistor chosen
Pbrake, conv 900 [W] Braking power of the converter
Pbrake, avg 291.3 [W] Average braking power of the cycle
Ebrake, conv 33 [kJ] Braking energy of the converter
Ebrake, 1 25.65 [J] Braking energy in the first period of time
Ebrake, 2 44.45 [J] Braking energy in the second period of time
Ebrake 70.1 [J] Total braking energy
tNBR 36.67 [s] Maximum braking time at nominal braking power
tbrake, 1 0.09069 [s] Duration of the first braking
tbrake, 2 0.15 [s] Duration of the second braking
tbrake, tot 0.2407 [s] Total braking time

40
Reports on the laboratories R. Desideri, L. Scocco

4 Report 4

A DC electrical drive has to cover the duty cycle (current vs time and speed vs time) reported in
figure ??. The DC motor is driven by a one quadrant DC/DC (chopper) converter based on IGBT
technology.

Figure 15: Current and speed profile.

The duty cycle has: a 0.4 s starting time with a 35 A current, a 2.6 s constant speed part (4005
rpm) with 25 A, and a rest time of 2 second with 0 A. In the last part the motor is coasting to
zero speed with the converter switched off (no pulsing).
Motor data:
• Nominal speed: 4100 rpm
• Rated Power: 2500 W
• Rated Voltage: 180 V
• Torque constant: 0.4 Nm
A
• Motor Resistance: 0.35 Ω
• Motor Inductance: 1.25 mH

41
Reports on the laboratories R. Desideri, L. Scocco

Chopper data:
• Input Voltage: 300 V
• Rated Power: 2800 W
• Switching frequency: fc = 8 kHz
• IGBT gate voltage: VGE = 17 V

Heat sink data:

K
• Thermal resistance: 0.4 W
• Thermal time constant: 15 min
• Ambient temperature: 45°C

For the IGBT data refer to the annexed data sheet for component SKM 75GB063D. The electric
and thermal verification are required for the suggested IGBT (SKM 75GB063D) answering the
following questions:
1. Check the voltage and current rated values for IGBT and diode.
2. Check the heat sink temperature during duty cycle (has to be lower than 80°C). Check the
IGBT and DIODE losses (conduction + switching). They have to be lower than the maximum
value specified in data sheet.
3. Check the IGB and DIODE maximum junction temperature (they have to be less than 125°C)

We’re going to refer to the following electrical model of the 1-quadrant chopper.

Figure 16: Electrical circuit of the DC motor-chopper electrical system.

The annexed catalogue reporting the physical characteristics of the IGBT and the diode is the
following

42
Reports on the laboratories R. Desideri, L. Scocco

Figure 17: Catalogue, table 1.

Figure 18: Catalogue, table 2.

43
Reports on the laboratories R. Desideri, L. Scocco

Figure 19: Catalogue, table 3.

Request 1
Once the working cycle and the circuit configuration are known, the first validation is about the
voltage class of the device and the conduction current. First of all, calculate the supply voltage
V1 needed for the application and then the duty cycle δ that the control has to provide to the
components of the chopper (IGBT and diode) to follow the working operations:

V1 (t)
δ=
V
Since the cycle is featured by time intervals in which the current is constant, we can neglect the
inductive contribution in the electrical equation of the DC motor (we also neglect the dynamics
associated to the step variation of the current profile).

V1 = RI1 + E = RI1 + Kt Ω
where Kt is the torque constant.
So, knowing the speed and current profiles of the motor, it’s possible to derive the voltage and the
duty cycle profile that the chopper has to guarantee.

44
Reports on the laboratories R. Desideri, L. Scocco

Voltage [V]
250

200 X3
Y 176.511
180.0110

150

100

50

12.25
0
0 0.4 1 1.5 2 2.5 3 3.5 4 4.5 5

Duty cycle
0.9

0.8

0.7
X3
Y 0.58837
0.6000

0.5

0.4

0.3

0.2

0.1
0.0408

0 0.4 1 1.5 2 2.5 3 3.5 4 4.5 5

45
Reports on the laboratories R. Desideri, L. Scocco

Now, the electrical validation of the chopper can be carried out. For IGBT transistor different
voltage classes exist: 600 V, 1200 V, 1700 V, 2500 V e 3300 V. Typically the safety margin that
is used is about twice the nominal supply voltage. In this application, the peak voltage that the
chopper has to provide is 180.011 V. It means that the first class (600 V) is suitable with a safety
margin equal to 3.33.
Considering the nominal current validation, we have to compare the nominal current of the device
with the maximum current of the working cycle. The maximum current of the cycle (35 A) is
lower of the nominal current the diode can stand at 25° C (100 A) and also lower than the nominal
current the diode can stand at 80° C (75 A).
So, both the IGBT and the diode are verified.

Request 2
Semiconductor devices cause losses that lead them to heat up during the working cycle. For this
reason, the device has to be equipped with a well dimensioned heat sink, so that the working
temperatures of the IGBT, the diode and the heat sink itself are lower than those specified by
manufacturers, in order to avoid deterioration and finally their breaking.
Once the losses profile has been identified, the thermal analysis is carried out using an equivalent
electric-thermal circuit that takes into account the dynamic phenomena, since the semiconductor
and heat sink time constants are very different.
Semiconductor devices time constants are very small and the period of each phase of working cycle
exceeds their value, thus they can be neglected and the junctions can be considered at thermal
steady state because their dynamics are too sudden.
Conversely, the thermal time constant of heat sink is very high (15 minutes) and since the overall
period of the working cycle is very short (5 s), we can also neglect its dynamics. So, even the heat
sink can be considered at thermal steady state.
Thus, a steady state thermal model is used and, in order to define losses, the r.m.s. current and
the average value of duty cycle are calculated. So, the steady state value of heat sink temperature
can be derived, verifying that it is lower than a maximum admittable value (80° C). The equivalent
thermal circuit of the converter is shown in figure ??:

Figure 20: Thermal model by exploiting the electrical analogy.

Power electronics devices have two different types of power losses: conduction losses (similar to
Joule losses) and switching losses (corresponding to the energy lost during the turn-on and turn-off
transitions).
Conduction losses can be derived multiplying the current flowing through the device for the voltage
across its terminals. Notice that the voltage-current relationship is not linear and depends on the
device (IGBT or diode). Finally, to compute conduction losses the following relationship will be
used, where voltage values are derived from voltage-current characteristics reported in datasheet.

46
Reports on the laboratories R. Desideri, L. Scocco

Pcond,IGBT = Vce (Ic )Ic

Pcond,DIODE = Vak (Id )Id
Switching losses are calculated considering that during a turn-on or turn-off transient of any semi-
conductor device, there is an energy dissipation. Once the energy used in these phases is known,
the calculation of power dissipation has to be conducted by dividing the total energy in a switching
period for the period itself. Since the switching period is equal to the inverse of the frequency, the
power loss is calculated by multiplying the energy dissipated in a working cycle for the switching
frequency. The energy values are provided in the datasheet by manufacturer, as a function of the
conduction current. In particular, calculation has to be set as shown below:

Pswitch,IGBT = [Eon (Ic ) + Eoff (Ic )]fc

Pswitch,DIODE = Erec (Id ) · fc
To make a thermal analysis, we have to first calculate the r.m.s. current value and the average
duty cycle:
s
1 T 2
Z
Irms = I (t)dt
T 0

1 T
Z
δavg = δ(t)dt
T 0

IGBT Losses
Data required to calculate power losses are reported in the attached datasheet. Since characteristic
curve depends on voltage Vce , let us consider the one related to a voltage equal to 17 V.
Once Irms is known, Vce voltage can be derived from datasheet, by using voltage-current
characteristic. Since the characteristics are expressed as a function of junction temperature, the
one related to the maximum temperature is used in order to make calculation with the highest
safety margin.

47
Reports on the laboratories R. Desideri, L. Scocco

Figure 21: Characteristic curves of the IGBT as a function of Irms current and the gate voltage
VGE .

In order to calculate the current, notice that IGBT works with a series of on/off phases, during
a period, determined by the switching frequency (8 kHz), where on time and off time are defined
by the duty cycle δ required at that given instant, based on the drive characteristic. Thus, for
conduction losses it is necessary to calculate an average conduction current through multiplying
the Irms current by the average duty cycle.

Pcond,IGBT = Vce (Irms )Irms δavg

Switching losses are derived using graphs of turn-on/-off energy as a function of current. Notice
that, typically, energies in graphs are expressed in mJ.

Figure 22: IGBT switching losses in terms of energy.

Thus the energy losses can be calculated as follows:

Pswitch,IGBT = Eon fc + Eoff fc

Finally, the IGBT total losses are:

Ptot,IGBT = Pcond,IGBT + Pswitch,IGBT

Diode Losses
The same procedure can be followed to calculate diode losses. However, for this device, the average
current flowing through it is inversely related to the duty cycle (the diode is on during the off phase)
according to the following relation:

Iavg,DIODE = (1 − δavg )Irms

The voltage across the diode terminals can be derived from datasheet as a function of the RMS
current considering the worst condition: so the curve corresponding to 125°C max duty cycle is
considered.

48
Reports on the laboratories R. Desideri, L. Scocco

Figure 23: Characteristic curves of the diode as a function of Irms and the operating temperature.

Thus, for conduction losses calculation is:

Pcond,DIODE = Vak (Irms ) · Irms (1 − δavg )

For diode switching losses consider only the turn-off phase. Notice that, typically in diode datasheet
the value Erec is not directly provided, as for the IGBT, but the corresponding value has to be
derived from characteristic data provided by the manufacturer (furthermore, typically this energy
value is lower with respect to the one of the IGBT). In figure ??, the voltage and current profile
during the on- and off- phases of the valves are shown. Notice that the power losses during diode
turn-off is only the dashed area defined as WCof f .

49
Reports on the laboratories R. Desideri, L. Scocco

Figure 24: Voltage, current and power profiles for a diode. The highlighted area represents the
energy dissipated by the diode.

The time derivative of current, provided by datasheet didtF ,is used as the diode characteristic
parameter, in order to calculate the switching losses. In this case, the curve corresponding to the
gate resistance RG = 15Ω is consdered, while IF is still the RMS current calculated before.

Figure 25: diF

dt parameter of the diode, as a function of the RG resistance and the Irms current

So, the energy losses of the diode during the switching off phase can be calculated as:
2 2
1 1 V · Irms · ∆t 1 V · Irms
Erec = V Irms toff = =
2 2 ∆I 2 di
dt
Pswitch,DIODE = Erec · fc
The total losses of the diode are:

Ptot,DIODE = Pcond,DIODE + Pswitch,DIODE

Once losses in the devices are known, considering the thermal model in figure ??, the steady state
temperature of the heat sink can be derived:

θHS = Rth,HS · (Ptot,IGBT + Ptot,DIODE ) + θamb

And then, a validation of the heat sink temperature (that has to be lower than 80° C) can be done.

Request 3
Finally, the temperature reached by each valve in the worst working condition can be verified.
To make this analysis, first of all we have to define which are the worst conditions for the two

50
Reports on the laboratories R. Desideri, L. Scocco

devices inside the working cycle. Analyzing power dissipations expressions of conduction losses,
since they depend on duty cycle, either the IGBT or the diode appear to be more stressed. In fact,
in IGBT the average current is directly dependent on duty δ = tTon , while in diode the average
current is proportional to (1-δ). Thus, in order to calculate IGBT power losses, we have to consider
a maximum δ while in order to calculate diode power losses, we have to consider a minimum δ
Reasonable values for the case under examination are 0.02 for minimum δ and 0.98 the maximum
one. The above values are determined in relation to the minimum turn-on and turn-off time of
semiconductors. In fact, the switching times (from on to off and vice versa) are different from zero,
then once given the one pulse it must last for a time sufficient to ensure both the switching on and
off (otherwise there is the risk to control the shutting down when it is not yet finished the turn on
phase). Considering that the switching frequency is equal to 8000 Hz (and thus the time period is
125 µs), the time corresponding to the 2% of duty cycle is equal to 2.5 µs. Turn on and turn off
times are in the order of microseconds.
The maximum temperature of each device can be calculated by summing the overtemperature of
each device (the product of the total power losses for the thermal resistance) to the heat sink
temperature. The Rth,IGBT e Rth,DIODE values, that are the thermal resistances between the
semiconductor junction and the external case of the device, for diode and IGBT respectively, are
defined in datasheet as Rth,JC .

IGBT Losses
Power losses calculation:

Imax = 35 A, δmax = 0.98

Pcond,IGBT,max = Vce (Imax )Imax δmax
Pswitch,IGBT,max = (Eon + Eoff )fc

Ptot,IGBT,max = Pcond,IGBT,max + Pswitch,IGBT,max

Verify if the maximum total power losses value calculated for the IGBT are lower than the maximum
limit provided by the datasheet. Furthermore, the temperature value can be derived:

θIGBT = Rth,IGBT · Ptot,IGBT,max + θHS

Finally, verify that the maximum junction temperature of the IGBT in the working cycle is lower
than 125° C.

Diode Losses
The method to calculate the power dissipation in diode is the same that for the IGBT:

Imax = 35 A, δmin = 0.02

Pcond,DIODE,max = Vak (Imax )Imax (1 − δmin )
2
1 V · Imax
Erec =
2 di
dt
Pswitch,DIODE,max = Erec · fc
Ptot,DIODE,max = Pcond,DIODE,max + Pswitch,DIODE,max
Verify if the maximum total power losses value calculated for the diode are lower than the maximum
limit provided by the datasheet. Furthermore, the temperature value can be derived:

θDIODE = Rth,DIODE · Ptot,DIODE,max + θHS

Finally, verify that the maximum junction temperature of the diode in the working cycle is lower
than 125° C.

51
Reports on the laboratories R. Desideri, L. Scocco

It is slightly not verified but one must remember that this condition is extreme and cannot happen
in reality, so the system is practically always verified. In the following graph we consider the actual
temperature reached by the diode, the IGBT and the heat sink varying with the duty cycle from
0.02 to 0.98, neglecting the penalty of considering the most onerous duty cycles for both devices
at the same time; and it can be seen that the maximum temperature never reaches 125° C

110
Heat sink temperature
IGBT temperature
105 Diode temperature

100
Temperature [°C]

95

90

85

80

75
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Duty cycle

52
Reports on the laboratories R. Desideri, L. Scocco

Results
Symbol Value Units
Rated voltage and current
nss 4005 [rpm]
Imax 35 [A]
I1 25 [A]
ωss 419.4 [rad/s]
Kemf 0.4 [V·s/rad]
Vmax 180 [V]
V1 176.5 [V]
δavg 0.3316 [-]
Heat sink temperature during duty cycle
IRMS 20.57 [A]
δ 0.3316 [-]
VCE (IIGBT ) 1.372 [V]
VAK (Idiode ) 1.337 [V]
Ploss, IGBT, cond 9.357 [W]
Ploss, diode, cond 18.38 [W]
Eon (IIGBT ) 0.0009508 [J]
Eoff (IIGBT ) 0.0009667 [J]
diF /dt(Idiode ) 1.044·109 [A/s]
Erec 6.078·10−5 [J]
Ploss, IGBT, switch 15.34 [W]
Ploss, diode, switch 0.4862 [W]
Ploss, IGBT, tot 24.7 [W]
Ploss, diode, tot 18.87 [W]
Ploss, tot 43.56 [W]
θHS 62.43 [°C]
θIGBT 71.07 [°C]
θdiode 73.74 [°C]
Heat sink temperature with max conditions
IIGBT 35 [A]
Idiode 35 [A]
δmax 0.98 [-]
VCE (IIGBT ) 1.657 [V]
VAK (Idiode ) 1.495 [V]
Ploss, IGBT, cond 56.83 [W]
Ploss, diode, cond 51.3 [W]
Eon (IIGBT ) 0.001439 [J]
Eoff (IIGBT ) 0.001362 [J]
diF /dt(Idiode ) 1.308·109 [A/s]
Erec 1.405·10−4 [J]
Ploss, IGBT, switch 22.41 [W]
Ploss, diode, switch 1.124 [W]
Ploss, IGBT, tot 79.24 [W]
Ploss, diode, tot 52.42 [W]
Ploss, tot 131.7 [W]
θHS 97.66 [°C]
θIGBT 125.4 [°C]
θdiode 129.1 [°C]

53
Reports on the laboratories R. Desideri, L. Scocco

5 Report 5
A PMDC motor has the following main data:
• Nominal speed: Ωn = 3000rpm = 314.2rad/s
• Nominal torque: Cn = 9.6N m
• Nominal current: In = 30A
• Motor constant: K = CInn = 0.32N m/A
• Armature inductance: L = 0.003H
• Armature resistance: R = 0.32Ω
• Rated E.M.F. : E=105 V
• Rated Voltage: Vn = 110V
• Shaft inertia: J = 0.0013kgm2

The motor is controlled via an 8 kHz chopper and a digital controller based on PI closed loop
regulators (one for current control and another for speed control). The following points have to be
solved:
1. calculate the gain of the current (proportional and integral) with a band pass width of 500
rad/sec and a phase margin of 80°;
2. calculate the gain of the speed PI (proportional and integral) with a band pass width of 100
rad/sec and a phase margin of 60°.
Repeat the previous points by considering the effects of non-idealities.
The current speed loop is executed with a Ts1 of 0.25 ms and the speed loop is executed with Ts2
equal to 1 ms.
The current value is sampled with a 12 bit Analog to Digital Converter.
The speed is measured via a 2000 pulses per revoulution sensor and the chopper is controlled with
10 bit digital comparator for generating the PWM pulses.

Block Diagram
Any constant flux electrical machine, controlled via a cascade control with industrial PID
regulators can be expressed with the block diagram shown in Figure ??.

Figure 26: Block diagram of a constant flux electrical machine control.

The relation between the reference torque and the reference current is expressed through the K(ie )
term, that is constant if the flux is constant. In the above scheme the following assumptions are
considered: the electronic converter, the current and speed measurement system and the electronic
system that implements the regulation are ideal, without delays and free from the effects of sam-
pling. The procedure able to tune regulators, based on linear systems control theory, valid under
the above mentioned hypothesis, is provided below.

54
Reports on the laboratories R. Desideri, L. Scocco

Request 1
Tuning of the Current Regulator
The current regulator block diagram can be schematized as in Figure ??.

Figure 27: Internal Loop.

The system under control transfer function is:

1
BI (s) =
R+L·s
while the controller transfer function is:

RI (s) = KP,I + KI,I /s

Thus, the internal open loop transfer function is:

G · H(s) = GI (s) = RI (s) · BI (s)

In order to properly tune the current controller, the following conditions have to be imposed:
1. |GI (s)| = 1
2. ∠(GI (s)) = −π + ϕm,I

   
 KI,I   1 
KP,I + · =1
 
 jωp   R + jωp,I L 
   
KI,I 1
∠ KP,I + +∠ = −π + ϕm,I
jωp R + jωp,I L
So,exploiting the Eulero formula, we can write:
 
KI,I
KP,I + · |BI (jωp,I )|ej·∠(BI (jωp,I )) = ej(−π+ϕm,I )
jωp
 
KI,I 1
KP,I + = · ej(−π+ϕm,I −∠(BI (jωp,I ))
jωp,I |BI (jωp,i )|
Separating the real from the imaginary part, we obtain the current regulator gains as a function
of band pass width ωp,I and phase margin ϕm,I :
1
KP,I = · cos(−π + ϕm,I − ∠(BI (jωp,i )))
|BI (jωp,i )|
ωp,I
KI,I =− · sin(−π + ϕm,I − ∠(BI (jωp,i )))
|BI (jωp,i )|

55
Reports on the laboratories R. Desideri, L. Scocco

The open loop transfer function of current regulator is defined as follows:

 
KI,I
GI (s) = KP,I + · BI (s)
s

After calculating proportional and integral gains, Bode diagrams of the open loop transfer function
can be drawn in order to verify band pass width and phase margin, shown in Figure ??.

Ideal Bode diagram - Current loop

20·log|GH I( )|
100
db(|GH|)

50

-50
10-1 100 101 102 103 104
[rad/s]
(GH I( ))
-90
(GH) [deg]

-95

-100

-105

-110
10-1 100 101 102 103 104
[rad/s]

Figure 28: Ideal Bode diagram of the current loop

Finally, it’s important to define the closed loop transfer function of the current regulator:

 
KI,I
KP,I + · BI (s)
s
LI (s) =  
KI,I
1 + KP,I + · BI (s)
s

Request 2

Tuning of the Speed Regulator

The tuning of speed regulator can be done as for the current regulator, by taking into account the
block diagram shown in Figure ??, where the current loop can be replaced with the closed loop
transfer function LI (s).

56
Reports on the laboratories R. Desideri, L. Scocco

Figure 29: External Loop.

The system under control transfer function BV (s), consists of the current loop and the mechanical
loop where the resistant torque can be neglected by considering it as a disturbance:

LI (s) · Kc 1 LI (s)
BV (s) = =
Kc J ·s J ·s
To consider the effects of resistance torque, an “a posteriori” analysis can be carried out, once the
loop is tuned. If the resistant characteristic is linear, or can be linearized, it can be considered
in the model by adding it to the inertia. Thus, under the assumption that the regulator is a PI
regulator, the proportional and integral gains KP,V and KI,V can be derived using band pass width
and phase margin.
Again, in order to properly tune the speed controller, the following conditions have to be imposed:
1. |GV (s)| = 1
2. ∠(GV (s)) = −π + ϕm,V

   
 KI,V   
KP,V +  · BV (jωp,V ) = 1
 
 jωp   
 
KI,V
∠ KP,V + + ∠ (BV (jωp,V )) = −π + ϕm,V
jωp
So,exploiting the Euler formula, we can write:
 
KI,V
KP,V + · |BV (jωp,V )|ej·∠(BV (jωp,V )) = ej(−π+ϕm,V )
jωp
 
KI,V 1
KP,V + = · ej(−π+ϕm,V −∠(BV (jωp,V ))
jωp,V |BV (jωp,V )|
Separating the real from the imaginary part, we obtain the current regulator gains as a function
of band pass width ωp,V and phase margin ϕm,V :
1
KP,V = · cos(−π + ϕm,V − ∠(BV (jωp,V )))
|BV (jωp,V )|
ωp,V
KI,V =− · sin(−π + ϕm,V − ∠(BV (jωp,V )))
|BV (jωp,V )|
As before, the open loop transfer function of speed regulator can be expressed as follows:
 
KI,V
GV (s) = KP,V + · BV (s)
s
Again, to verify the tuning accuracy, the Bode diagram of the open loop transfer function of speed
regulator can be drawn. The result is shown in Figure ??.

57
Reports on the laboratories R. Desideri, L. Scocco

Ideal Bode diagram - Speed loop

20·log|GH V( )|
150
100
db(|GH|)

50
0
-50
-100
10-1 100 101 102 103 104
[rad/s]
(GH V( ))
-100
(GH) [deg]

-120

-140

-160

-180
10-1 100 101 102 103 104
[rad/s]

Figure 30: Ideal Bode diagram of the speed loop

Finally, the closed loop Transfer Function of the entire system is:
 
K
KP,V + I,Vs · BV (s)
LV (s) =  
K
1 + KP,V + I,V s · BV (s)

Since we are using a cascade control, attention should be paid when choosing the dynamic specifi-
cations (the outer band pass width has to be 5-10 times lower with respect to the inner one) and
when considering non-idealities, discussed in next paragraphs.

58
Reports on the laboratories R. Desideri, L. Scocco

Non-ideality effects
In order to evaluate (for both the regulators tuning) the presence of real electronic converter and
measurement systems, the following complete block diagram is now considered in the following
Figure ??

Figure 31: Real Block Diagram.

Electronic Converter The electronic converter is made of switching devices, thus it is a time-
variant system, but it can be easily modeled as linear. Thus, we are going to find an equivalent
transfer function that takes into account the delays in providing the required voltage and the effects
of harmonic distortion. Considering the chopper shown in Figure ??.

Figure 32: Chopper model.

If the chopper switching frequency is fc, the switching period is expressed as:

1
Tc =
fc
The commutations are made by comparing a saw tooth signal with a reference signal,as shown in
Figure ??.
If at time t1 , where the commutation has already begun, the reference decreases, it is not possible
to immediately modify the power on time, but we have to wait for the next commutation at time
t2 . The same is for an increasing in reference signal.
Statistically, it is possible to assume an average delay time Tmr equal to half the switching period:

1
Tmr = Tc
2
This is the average time that the converter needs to modify the output level, following a reference
variation.

59
Reports on the laboratories R. Desideri, L. Scocco

Figure 33: Chopper commutation.

Thus, the converter transfer function taking into account its time delay can be expressed as in the
following:

C(s) = g · e−s·Tmr
If the band pass width is enough lower than the switching frequency, we can linearize this expression
and obtain:
g
C(s) =
1 + s · Tmr
The value of gain g is the ratio between the maximum output voltage, when the chopper duty cycle
is equal to 1, and the corresponding reference voltage signal vref. This value is the maximum value
of saw tooth signal.

Current and Speed Transducers The measurement systems represent physical quantities
with voltage signals that the control system can manage. The measurement block gain, typically
expressed as H(s), depends on the ratio between the maximum value achievable by the corre-
sponding signal and the maximum value of the physical measured quantity.

This is true if the controller is made of analogic systems. If we are using digital systems, the
gain is the ratio between the maximum value of the measured quantity and the digital value
detected by the digital system. If a per unit notation (very useful in real applications) is adopted,
a rated value of physical quantity is chosen to correspond to value 1 of the digital system.
Thus, when the machine is working in nominal condition, all the quantities are equal to 1 in
the control system. Each transducer has a finite band pass width and a filtering system, thus,
it can be expressed with an equivalent transfer function with one or more poles in the denominator.

Considering the current transducer, and using a per unit, p.u., normalization, the current
transducer gain is 1 over the nominal current, thus HI = I1n = hI . This means that when the
sampling current is equal to 1, in the motor the nominal current flows.

The current transducer has to be able to measure up to the maximum current with a margin of
20% for the regulation dynamic, thus the maximum measurable current and the full scale of the
transducer is Imax,m = 3 · Imax , corresponding to a per unit value of 3, that is three times higher
than the nominal current.
Considering the speed transducer, and using a per unit, p.u., normalization, the gain is 1 over the
nominal speed, thus Hω = Ω1n = hω . the maximum measurable speed and thus the full scale of the
transducer has to be equal to the maximum one with a margin of 10% : Ωmax,m = 1.1Ωmax .

Digital Implementation Effects Typically, most of the control systems exploit digital elec-
tronic technologies to implement regulators. The so obtained structure has the advantages of
being flexible, reprogrammable and easily modifiable. Furthermore, controllers are cheap, not so
complex and they need only a few space.
Considering the regulator tuning, notice that the overall system (regulator-converter-motor) has a

60
Reports on the laboratories R. Desideri, L. Scocco

discrete time part and a continuous time part, thus, from a time point of view, we have a hybrid
overall system. The calculations are different for the two parts, then, before tuning, we have to
transform the discrete time part in continuous time or vice versa obtaining a consistent model like
the following one in Figure ??:

Figure 34: Complete Block Diagram including digital implementation.

Notice, in Figure ??, that typically, the outer and inner loop controllers have different sampling
time periods, since the outer loop dynamic has to be slower, thus it has a higher period with
respect to the inner one.

The presence of a digital system causes the need to model the effects due to the analog to digital
and the digital to analog conversion through a transfer function with a pole equal to 1 over the
time of the corresponding loop.
The transfer functions for speed loop and for current loop are the following:
1
TI (s) =
1 + s · Ts1
1
TV (s) =
1 + s · Ts2
Now, we’re going to tune again the current and speed controls considering all these non-idealities;
for the calculation of the Transfer Functions, we’ll refer to the following block diagram in Figure
??:

Figure 35: Final Block Diagram including digital implementation.

Considering the same permanent magnet DC motor, now we consider the following additional data:
• Chopper switching frequency: fc = 8000Hz
• Speed Loop Sampling time: Ts,V = 0.001ms
• Current Loop Sammpling time: Ts,I = 0.0002ms

61
Reports on the laboratories R. Desideri, L. Scocco

Request 3
Current Regulator Tuning
The tune method is the one explained in the previous paragraph where the transfer functions have
to be revised because of the presence of non-ideality elements. For example, the relation expressing
the open loop transfer function of the controlled system becomes the one indicated in the following:
1 1 g 1
BI (s) = · C(s) · HI (s) · TI (s) = · · hI ·
Rs + s · Ls Rs + s · Ls 1 + s · Tmr 1 + s · Ts2
With the same mathematical steps, the following expressions are obtained:
1
KP,I = · cos(−π + ϕm,I − ∠(BI (jωp,I )))
|BI (jωp,I )|
ωp,I
KI,I =− · sin(−π + ϕm,I − ∠(BI (jωp,I )))
|BI (jωp,I )|
The open loop transfer function of current regulator is defined in the following.
   
KI,I KI,I 1 g 1
GI (s) = KP,I + · BI (s) = KP,I + · · · hI ·
s s Rs + s · Ls 1 + s · Tmr 1 + s · Ts2
Once the current regulator is tuned, Bode diagrams of the open loop transfer function can be
drawn in order to verify band pass width and phase margin. ??

Real Bode diagram - Current loop

20·log|GH I( )|

60
40
db(|GH|)

20
0
-20

10-1 100 101 102 103 104

[rad/s]
(GH I( ))

-100
(GH) [deg]

-120
-140
-160
-180

10-1 100 101 102 103 104

[rad/s]

Figure 36: Real Bode diagram of the current loop

Finally, it’s important to define the closed loop transfer function of the current regulator:

   
KI,I KI,I 1 g 1
KP,I + · BI (s) KP,I + · · ·
1 s s Rs + s · Ls 1 + s · Tmr 1 + s · Ts2
LI (s) =   =  
HI (s) KI,I KI,I 1 g 1
1 + KP,I + · BI (s) 1 + KP,I + · · · hI ·
s s Rs + s · Ls 1 + s · Tmr 1 + s · Ts2

62
Reports on the laboratories R. Desideri, L. Scocco

Request 4
Speed Regulator tuning
Using the closed loop transfer function of current regulator, it is possible to define the transfer
function of the system under control for the speed regulator:

LI (s) · Kc 1 LI (s)
BV (s) = · TV (s) · hω · = · TV (s) · hω
Kc J ·s J ·s

Again, following the previous mathematical steps, the expressions that allow to tune the speed
controller are the following ones:

1
KP,V = · cos(−π + ϕm,V − ∠(BV (jωp,V )))
|BV (jωp,V )|

ωp,V
KI,V = − · cos(−π + ϕm,V − ∠[BV (jωp,V )))
|BV (jωp,V )|

The open loop transfer function of speed regulator can be expressed as in the following:
 
KI,V
GV (s) = KP,V + · BV (s)
s

In order to verify the tuning accuracy, the Bode diagram of the open loop transfer function of
speed regulator can be drawn. The result is shown in Figure ??

Real Bode diagram - Speed loop

20·log|GH V( )|
100

50
db(|GH|)

-50

10-1 100 101 102 103 104

[rad/s]
(GH V( ))

-150
(GH) [deg]

-200

-250

-300

-350
10-1 100 101 102 103 104
[rad/s]

Figure 37: Real Bode diagram of the speed loop

63
Reports on the laboratories R. Desideri, L. Scocco

Results
Symbol Value Units Description
Ideal case
A
 
KP,I 1.422 Proportional gain current loop
V 
A1
KI,I 287.8 Integral gain current loop
V s
Nm 1

KP,V 0.1166 Integral gain speed loop
 rad/s s 
Nm 1
KI,V 4.77 Integral gain speed loop
rad/s s
Non ideal case
Tc 0.000125 [s] Switching period
Tmr 6.25·10−5 [s] Average time delay
g 110 [-] Gain: rated voltage divided by unitary reference voltage
hI 0.03333 [-] Current transducer gain
A
 
KP,I 0.4107 Proportional gain current loop
V 
A1
KI,I 47.9 Integral gain current loop
V s 
Nm
KP,V 1.336 Integral gain speed loop
 rad/s 
Nm 1
KI,V 31.89 Integral gain speed loop
rad/s s

64
Reports on the laboratories R. Desideri, L. Scocco

6 Report 6
Battery sizing for an electric city car

A city car is equipped with a battery power controlled electrical drive. The car can adopt lead
acid battery or lithium-ion battery. The city car is for urban-suburban purpose and has a load
capability of two persons, 40 kg of baggage and can afford a slope of 10%. The max speed is 90
km/h with a cruise speed of 50 km/h. The required range is 100 km in urban drive cycle and
the maximum allowed weight (including batteries) is 700 kg. For energy consumption evaluation
consider the simplified urban drive cycle reported below.

E.C.E regulations - simplified urban drive cycle

50

40
Speed [km/h]

30

20

10

0
7 47 54 84 98 198 212 240
Time [s]
Acceleration profile
4
Acceleration [km/h 2 ]

-2

-4
7 47 54 84 98 198 212 240
Time [s]

Figure 38: Urban cycle profile.

Both for lead acid and lithium battery the following points must be solved. Requests:
1. Calculate the drive cycle energy consumption (Wh/cycle).
2. Calculate the energy in the battery for the required range and verify if the battery weight is
compliant with the maximum allowed one.
3. Calculate the energy consumption for flat ground at max speed and the corresponding range.
4. Evaluate the cost for recharging the batteries.

Mechanical Data
• Cx = 0.3 Drag coefficient
• S = 2.4 m2 Frontal area
• vnom = 50 km/h nominal speed
• vmax = 90 km/h max speed
2
• a = 1 m/s acceleration and deceleration of the vehicle
• mt = 700 kg full load allowed mass of the vehicle
• mempty = 300 kg vehicle mass without load

65
Reports on the laboratories R. Desideri, L. Scocco

• ηt = 0.95 Transmission efficiency

• ηm = 0.92 Motor efficiency
• p = 10% Max slope
• µt = 0.015 tyres rolling resistance coefficient
2
• g = 9.81 m/s gravity acceleration
3
• ρ = 1.2258 kg/m air density
• mpeople = 2 · 75 kg Mass of 2 people on board
• mbag = 40 kg baggage mass

Electrical Data
• Vbatt = 80 V battery voltage
• Espec = 30 Wh/kg Pb-acid specific energy
• Pspec = 80 W/kg Pb-acid specific power
• Espec = 120 Wh/kg Li-ion specific energy
• Pspec = 240 W/kg Li-ion specific power
• QkWh = 0.15 e/kWh Electric energy cost
• N = 4000 rpm Motor rated speed
• ηbatt = 0.8 Charging-discharging battery and converter efficiency

Assumptions
1. Inertia is considered only when acceleration is positive (we suppose we do not have a regen-
erative braking)
2. ηbatt was considered only for Lithium Ion batteries, in order to keep a minimum of 20% of
State of Charge
3. The efficiency of the battery doesn’t influence the power (only the energy)
4. Consequently, when recharging the batteries, it won’t be necessary to replenish the remaing
20% of energy always present in the Lithium battery, therefore the cost will be calculated
accordingly.

Request 1
For energy consumption evaluation consider the simplified urban drive cycle, defined according to
E.C.E. regulation in Figure ??.

To evaluate the energy needed to perform this cycle, the forces to which the city car is subjected
during the cycle have to be analyzed.
In the acceleration and deceleration phases there is an inertia force given by:

Fin = mtot · a(t)

In sloping conditions, there is also a gravitational contribution:

Fg = mtot · g · sin(θ)
where θ can be easily derived from the slope as:

θ = arctan(p)
Other resistance forces are the rolling resistance and the aerodynamic drag:

Frr = mtot · g · cos(θ) · µt

1
Faero = ρCx Sv 2 (t)
2
where µt is the rolling resistance coefficient, ρ is the air density, Cx is the drag coefficient and S
is the frontal area of the car. From the above equations it can be noticed that the inertia force,
the weight contribution and the rolling resistance depend on the overall mass of the vehicle. The
total mass of the car is the sum of the unload vehicle (300 kg), the people (75 kg), the luggage (40

66
Reports on the laboratories R. Desideri, L. Scocco

kg) and the battery weight. We don’t known yet the weight of the battery, because we still have
to select and size it.
Thus, we are going to follow an iterative procedure, starting from the speed data and assuming a
value of total force, and therefore a value of total power demanded by the motor.

Ftot = Fin + Fg + Frr + Faero

Ptot = Ftot · v(t)

Notice that the contribution of Fg has to be considered only in Request 2, where the calculation
on the urban cycle with 10% slope will be performed.
The energy to be provided to perform a single speed cycle is the area under the power curve:
Z Tcycle
Ecycle = Ptot (t)dt
0

Considering battery, motor and mechanical transmission efficiency, the energy demanded to the
battery is :

Ecycle
Ebatt,cycle =
ηbatt · ηtrans · ηmot
Now, it’s necessary to calculate the distance that the vehicle travels per each urban cycle:
Z Tcycle
s= v(t)dt
0

and the number of cycles performed knowing that the required range of the battery pack has to
be at least 100km:

100 km
ncycles =
s
Now we’re able to size the battery pack, considering the specific energy:

Ebatt,cycle · ncycles
mbatt,Pb,E =
Espec,P b

Ebatt,cycle · ncycles
mbatt,Li,E =
Espec,Li

On the other hand, if we consider the specific power, we’ve to satisfy the peak power achieved
during the urban cycle.

max(P (t))
mbatt,Pb,P =
Pspec,Pb

max(P (t))
mbatt,Li,P =
Pspec,Li

So, we choose the battery pack mass as:

mbatt = max(mbatt,E , mbatt,P )

The above mentioned procedure has to be repeated as long as the vehicle is able to guarantee the
imposed autonomy and, in any case, the maximum allowable load is not exceeded.

67
Reports on the laboratories R. Desideri, L. Scocco

Request 2
Once the battery is sized, it’s necessary to add the gravitational force contribution and calculate
the new requested energy following the previous procedure.
Fixing the battery pack chosen at the end of the previous point, we’ve to recalculate the energy
required per each urban cycle and the number of urban cycles the vehicle can travel:
Etot,batt
nslope =
Eslope

where:

Etot,batt = Ebatt,cycle · ncycle

And finally the autonomy in a sloping path is:

sslope = s · nslope
where s is the length of the urban cycle.

Request 3
Let’s now calculate the energy consumption in an extraurban path with no slope and at maximum
speed. In these conditions the inertia forces and the component of gravitational force are equal to
zero. So, the total force is only:

Ftot = Frr + Faero

that is constant in time.
So, also the power will be constant. Since the net energy available from the battery is:
Z T
Eextra = Ebatt,cycle · ncycle = P (t)dt
0
we easily obtain:
Eextra
T =
P
Once we know the value of T, the new value of autonomy for an extraurban path can be derived:

sextra = vmax · T

Request 4
The energy to be reintegrated is:

Erecharge = Ebatt,cycle · ncycle

where Erecharge has to be calculated in kWh. Finally the total recharging cost is:

Crecharge = QkW h · Erecharge

68
Reports on the laboratories R. Desideri, L. Scocco

Results

Symbol Value Units Description

Mass of the batteries, no slope
θ 0 [°] Slope of the road
d 1.91 [km] Distance travelled per cycle
ncycle 52.36 [-] Required cycles for 100 km of autonomy
mbatt, Pb, E 235.3 [kg] Lead battery mass, considering energy consumption
mbatt, Li, E 61.43 [kg] Lithium battery mass, considering energy consumption
mbatt, Pb, P 165.2 [kg] Lead battery mass, considering max power
mbatt, Li, P 46.07 [kg] Lithium battery mass, considering max power
mbatt, Pb 210 [kg] Lead battery mass
mbatt, Li 61.43 [kg] Lithium battery mass
Ecycle, Pb 131.6 [Wh] Energy required per cycle with the lead battery
Ecycle, Li 140.8 [Wh] Energy required per cycle with the lithium battery
Etot, Pb 6300 [Wh] Total energy in the lead battery
Etot, Li 7372 [Wh] Total energy in the lithium battery
ncycle, Pb 47.88 [-] Number of cycles completable with the lead battery
ncycle, Li 52.36 [-] Number of cycles completable with the lithium battery
Autonomy with slope
θ 5.711 [°] Slope of the road
Ecycle, Pb, slope 546 [Wh] Energy required per cycle with the lead battery
Ecycle, Li, slope 548.9 [Wh] Energy required per cycle with the lithium battery
ncycle, Pb, slope 11.54 [-] Number of cycles completable with the lead battery
ncycle, Li, slope 13.43 [-] Number of cycles completable with the lithium battery
dPb, slope 22.03 [km] Autonomy of the lead battery
dLi, slope 25.65 [km] Autonomy of the lithium battery
Autonomy at the maximum speed
θ 0 [°] Slope of the road
d 6 [km] Distance travelled per cycle
vmax 25 [m/s] Maximum speed
Ecycle, Pb, max 722.4 [Wh] Energy required per cycle with the lead battery
Ecycle, Li, max 850.8 [Wh] Energy required per cycle with the lithium battery
ncycle, Pb, max 8.721 [-] Number of cycles completable with the lead battery
ncycle, Li, max 8.664 [-] Number of cycles completable with the lithium battery
dPb, max 52.33 [km] Autonomy of the lead battery
dLi, max 51.99 [km] Autonomy of the lithium battery
Cost of recharge
Erecharge, Pb 6300 [Wh] Energy to recharge the lead battery
Erecharge, Li 5898 [Wh] Energy to recharge the lithium battery
Crecharge, Pb 0.945 [e] Cost to recharge the lead battery
Crecharge, Li 0.8847 [e] Cost to recharge the lithium battery

69
Reports on the laboratories R. Desideri, L. Scocco

7 Report 7
AC brushless operational range identification
A permanent magnets AC motor has the following main data:
• Vn = 28 V Concatenated Voltage
• n = 4 Pole pairs number
• Ωn = 141 rads Nominal speed
• Rs = 0.05Ω Winding equivalent resistance
• Ls = 0.00019H Winding equivalent inductance
• Ke = ψm = 0.042W b Permanent magnet magnetic flux (rational)
• Kt = nKe Torque constant
• Pf e,n = 6.8W Iron losses @nominal speed

The motor is controlled by a three-phase inverter with the folowing features:

• Vin = 48V Input voltage
• In,conv = 75ARM S RMS current provided by the converter
• fmax = 250Hz Max current frequency
• ηinv = 0.95 Inverter efficiency

The motor moves a scooter. Here the mechanical data:

• z=6 Reduction ratio
• r=0.25m Wheel radius
• ηtrans = 0.97 Transmission efficiency

A digital control is implemented and it keeps ε = π/2 (angle between the statoric current and the
permanent magnets flux). The torque constant Kt and magnetic flux Ke are given in the rational
phasor domain. Requests:
1. Draw the operational range of the motor considering Imotor = Idrive-max .
2. Draw the torque and the power of the wheel.

70
Reports on the laboratories R. Desideri, L. Scocco

Request 1
Operating range without field weakening
If the control algorithm operates with an angle between current and permanent magnet flux ε =
π/2, the space vector diagram of Figure ?? can be obtained.

Figure 39: Qualitative Space Vectors diagram.

Under this hypothesis, the torque-current relation is the following:

Tr
I=
KT

The stator voltage space vector (expressed in the “rational” form) can be calculated as:

V = E + Rs I + jωLs I

In this case, the current space vector I and the back e.m.f. space vector E assume the same
orientation. The space vector E magnitude can be expressed as follows:

E = Ke ω = Ψm ω

The voltage V magnitude can be calculated:

p
Vs = (E + Rs I)2 + (ωLs I)2

For continuous duty service the current limit value is the nominal one. The flux is constant and
assumes the value of the permanent magnets one and is normally expressed as back e.m.f. constant
Ke . √
Considering the stator current I equal to the nominal one in “rational representation” (I = 3In,inv )
and the voltage equal to the phase to phase nominal RMS Vn (that already corresponds to a rational
representation), we can rewrite the space vector E magnitude as:
q √ √
Vn = (ωLs 3In )2 + (Rs 3In + Ke ω)2

It is now possible to find the base speed ωbase corresponding to the limit of the nominal operating
range without field weakening. Making the calculation, it’s possible to find that:

rad
 
ωbase = 455.83 = 1088.2 [rpm]
s

71
Reports on the laboratories R. Desideri, L. Scocco

Field Weakening of PMS motor

Considering the Figure ?? and neglecting the value of Rs it is possible to see that, increasing the
motor electrical speed ω, the angle δ and the voltage amplitude are increasing.
We now split the current and voltage space (rational) vectors in the real and imaginary components
of the rotating reference frame:
V = Vd + jVq

I = Id + jIq

So, considering a generic  angle we obtain the following expressions for the voltage rational phasor
components:

vsd = Rs isd − ωLs is q

vsq = Rs isq + ωLs is d + E

and by making an energy balance we can demonstrate that the expression of the torque is:

T = nΨm · isq

Considering now to introduce a negative value of isd , the diagram of Figure ?? is obtained.

Figure 40: Qualitative Space Vectors diagram considering an injection of a isd < 0 component.

Observing Figure ??, it is possible to notice that a negative value of isd yields to a reduction of
the stator voltage amplitude and an increasing of the current amplitude.
If the current amplitude exceeds the current limit value, the isq component has to be reduced, thus
the torque value is reduced.
For a proper use of the motor, exploiting the maximum possible torque for every speed, the following
control strategy can be adopted:
• For speed lower than the base one ωbase , it is convenient to set isd = 0 and isq equal to the
maximum (imax ), assuring the maximum constant torque. (see Figure ??(a)). Motor voltage
increases proportionally to speed.
• For speed equal to ωbase , the motor voltage is equal to the nominal one and corresponds to
the maximum voltage available as output of the power supplier (inverter) (see Figure ??(b)).
• For speed ω > ωbase , the isq has to be reduced since a negative isd component is necessary.
The motor current is has to be compliant to the following:
q
is = i2sd + i2sq ≤ imax

In this way Vs = Vm,max and is ≤ imax , with an isq < imax always. See Figure ??(d).

72
Reports on the laboratories R. Desideri, L. Scocco

Figure 41: Qualitative Space Vectors diagrams considering different operating conditions.

73
Reports on the laboratories R. Desideri, L. Scocco

Results
In order to obtain the following graphs representing the operating range of this PMSM, we have
already taken into account the following conventions. Knowing that ω is the current angular
frequency and Ω is the shaft angular frequency, we have that:

ω(f ) = 2πf

ω(f )
Ω(f ) =
n
where n is the pole pairs number.

Rational currents

129.9038

100

50
Current [A]

I max
Id
-129.9035 Iq

0 50 72.5482 100 150 200 250

Frequency [Hz]

74
Reports on the laboratories R. Desideri, L. Scocco

Phase currents
75

57.7350

28.8675
Current [A]

I max
Id
Iq
-74.9998
0 50 72.5482 100 150 200 250
Frequency [Hz]

Voltages
27.9255
25.5314
22.0375
Voltage [V]

Vd
6.4952
Vq
Vs
0

-6.5753

-11.3129

-17.2732

0 50 72.5482 100 128 150 200 250

Frequency [Hz]

75
Reports on the laboratories R. Desideri, L. Scocco

Request 2
The calculation has to be performed as a function of the frequency f , that starts from a minimum
of 1 Hz and reaches the maximum inverter output frequency.
ηinv
Vmax (f ) = Vin √ Maximum inverter output voltage
2
√
Iqmax = 3Imax Iq max value in rational representation
For each step as a function of f , the calculation starts with Iq = Iqmax .
ω(f )
1. ω(f ) = 2πf ; Ω(f ) =
n
2. Vq (f ) = Ke ω(f ) + Rs Iq (f )
3. Vd (f ) = p
−Ls ω(f )Iq (f )
4. Vs (f ) = Vd (f )2 + Vq (f )2
Ω(f ) · r · 3.6
5. speed(f ) = [km/h]
z
If Vs (f ) > Vmax then an Id < 0 has to be injected repeating the calculation until
Vs (f ) ' Vmax . The steps for this calculation are:
q √
6. Iq (f ) = (Imax 3)2 − Id (f )2
7. Vq (f ) = Ke ω(f ) + Rs Iq (f ) + ω(f )Ls Id (f )
8. Vd (f ) = p
−ω(f )Ls Iq (f ) + Rs Id (f )
9. Vs (f ) = Vd (f )2 + Vq (f )2
Step from 6 to 9 are repeated until Vs (f ) ' Vmax decreasing negative Id .
After condition Vs (f ) ≤ Vmax is verified for every frequency the following points are calculated,
in order to find the electric power and current absorbed by the inverter, the torque, power and
efficiency curves both for the shaft and for the wheel.

P (f ) = Vd (f )Id (f ) + Vq (f )Iq (f ) Motor input power

2
Pcu (f ) = Rs (Id (f ) + Iq (f ) )2
Copper losses
ω(f )
Pf e (f ) = Po Iron losses
nΩn
Pshaf t (f ) = (P (f ) − Pcu (f ) − Pf e (f )) Shaft mechanical power
Pshaf t (f )
Tshaf t (f ) = Shaft torque
Ω(f )
Pwheel (f ) = Pshaf t (f )ηtransm Wheel power
P (f )
Pinv (f ) = Inverter input power
ηinv
Pinv (f )
Iinv (f ) = Inverter input current
Vinv
Pwheel (f )
ηtot (f ) = Overall efficiency
Pinv (f )
Pwheel (f )z
Twheel (f ) = Wheel torque
Ω(f )
Id (f )
Idf (f ) = p Phase Id current
(3)
Iq (f )
Iqf (f ) = p Phase Iq current
(3)
q
Isf (f ) = Idf (f )2 + Iqf (f )2 Phase current (total)

76
Reports on the laboratories R. Desideri, L. Scocco

Results

Shaft efficiency
0.7276
0.6
[-]

0.4
0.2
0
0 50 100 150 200 250
Shaft power
2785.8894

2000
[W]

1000

0
0 50 100 150 200 250
Shaft torque
21.7756
[Nm]

10

0
0 50 72.5482 100 150 250
Frequency [Hz]

Wheel efficiency
0.7058
0.6
[-]

0.4
0.2
0
0 50 100 150 200 250
Wheel power
2702.3127
2000
[W]

1000

0
0 50 100 150 200 250
Wheel torque
126.7341
100
[Nm]

50

0
0 50 72.5482 100 150 250
Frequency [Hz]

77
Reports on the laboratories R. Desideri, L. Scocco

Wheel efficiency
0.7058
0.6
[-]

0.4
0.2
0
10 17.0938 23.5619 30 40 50 58.9049
Wheel power
2702.3127
2000
[W]

1000

0
10 17.0938 23.5619 30 40 50 58.9049
Wheel torque
126.7341
100
[Nm]

50

0
10 17.0938 23.5619 30 40 50 58.9049
Speed [km/h]

Figure 42: Wheel efficiency, power and torque as a function of the scooter speed.

78
Reports on the laboratories R. Desideri, L. Scocco

Powers and power losses

P
3637.2906 P shaft
P wheel
P cu

2785.8894 P fe

2500
Power [W]

2000

1500

843.75

18.9387
0 72.5482 101 250
Frequency [Hz]

Current absorbed by the inverter

80

70

60

50
I inv [A]

40

30

20

10
0 50 100 150 200 250
Frequency [Hz]

79
Reports on the laboratories R. Desideri, L. Scocco

8 Report 8
Given the following data:
Locomotor:
• Full load mass: mv = 40.6 t without motors
• Wheels diameter: D = 0.92 m

Performance reference - Plane road:

• p = 0.5‰
• Maximum speed: vm = 100 km/h
• Starting acceleration: a0 = 0.9 m/s2
• Residual acceleration: ar = 0.05 m/s2

Performance reference - Ascent:

• p = 15‰
• Maximum speed: vm = 90 km/h
2 2
• Starting acceleration: a0 = 0.9 m/s − g sin(tan−1 (p)) = 0.7529 m/s
• Residual acceleration: ar = 0.05 m/s2

The overall resistance force acting on the vehicle due to the aerodynamic drag and the kinematic
pairs (wheel-rail contact and pivot-bearing bushings), can be calculated as:

Fr = (A + B · v) · M + Cv 2

where M is the mass (in tons) of the vehicle, v is the speed [km/h], and the coefficients are:
• A=0.00084895
• B=0.0001
• C=00002674

Figure 43: Propulsion Scheme

80
Reports on the laboratories R. Desideri, L. Scocco

Engine and drive specifications:

The propulsion system consists of two three-phase asynchronous motors operated independently
each from a IGBT inverter (Figure ??).
• Nominal output power of the energy source: Pn = 210 kW
• Voltage at the DC bus line (input for the inverter): VDC = 700 V
• Maximum number of engine revolutions: nmax = 4100 rpm
• Number of pole pairs: np = 2

Calculate:

• The engine power and the traction characteristics in relation to the performance required.
• The characteristics of the electric motor traction in relation to the performance required.
• The size and the weight of the motor.

Request 1
Performance Analysis
In order to make a first estimation of the power of the motors, let’s draw first all the resistant
power curves as a function of speed, for the two different services reported in data (in slope and
flat ground).
The resistant forces on the vehicle are:
• Aerodynamic forces
• Forces related to kinematic torques
• Component of the weight on the inclined plane
• Inertia force related to the residual acceleration

Aerodynamic forces and resistant forces related to kinematic torques are calculated through the
following empiric relation, as a function of mass M (in tons) of vehicle and speed v (in km/h):

Fr = (A + B · v) · M + Cv 2

While for the mass of the vehicle we need to add to the value provided by data a contribution that
takes into account the equivalent masses related to the propulsion system. Thus, let’s assume:

M = 43 [ton]

The weight component and the inertia force related to the residual acceleration, that is the accel-
eration that the vehicle has at the maximum speed, are:

Fp = M · g · p

Far = M · ar

Where p is the slope (p/1000). The resistant power, related to one drive, can be calculated with
the equation:

Fr + Fp + Far
Pr = ·v
2
Computing the calculation for both the situation and plotting the obtained values as a function of
speed, the following graph can be obtained:

81
Reports on the laboratories R. Desideri, L. Scocco

Resistive power for one drive

140
No slope
Ascent
120

100

80
P r [kW]

60

40

20

0
0 10 20 30 40 50 60 70 80 90 100
v [km/h]

Figure 44: Resistant power vs Speed

From the graph it can be noticed that the hardest situation is the uphill situation, that requires
for the single motor a power of at least 140 kW (138.3 kW for accuracy); thus, consider a minimum
power for the sizing of at least 145 kW. The nominal power provided from the primary source is
210 kW.
Thus, assuming for the motor and the inverter the following efficiency values, that are typical
values for this kind of devices used in traction applications:
• ηmotor = 0.93
• ηinverter = 0.99
The maximum value for the traction power assumed to be available for each motor is:

Pmax = Pn · ηmotor · ηinverter

The rated power of the motor is set to 190 kW, in order to have the best use of the power available
on board, also achieving a good margin on the power required by the (Figure ??). This margin
will, in particular, ensure the residual acceleration value at the maximum speed required in the
data, i.e. the characteristic curve of the vehicle (power as a function of speed) will always remain
above the corresponding curve of the load.

Traction curve of the vehicle Assuming the transmission efficiency equal to 1, from the nom-
inal power previously set and from the required maximum speed the traction force for each drive
at maximum speed can be calculated (the part of the graph at constant power):

Pm
Fv,max =
vmax

82
Reports on the laboratories R. Desideri, L. Scocco

According to the starting acceleration, the starting force can be calculated. It is kept constant
until the speed at which the nominal power is reached. To compute the calculation we need to
take into account the overall mass:

2
4 Dwheel
Mtot = mv + 2 · mmotor + 2 · [2(Jmotor + Jtransm ) · K 2 + mwheelset · ]
Dwheel 4

where:
• mass of the wheelset: mwheelset = 0.08 · mv
• First attempt value of the motor inertia: Jmotor = 4 kgm2
• First attempt of the motor mass: Mmotor = 600 kg

Once the real dimension of the motor has been calculated (at the end of Request 3), the procedure
has to be iterated up to convergence.

Maximum angular speed of the motor (electric rad/s):

N · 2π
ωm,max = n ·
60
Gear Ratio:
ωm,max · D
K=
n · vmax · 2
Moment of inertia of transmission:
Jtransm = 2 [kgm2 ]

For the computation of the resistant force contribution due to the weight on the inclined plane, it
should be noted that the "real" mass of the vehicle provided in the data did not take into account
the mass of the motors, thus:

Mv = mv + 2 · mmotor

The traction starting force of each drive is:

Mtot · a0 + Mv · g · p
F0 =
2
where:

• a0 = 0.9 sm2 in case of plane road;

• a0 = 0.9 sm2 − g · sin[tan−1 (p)] in case of ascent.

From this force the nominal speed of the vehicle can be derived, that is the speed at which the
nominal power is reached:

Pm
vn =
F0
Once these values are known, the traction characteristic for the railway car can be plotted.

Finally, the elasticity of the vehicle, that is the ratio between the maximum speed and the minimum
speed at constant power, can be calculated:

vmax
ev =
vnom

83
Reports on the laboratories R. Desideri, L. Scocco

Request 2
Traction curve of the electric motor
Let’s now consider the sizing of the motor, analyzing the traction characteristic achievable with
our requirements. The motor traction characteristic will have the same qualitative profile of the
traction characteristic of the railway.
From the input voltage of the inverter the nominal phase voltage of the motor considering the
six-step modulation, can be calculated:
√ VDC
Vnf = 6·
π
And then the nominal current (rational) can be estimated:

Pm
In = √
3 · Vnf · cos(φ) · ηmotor
Considering the traction characteristic of the motor, the parameters can be calculated as:
ωmax
Ωm,max =
n
Pn
Tm,max = Tm @Ωm,max =
Ωm,max
In the previous equation the maximum angular speed of the motor, calculated before, is indicated.
From it the maximum working frequency can be calculated and then, evaluated the frequency
operating range for the converter:
ωm,max
fmax =
2π
In order to achieve the best exploitation of the performance of the motor it is now required that the
torque value delivered at the maximum angular speed is equal to the maximum torque delivered
by the motor, that from the 4 parameters equivalent circuit has the expression:
2
3Vnf
Tm,max = Tm @Ωm,max ∼
=
2ΩXk
Thus:
2
3Vnf
Tm,max = Tm @Ωm,max ∼
=n 2
2ωm,max Lk
In the previous expression, all the parameters are known except for Lk, that represents the total
dispersion inductance. This data is physically determined by some sizing parameters of the active
parts of the motor, and in particular it is directly connected to the air gap value and the design of
the slots.
Thus, the so calculated value represents a very important indication for the manufacturer, in
particular in order to be able to verify the equivalence of the two torque terms previously mentioned.
Reversing the expression (20) the following value can be obtained:

3Vnf
Lk = n 2
2ωm,max Tm @Ωm,max
Sizing the motor according to the above mentioned instructions, it can also be verified that the
motor elasticity is the same as that of the railway car, that is:

vmax Ωm,max
ev = = em =
vn Ωm,n
For the calculation of parameters referred to nominal operation, i.e. torque and angular speed at
nominal power (at the end of the constant power section) the following equations can be used:

84
Reports on the laboratories R. Desideri, L. Scocco

Ωm,max
Ωm,n =
em

Tm,n = Tm,max · em
It is also possible to obtain the rated operating frequency from the nominal speed of the motor:
Ωm,n
fn = n
2π
Once all the parameters are known, the traction characteristic of the motor can be plotted, as was
for the traction characteristic of the railway car.

Request 3
Calculation of size and weight of the motor
Knowing the nominal torque value and using the relationship between the torque and the motor
nameplate, the rotor diameter can be obtained:
πfa
Tm,n = √ · B · An · ff · Dr3 · ηmotor · cos(φn )
2 2
For the parameters in this equation, as for other machines with similar architecture, the following
values can be considered:
• Maximum magnetic field in the air gap: B = 0.85T
• Nominal current density: An = 3I nU
πDr = 47000 m
A

where:
– Number of coils in series per each phase: U
– Rotor diameter: Dr
• Power factor: cos(φn ) = 0.84
• Winding factor: fa = 0.96
• Shape factor of the rotor. ff = Dlr = 1

Once the nominal torque is known, from the previous expression it is possible to calculate the rotor
diameter:
s √
3 2 2Tm,n
Dr =
π · fa · ff · B · An · ηmotor · cos(φn )
And from it, using an adequate coefficient, the external diameter of the motor:
Dr
Dext =
0.7
From the imposing shape factor, the length of the rotor can be derived:

1 = ff · Dr

Whereas for the external length of the motor, assuming that it is self-cooling trough a cooling fan
directly mounted on the rotor, it is:

lext = 1.1 · Dext

Finally, the mass of the motor and the moment of inertia of the rotor, onche that the volume have
been calculated, can be estimated assuming an appropriate average density for the mass of the
motor:
2 2
π · Dext ·l π · Dext ·l
mmotor = δ = 1.6 · 8000
4 4

85
Reports on the laboratories R. Desideri, L. Scocco

Dr2
Jrotor = mmotor
8

Results
The final results of the calculations are:

Symbol Value Units Description

No slope Ascent
pslope 0.005 0.015 [-] Slope
vmax 100 90 [m/s] Maximum speed
a0 0.9 0.7529 [m/s2 ] Starting acceleration
ar 0.05 0.05 [m/s2 ] Residual acceleration
Power required to the motor
Pn 210 210 [kW] Nominal power primary source
N 4100 4100 [rpm] Maximum number of engine revolutions
Ωmax 429.4 429.4 [rad/s] Maximum mechanical speed
ωmax 858.7 858.7 [rad/s] Maximum electrical speed
fmax 136.7 136.7 [Hz] Maximum electrical frequency
M 43 43 [t] Mass for the preliminary design
Fr 3.141 2.589 [kN] Resistant forces at the max speed
Fp 2.109 6.327 [kN] Resistant force due to the weight
Far 2.15 2.15 [kN] Inertia force
Ftot 7.4 11.07 [kN] Total resistant force
Pr 102.8 138.3 [kW] Resistant power at the max speed
Pm 190 190 [kW] Mechanical Power from the electric motor
F (vmax ) 6.84 7.6 [kN] Force of the motor at maximum speed
K 7.11 7.9 [-] Transmission ratio
Characteristics of the motor
Mmotor 1034 843 [kg] Resultant motor mass
Jmotor 17.65 12.55 [kg·m2 ] Resultant motor inertia
Jtransm 2 2 [kg·m2 ] Transmission inertia
Mv 42.67 42.29 [t] Total mass of the vehicle
Mtot 55.3 54.11 [t] Equivalent mass
F0 25.93 23.48 [kN] Starting force
vn 7.326 8.091 [m/s] Nominal speed
ev 3.791 3.09 [-] Elasticity factor
F (vn ) 25.93 23.48 [kN] Force at nominal speed
Operating range
Ωn 113.2 139 [rad/s] Nominal mechanical speed
ωn 52.09 63.92 [rad/s] Nominal electrical speed
fn 36.05 44.23 [Hz] Nominal electrical frequency
Tm, n 1.678 1.367 [kNm] Nominal torque
Tm, max 0.4425 0.4425 [kNm] Maximum torque
A 243.2 243.2 [kW] Apparent power
Vdc 700 700 [V] Input voltage for the inverter
Vph 545.8 545.8 [V] Phase voltage
Iph 138.1 138.1 [A] Phase current
In 257.3 257.3 [A] Nominal current (rational)
Lk 2.739 2.739 [mH] Total dispersion inductance
Size of the motor
Dr 0.3694 0.3451 [m] Rotor diameter
Dext 0.5278 0.493 [m] External diameter
l 0.3694 0.3451 [m] Rotor length
lext 0.5805 0.5422 [m] External length

86
Reports on the laboratories R. Desideri, L. Scocco

Operating range: force, no slope

30

25

20
Force [kN]

15

10

0
0 10 20 30 40 50 60 70 80 90 100
Speed [km/h]

Operating range: force, ascent

25

20

15
Force [kN]

10

0
0 10 20 30 40 50 60 70 80 90
Speed [km/h]

87
Reports on the laboratories R. Desideri, L. Scocco

Operating range: torque, no slope

1.8

1.6

1.4
Torque [kNm]

1.2

0.8

0.6

0.4
0 50 100 150 200 250 300 350 400
Speed [rad/s]

Operating range: torque, ascent

1.4

1.3

1.2

1.1
Torque [kNm]

0.9

0.8

0.7

0.6

0.5

0.4
0 50 100 150 200 250 300 350 400
Speed [rad/s]
While, the results in to the first iteration, using the guess values for the mass and the inertia of
the motor, are:

88
Reports on the laboratories R. Desideri, L. Scocco

First iteration - Guess mass and inertia

Symbol Value Units Description
No slope Ascent
Characteristics of the motor
Mmotor, guess 600 600 [kg] Resultant motor mass
Jmotor, guess 4 4 [kg·m2 ] Resultant motor inertia
Mv 41.8 41.8 [t] Total mass of the vehicle
Mtot 47.91 48.59 [t] Equivalent mass
F0 22.59 21.37 [kN] Starting force
vn 8.412 8.893 [m/s] Nominal speed
ev 3.302 2.811 [-] Elasticity factor
F (vn ) 22.59 21.37 [kN] Force at nominal speed
Operating range
Ωn 130 152.7 [rad/s] Nominal mechanical speed
ωn 59.81 70.25 [rad/s] Nominal electrical speed
fn 41.39 48.61 [Hz] Nominal electrical frequency
Tm, n 1.461 1.244 [kNm] Nominal torque
Tm, max 0.4425 0.4425 [kNm] Maximum torque
Size of the motor
Dr 0.3528 0.3344 [m] Rotor diameter
Dext 0.504 0.4777 [m] External diameter
l 0.3528 0.3344 [m] Rotor length
lext 0.5544 0.5254 [m] External length

89
Reports on the laboratories R. Desideri, L. Scocco

9 Report 9
Given the following asynchronous motor:
• Concatenated voltage: Vn = 220 V
• Pole pairs number: n = 2
• Nominal torque: Tn = 14 Nm
• Nominal frequency: fn = 50 Hz
• Nominal Mechanical speed: Ωn = 2πfn /n
• Rs = 0.68 Ω
• Iron losses in nominal condition: Pf e,n = 100 W

The 4-parameters-circuit with statoric inductance:

• Lks = 7.8 mH
• Rr∗∗ = 0.6065 Ω
• M = 0.6042 H
• Magnetizing current corresponding to the nominal flux: Im = 5.38 A

The 4-parameters-circuit with rotoric inductance:

• Rr∗ = 0.7624 Ω
• Lkr = 8.7 mH
• Ls = 72 mH

The drive used is a three-phase inverter with fmax = 250 Hz.

The control acts by imposing a constant flux from 2 Hz till the nominal frequency (first area) and,
for higher frequencies, it keeps a constant voltage equal to Vn (second area).
You are asked to:
• Draw the operational range of the given motor, under the hypothesis that the output torque
equals to the nominal for the first area, and that in the second area the power is equal to the
nominal one. Plot the diagrams of: voltage module Vs, current module absorbed Is, delivered
torque, factor of power and performance.
• Draw the diagrams of the maximum torque and current absorbed in the maximum torque
condition.

90
Reports on the laboratories R. Desideri, L. Scocco

Request 1
Definition of current, efficiency, voltage, power factor values for an asynchronous
motor supplied by an inverter in nominal conditions
The operating range of an asynchronous motor supplied by a converter can be divided into two
parts: a first one at constant flux and a second one at constant voltage (the maximum provided by
the converter). The two regions are analyzed using two different electric circuits. In this regard,
the unknowns will be defined independently.

Constant flux region (2 < f < fn ) First of all, let’s take into consideration the constant flux
region. The equivalent electric circuit is the 4 parameters circuit shown in the following figure:

Figure 45: 4-parameters equivalent circuit with Lks stator inductance

Now, from the known parameters, let’s determine the current I s and the voltage V s needed to
reach the imposed nominal torque Tn . First of all, determine the voltage upon the magnetizing
inductance through the current I m .

E = jωM · I m = j2πf M I m

Once the voltage E is known, it is possible to find the rotor current from the mechanical power
definition:

Pm = Tn Ω = 3E · I s = 3E · Ir

Tn Ω Tn 2πf /n
Ir = =
3E 3E

Notice that the factor "3" is due to the fact that both Ir and E are non-rational phasors.
Knowing the rotor current and the magnetizing current, it’s possible to find the input statoric
current I s :

Is = Im − Ir

Now, the following equation that allows to calculate the required voltage, can be derived:

V s = Rs I s + jωLks I s + E

This equation can be expressed by the following phasor diagram:

91
Reports on the laboratories R. Desideri, L. Scocco

Figure 46: Phasor Diagram

Constant voltage region (fn < f < fmax ) Now the supply voltage is constant and equal to
the maximum one:
220
V s,max = V s (fn ) = j √ V
3
From this voltage, the magnetizing current can be derived:

Vs
I s,0 = I m =
jωLs
For the other values of current, first we have to determine the ratio y = Rr
x computing a quadratic
equation, placing the torque T equal to:
3y
T = · I ∗2 (4)
Ωn r
where:

Vs2
Ir∗2 = (5)
(RS + y)2 + ω 2 L2kR
Replacing ?? into ?? and writing the equation as a function of y, we have ??:

3Vs2
 
y 2 + 2 · RS − y + (ω 2 L2kR + RS2 ) = 0 (6)
Tn Ωn
This is a quadratic expression in the form:

Ay 2 + By + C = 0
where:
• A=1

92
Reports on the laboratories R. Desideri, L. Scocco

• B = −20.6233
• C is a function of the electrical frequency.

Equation ?? has two solutions that are real since the demanded torque is lower than the
maximum deliverable at that frequency and they correspond to the stable or the unstable
solution on the mechanical characteristic. Of course, we have to choose the stable one, that
is the solution with the maximum positive value expressed by the lowest value of slip (near to zero).

Figure 47: 4-parameters equivalent circuit with Lkr rotor inductance.

Once we know the value of y, solving the circuit in Figure ?? it is possible to determine the rotor
current I R and then the stator current I S .

Vs
Ir =
Zt

Z t = RS + y + jωLkR

I s = I s,0 + I r

Figure 48: Phasor diagram.

93
Reports on the laboratories R. Desideri, L. Scocco

The phasor diagram that describes currents and voltages in the second region of the operating
range is shown in Figure ??.

Considering the second region of the operating range, we need to verify that the computed torque
is lower than the maximum one. To do this validation, we need to calculate the slip at a given
frequency:

Rr
xi =
yi
and compare it with the maximum slip expressed in the following expression:

Rr
xmax =
ωLkR
If the calculated slip is higher than the maximum one, then the torque is equal to the maximum
one that has a profile inversely proportional to ω 2 :

3Vs2 · n
Tmax =
2ω 2 LkR

Power and Efficiency For both regions of the operating range voltage and current provided to
the motor have been calculated. Now we want to calculate losses and the efficiency of the machine.
The power absorbed by the motor can be expressed by the following expression.

Pabs = 3Re(V s Is )
But now we’ve to take into account of iron losses and resistive losses. The iron losses can be
calculated as described in the following equations:
 2
ω Ψ
Pf e = Pf e,n
ωn Ψn
E
Ψ=
ω
 2
ωn E
Pf e = Pf e,n
ω En
Thus, in the first region, since the flux is constant, we have that:
ω
Pf e = Pf e,n
ωn
whereas in the second region, since the voltage is constant:
ωn
Pf e = Pf e,n
ω
Thus, the output power can be expressed as in the following equation:

Pout = Pabs − 3RS Is2 − 3RR Ir2 − Pf e

Finally, the efficiency of the machine can be calculated with the ratio between output and absorbed
power:

Pout
η=
Pabs
The power factor is the cosine of the angle between the voltage V s and current I s phasors:
  
|V s |
cos ϕ = cos tan−1
|I s |

94
Reports on the laboratories R. Desideri, L. Scocco

Maximum torque profile and current value at maximum torque

Once the characteristic of the motor in nominal conditions is known, it is possible to determine
the operating range in maximum conditions.
First of all, the second region of the operating range (f > fn ), where the maximum torque can be
calculated as in the following:

3Vs2 · n
T =
2ωi2 LkR

the slip at maximum torque as:

Rr
xmax =
ωLkR

and then solving the equivalent circuit in Figure ?? computing the rotor Ir and stator Is current:

Rr
Z t = RS + + jωLkR
xmax

V s,max
I r,max =
Zt

where:

220
V s,max = j √ V
3

Considering the first region (2 < f < fn ), assuming a constant flux Ψ, the torque is constant and
equal to:

3Vs2 · n
Tmax =
2 · (2πfn )2 LkR

Finally, the rotor and stator current (constant too) can be obtained, still following the 4-parameters
equivalent circuit with the stator inductance:

Tmax
I r,max = −j
3nM Im

So, the statoric current phasor is defined as

I s,max = I m,0 − I r,max

while for f > fn , following the 4-parameters equivalent circuit with rotor inductance, the statoric
current phasor is defined as:

I s,max = I s,0 + I r,max

where:

Vs
I s,0 =
jωLs

95
Reports on the laboratories R. Desideri, L. Scocco

Results

Symbol Value Units

fmin fn fn + df fymax fTmax fmax
Vs 9.925 126.9 127 127 127 127 [V]
Ir 6.756 6.756 6.016 8.435 7.913 6.411 [A]
Ir, max 27.19 27.19 29.04 8.435 7.913 6.411 [A]
Is 8.636 8.636 8.744 9.516 8.927 7.231 [A]
Is, max 27.72 27.72 32.85 9.516 8.927 7.231 [A]
Is0 5.38 5.38 5.614 1.49 1.395 1.123 [A]
cos(ϕ) 0.6564 0.06787 0.06868 0.07471 0.07011 0.05684 [-]
T 14 14 14 3.714 3.477 2.255 [Nm]
Tmax 56.34 56.34 56.34 3.967 3.477 2.255 [Nm]
y 20.26 10.3 11 13.67 [-]
x 0.03764 0.074 0.06928 0.05579 [-]
xmax 0.2789 0.074 0.06928 0.05579 [-]
Pin 240.1 2351 2273 2344 2195 1769 [W]
Pcu 152.1 152.1 156 184.7 162.6 106.7 [W]
Pfe 4 100 99.98 26.53 24.84 20 [W]
Pout 83.96 2099 2017 2133 2008 1642 [W]
η 0.3497 0.8928 0.8874 0.9099 0.9146 0.9284 [-]

Supplying voltage module

140

120

100

80
V s [V]

60

40

20

0
0 50 100 150 200 250
Frequency [Hz]

96
Reports on the laboratories R. Desideri, L. Scocco

Current absorbed in module

10

9.5

9
I s [A]

8.5

7.5

7
0 50 100 150 200 250
Frequency [Hz]

Delivered torque
14

12

10
T [Nm]

2
0 50 100 150 200 250
Frequency [Hz]

97
Reports on the laboratories R. Desideri, L. Scocco

Power factor
0.7

0.6

0.5

0.4
cos( )

0.3

0.2

0.1

0
0 50 100 150 200 250
Frequency [Hz]

Input and output power

2000
Power [W]

1500

1000
P in
500
P out

0
0 50 100 150 200 250
Frequency [Hz]
Power losses
200
P cu
150 P fe
Power [W]

100

50

0
0 50 100 150 200 250
Frequency [Hz]

98
Reports on the laboratories R. Desideri, L. Scocco

Efficiency
1

0.9

0.8

0.7

0.6

0.5

0.4

0.3
0 50 100 150 200 250
Frequency [Hz]

Slip
0.3
x
x max
0.25

0.2

0.15

0.1

0.05

0
50 100 150 200 250

99
Reports on the laboratories R. Desideri, L. Scocco

Maximum statoric current

35
32.8459

30
27.7156

25
I s, max [A]

20

15

10

7.2312
5
0 50 100 150 200 250
Frequency [Hz]

Maximum torque
60
T
T max
50

40
T [Nm]

30

20

10

0
0 50 100 150 200 250
Frequency [Hz]

100