Lesson 1

Exercise 1 Air gap in magnetic circuit

An iron core, seen in figure 1, has the following parameters:
μ0 = 1.25664 ∙ 10−6 mkgc −2 A−2
μr, air = 1, μr,iron = 5,000

lab = lef = 0.07 m

laf = lbe = 0.03 m
lair = lcd = 0.006 m
w = lhj = 0.01 m

d = ljk = 0.01 m

Nt = 250 turns

The steel (iron) saturates at B = 2 T and we assume that the cross-sectional area of the airgap is
equal to the area of the steel cross-section, Aair = Airon = lhj × ljk

a. Find the necessary current in order to produce a magnetic flux Φ = 3.1×𝟏𝟎−𝟒 Wb in the
airgap.
We use Ampere’s law:
n

NI = ∑ Hi li
i=1

As long as we have the same cross-section along the magnetic path and the same material, we
can keep that as one term, which leads us to:
NI = Hir ld→c + Hair lair
Φ
Since we only have soft magnetic materials, we know that B = μ0 μR H and B =
A

Bir Bair Φir Φair

NI = ld→c + lair = ld→c + l
μ0 μR,ir μ0 μ0 μR,ir Air μ0 Aair air
ld→c lair
(Here we can see the impact of reluctance, since μ = R ir and μ = R air)
0 R,ir Air
μ 0 Aair

We also made the assumption that Air = Aair = lhj × ljk = 10−4 m2 and since the flux is the
same throughout a magnetic circuit (if we have everything in series) then Φir = Φair = 3.1 ∗
10−4 Wb
Φir Φair Φair Φair Φair ld→c
μ0 μR,ir Air ld→c + μ0 Aair lair μ0 μR,ir Aair ld→c + μ0 Aair lair μ0 Aair (μR,ir + lair )
I= = =
Nt Nt Nt
 3.1 ∗ 10−4 2 ∗ 0.07 + 2 ∗ 0.03 − 0.006
( + 0.006)
μ0 0.012 5,000
I=
250
3.1 ∗ 10−4
−6 −4 (0.0000388 + 0.006)
I = 1.25664 ∙ 10 ∗ 10 = 59.59 A
250

b. What would the flux density be in the air-gap?

The flux density is:
Φ 3.1 ∗ 10−4
B= = = 3.1 T
A 10−4
c. Is this flux density possible to achieve? What would you need to change, without changing
the physical appearance of the core, in order to keep the flux density below saturation?
Flux density should only be 2 T. Given that we cannot change the area, we need to reduce the
flux to Φ = BA = 2 ∗ 10−4 Wb. The easiest way to achieve this is by decreasing the current. We
have that:
Φair ld→c 2 ∗ 10−4
(
μ0 Aair μR,ir + l air ) −6 −4 (0.0000388 + 0.006)
I= = 1.25664 ∙ 10 ∗ 10 = 38.4 A
Nt 250
d. Compare the MMF drop for the different parts in the circuit. Discuss it in relation with the
relative permeability of different materials
We know that mmf = Φ𝐑 = NI = Hl, and since the flux is the same in both iron and air, it is
0.0000388 0.006
enough to compare the reluctance of the two parts, R ir = −10 and R air = −10 .
1.25664∙10 1.25664∙10
We see that even if the travel length is much longer in the iron, since iron has such a high
relative permeability, most of the total reluctance (99.4 %) comes from the air-gap, which
shows how important it is to be able to decrease the air-gap.

e. How does length 𝐥𝐣𝐤 affect the flux and the flux density?

Depends on how you define the question. If the flux stays the same then the flux density will be
affected by the change in area. If the current stays the same and ljk change, then the flux will be
higher but flux density will be the same.
Exercise 2 – Series parallel magnetic circuit
For the series-parallel magnetic circuit in Figure 2, find the length of the air-gap lair , so that the
flux in the air-gap is Φair = 2.8 ∙ 10−4

μ0 = 1.25664 ∙ 10−6 mkgc −2 A−2

μr, air = 1, μr,iron = 5,000

lab = lbg = lgh = lha = lcf = 0.18 m

lbc = lfg = 0.12 m

lair = lde = ? m
Nt = 100 turns
I = 100 A

The thickness in all parts of the magnetic circuit is the same, bg = 4.2 ∗ 10−4 m2, except for the
middle part, where the thickness is 2.0 ∗ 10−4 m2
a) Find the length of lair . Assume that the iron does not saturate.
First we need to think about how the flux split between the two different legs. Just as with
Kirchhoff’s current law, we have that:
n

ΦT = ∑ Φi
i=1

Which gives us that ΦT = Φ1 + Φ2 . Kirchhoff’s voltage law gives us that mmfΦ1 = mmfΦ2 which
in turn gives us that:
NI = Hir,T lgb + Hir,2 lbg = Hir,T lgb + Hir,1 (lbcfg − lair ) + Hair lair

Similarly, as in exercise 1, we can rewrite this in terms of first flux density and then flux.
 Bir,T Bir,2 Bir,T Bir,1 Bair
NI = lgb + lbg = lgb + (lbcfg − lair ) + l
μ0 μR,ir μ0 μR,ir μ0 μR,ir μ0 μR,ir μ0 air
ΦT Φ2 ΦT Φ1 Φair
NI = lgb + lbg = lgb + (lbcfg − lair ) + l
μ0 μR,ir AΦT μ0 μR,ir AΦ2 μ0 μR,ir AΦT μ0 μR,ir AΦT μ0 Aair air

We will start with:

ΦT Φ2 Φ1 + Φ2 Φ2
NI = lgb + lbg = lgb + l
μ0 μR,ir AΦT μ0 μR,ir AΦ2 μ0 μR,ir AΦT μ0 μR,ir AΦ2 bg
Φ1 Φ2 Φ2 Φ1 Φ2 lgb lbg
= lgb + lgb + lbg = lgb + ( + )
μ0 μR,ir AΦT μ0 μR,ir AΦT μ0 μR,ir AΦ2 μ0 μR,ir AΦT μ0 μR,ir AΦT AΦ2

We want the flux Φ1 = 2.8 ∗ 10−4 Wb, so now we only have to solve for Φ2 .
2.8 ∗ 10−4 Φ2 3 ∗ 0.18 0.18
100 ∗ 100 = −4
(3 ∗ 0.18) + ( −4
+ )=
μ0 ∗ 5,000 ∗ 4.2 ∗ 10 μ0 ∗ 5,000 4.2 ∗ 10 2 ∗ 10−4
7.2 ∗ 10−5 0.4371
104 = + Φ2
μ0 μ0
104 μ0 − 7.2 ∗ 10−5
∴ = Φ2 = 2.86 ∗ 10−2
0.4371
We can now insert this flux into the right-hand side of the earlier equation to calculate the
length of the airgap.
ΦT Φ1 Φair
NI = lgb + (lbcfg − lair ) + l
μ0 μR,ir AΦT μ0 μR,ir AΦT μ0 Aair air
Φ1 + Φ2 Φ1 Φ1 Φ1
= l + l − l + l
μ0 μR,ir AΦT gb μ0 μR,ir AΦT bcfg μ0 μR,ir AΦT air μ0 AΦT air

Φ1 + Φ2 Φ1 Φ1 −1
= lgb + lbcfg + lair ( ( + 1))
μ0 μR,ir AΦT μ0 μR,ir AΦT μ0 AΦT μR,ir
Φ +Φ Φ
NI − μ μ1 A 2 lgb − μ μ 1A lbcfg
0 R,ir ΦT 0 R,ir ΦT
lair =
Φ1 −1
μ0 AΦT (μR,ir + 1)

2.8 ∗ 10−4 + 2.86 ∗ 10−2 2.8 ∗ 10−4

104 − (3 ∗ 0.18) − (2 ∗ 0.12 + 0.18)
μ0 ∗ 5,000 ∗ 4.2 ∗ 10−4 μ0 ∗ 5,000 ∗ 4.2 ∗ 10−4
=
2.8 ∗ 10−4 −1
( + 1)
μ0 ∗ 4.2 ∗ 10−4 5,000
2.8 + 2.86 ∗ 102 2.8
4.2 ∗ 104 ∗ μ0 − (3 ∗ 0.18) − (2 ∗ 0.12 + 0.18)
5,000 5,000
=
−1
2.8 ( + 1)
5,000
0.052777 − 0.0311904 − 0.0002576
= = 0.007619 m = 7.6 mm
2.79944
So, the length of the airgap needs to be 7.6 mm in order to establish a flux of 2.8 ∗ 10−4 Wb.

b) Now take saturation into consideration. Given the same length of the airgap as in a), what
would the flux and the flux density be in the airgap?
If we have to take saturation into consideration, then we know that the flux density inside the
second leg must be B2 = 2 T. We know this since the area of leg 2 is smaller than the rest of the
Φ2
iron core. Since B2 = AΦ2
we get Φ2 = B2 AΦ2 = 2 ∗ 2 ∗ 10−4 = 4 ∗ 10−4 Wb. We put this into:

Φ1 + Φ2 Φ1 Φair
NI = lgb + (lbcfg − lair ) + l
μ0 μR,ir AΦT μ0 μR,ir AΦT μ0 Aair air
Φ2 Φ1 Φ1
NIμ0 AΦT = lgb + lgb + (l − l ) + Φ1 lair
μR,ir μR,ir μR,ir bcfg air

Φ2 lgb + (lbcfg − lair )

NIμ0 AΦT = lgb + Φ1 ( + lair )
μR,ir μR,ir
Φ 4 ∗ 10−4
NIμ0 AΦT − μ 2 lgb 104 μ0 4.2 ∗ 10−4 − (3 ∗ 0.18)
Φ1 =
R,ir
= 5000
lgb + (lbcfg − lair ) 3 ∗ 0.18 + (2 ∗ 0.12 + 0.18 − 0.00076)
( + l ) ( + 0.00076)
μR,ir air 5,000

= 0.0054993 Wb
Φair 0.0054993
This gives us a flux density in the airgap of Bair = = = 13.09 T, and this is much
Aair 4.2∗10−4
higher than allowed, since we have the same flux density in the iron parts then as well. We
therefore know that the maximum flux is limited by the total iron core saturation, leading to
Φ1 + Φ2 = BT AΦT = 2 ∗ 4 ∗ 10−4 = 8 ∗ 10−4 . We know that half of this will travel down leg 2,
4∗10−4
Φ2 = 4 ∗ 10−4 Wb, which then gives us that Φ1 = 4 ∗ 10−4 Wb and Bair = = 0.952 T.
4.2∗10−4

c) How much current would we need in order to achieve Φ𝐓 = 𝟐 𝐓 in the iron core (assume
saturation in the iron)?
ΦT ΦT − Φ2 ΦT − Φ2
NI = lgb + (lbcfg − lair ) + l =
μ0 μR,ir AΦT μ0 μR,ir AΦT μ0 AΦT air
BT BT BT BT (lbcfg − lair ) lair μR,ir
lgb + (lbcfg − lair ) + lair = (lgb + + )=
μ0 μR,ir 2μ0 μR,ir 2μ0 μ0 μR,ir 2 2
2 2 ∗ 0.12 + 0.18 0.00076 0.00076
(3 ∗ 0.18 + − + 5,000 ∗ )
5,000 μ0 2 2 2
2 2 ∗ 2.64962
= (0.54 + 0.21 − 0.00038 + 1.9) = = 843.4 At
5,000 μ0 5,000 μo
Since we have 100 turns, this gives us:
843.4
I= = 8.434 A
100
d) What and how would we need to change the magnetic circuit in order to get 2 T in the
airgap (assume saturation in the iron)?
Do you have any suggestions? What could we do in order to increase the flux density in the
airgap?
Since we have already proven that it is impossible to get 2 T in the airgap with the current
configuration, we need to change the dimensions of iron core. The easiest way is to just remove
the leg in the middle. Since the cross-sectional area is the same throughout the rest of the core, if
we then achieve 2 T in the core we will also have 2 T in the airgap.
However, assuming this is not allowed, then we need to instead increase the area of the rest of
the core, along the length of lghab.

We know that the flux density will be 2 T in the airgap and the surrounding iron, giving us Φ1 =
B1 A1 = 2 ∗ 4.2 ∗ 10−4 = 8.4 ∗ 10−4

According to KCL we have that ΦT = Φ1 + Φ2 , meaning that:

ΦT = 4 ∗ 10−4 + 8.4 ∗ 10−4 = 12.4 ∗ 10−4 Wb

The flux density is still not allowed to be higher than 2 T, giving us the following area:
ΦT 12.4 ∗ 10−4
A= = = 6.1 ∗ 10−4 m2
BT 2
We need to increase the area of the main leg in the iron core by 1.9 ∗ 10−4 m2 in order to not
saturate it and produce a flux density of 2 T in the airgap.

Exercise 3 Operation of permanent magnets

If we reuse the iron core from exercise 1, but this time we insert a permanent magnet into the
airgap, with the north side pointing upwards, towards point c. We state the parameters of the
iron core again, together with those from the permanent magnet.

μ0 = 1.25664 ∙ 10−6 mkgc −2 A−2

μr, air = 1, μr,iron = 5,000

lab = lef = 0.07 m

laf = lbe = 0.03 m
lair = lcd − lPM = 0.003 m
w = lhj = 0.01 m

d = ljk = 0.01 m

Nt = 250 turns
lPM = 0.005 m
BR = 1.30 T
μr,PM = 1.08

a) Draw the hysteresis loop (BH curve) of the second quadrant for the given magnet.
A lot of hard permanent magnet materials (especially Neodynium-magnets normally used in
EVs and other motors) are almost completely linear in the second quadrant. We already know
that BR = 1.30 T, which is the place where the hysteresis curve cross the y-axis (when Hm , the
magnets internal field, is equal to 0). We know that:
BPM = μ0 μR Hm + Br
The other point we need to find is when the flux density in the magnet, BPM , is equal to 0, which
leads to:
 0 = μ0 μR Hm + Br
−Br 1.3 kA
∴ Hm = = − = −958.9
μ0 μR 1.08μ0 m
We can now draw the loop in the second quadrant as a straight line between the two points:

b) Find the magnetic flux, Φ𝐚𝐢𝐫, and the magnetic flux density B𝐚𝐢𝐫 if there is no current.
We will start by looking into how the permanent magnet fits within Ampere’s law. We have that
the flux density of a hard magnetic material can be described with:
BPM = μ0 μR Hm + Br
There are two ways that we can look at this, but they both yield the same answer. We will start
with Ampere’s law:

Nt I = Hir ld→c + Hair lair + HPM lPM

Here, Nt I = 0. We can transform the rest of the expression the same way we did in exercise 1.
Bir Bair Φir Φair
0= ld→c + lair + HPM lPM = ld→c + l + HPM lPM
μ0 μR,ir μ0 μ0 μR,ir Air μ0 Aair air

We now want to find an expression for HPM .

BPM Br
HPM = −
μ0 μR μ0 μR
BPM lPM Br lPM lPM Br lPM
HPM lPM = − = ΦPM −
μ0 μR μ0 μR μ0 μR APM μ0 μR
Br lPM
Hm lm = − + ΦPM 𝐑
μ0 μR
This we can then insert into the previous expression:
Φir Φair ΦPM Br lPM
0= ld→c + lair + lPM −
μ0 μR,ir Air μ0 Aair μ0 μR APM μ0 μR

Since we assume no leakage flux and the area is the same for all parts, Air = Aair = APM , we
know that Φir = Φair = ΦPM = ΦT
ΦT ΦT ΦT Br lPM
0= ld→c + lair + lPM −
μ0 μR,ir AΦT μ0 AΦT μ0 μR AΦT μ0 μR

ΦT ld→c lPM Br lPM

0= ( + lair + )−
μ0 AΦT μR,ir μR,PM μ0 μR

Bl −1.30 ∗ 0.005
μ0 AΦT ∗ μr μPM 10−4 ∗
ΦT = 0 R
= 1.08
ld→c lPM 2 ∗ 0.07 + 2 ∗ 0.03 − 0.006 0.005
μR,ir + lair + μR,PM 5,000
+ 0.001 + 1.08

ΦT = −0.00010618
ΦT −1.0618 ∗ 10−4
BT = = = −1.06 T
AT 10−4

c) Find the magnetic flux, Φ𝐚𝐢𝐫, if 𝐈 = 𝟏𝟎 𝐀.

We can use the same expression, and just exchange Nt I = 0 with Nt I = 250 ∗ 10
ΦT ld→c lPM Br lPM
2500 = ( + lair + )−
μ0 AΦT μR,ir μR,PM μ0 μR

Bl
2500μ0 AΦT μ0 AΦT μr μPM
0 R
+ = ΦT
ld→c lPM ld→c l
(μ + lair + μ ) (μ + lair + μ PM )
R,ir R,PM R,ir R,PM

2500μ0 AΦT
− 0.00010618 = ΦT
ld→c lPM
(μ + lair + μ )
R,ir R,PM

ΦT = 0.00005542465 − 0.00010618 = −5.0755 ∗ 10−5

ΦT −5.0755 ∗ 10−5
BT = = = −0.508 T
A ΦT 10−4

What is interesting here is that we were able to lower the effect on the flux coming from the
magnet, compare the flux and flux density in b) and c). This is a method that we utilize a lot in
electric vehicle motor operation, and there we call it flux weakening (or field weakening).
Exercise 4 Magnetic circuits in electric machines
Now we will take what we have learned and implement it on a real motor model.
The phases are divided into the following:
Phase A, yellow
Phase B, green
Phase C, blue
Dark colors indicate positive direction (current out from the paper) and lighter colors indicate
negative direction (current into the paper).
The motor has the following specifications:
Permanent magnet:
Length (along flux direction): l_m = 0.006 m
Magnet remanence: B_r = 1.4 T

Stator:
Stator tooth width w_T = 0.006 m
Stator yoke width = 0.066 m
Current phase A, I_A =100 A
Current phase B, I_C =-50 A
Current phase C, I_C =-50 A