MANE-4030 Elements of Mechanical Design

Homework 1 Solutions

1. The engineer should refuse to do the repair unless the owner/operator of the turbine
can demonstrate satisfactorily to the engineer (through analysis and/or test) that the
turbine will be safe to operate. According to the NSPE Code of Ethics for Engineers,
Section II.1.a. “If engineers’ judgement is overruled under circumstances that endanger
life or property, they shall notify their employer or client and such other authority as
may be appropriate.” Engineers should never make a repair that they deem unsafe.
If the owner/operator and the supervisor insist on the potentially unsafe repair and
go forward with it, the engineer should notify others (higher level people at either her
company or the owner/operator company) of her concern. She should also consider
looking for another job if her supervisor is willing to perform unsafe repairs.

2. Based on the information given, the rating numbers assigned to each of the eight rating
factors might be:

Rating Factor Selected Rating Number (RN)

1. Accuracy of loads knowledge 0
2. Accuracy of stress calculation -2
3. Accuracy of strength knowledge -2
4. Need to conserve -3
5. Seriousness of failure consequences +4
6. Quality of manufacture 0
7. Operating conditions +2
8. Quality of inspection/maintenance +2

From (2-85):
t = 0-2-2-3+4+0+2+2=+1

and, since t ≥ −6, from (2-86),

(10 + 1)2
nd = 1 + = 2.2
100

3. To determine the stress, we need the force and cross-sectional area. The force amplitude
is variable, so the stress amplitude will also be variable. The mean and standard
deviation of the amplitude of the applied stress is:

1
 2, 000 lb
µ̂σ = = 42, 667 psi
(0.375 in)(0.125 in)
400 lb
σ̂σ = = 8, 533 psi
(0.375 in)(0.125 in)

A plot of the probability distributions for the stress amplitude and fatigue strength are
shown in Fig. 1. Now we find the mean and standard deviation of the random variable

Figure 1: Sketch of probability distributions of stress and strength of tensile member.

y = S − σ.

µ̂y = µ̂S − µ̂σ = 68, 000 psi − 42, 667 psi = 25, 333 psi
q p
σ̂y = σ̂S2 + σ̂σ2 = (3, 200 psi)2 + (8, 533 psi)2 = 9, 114 psi

A plot of the distribution of y is shown in Fig. 2. I need to find the parameter X to use
in Table 2.9. The way I like to think of this is that I need to know how many standard
deviations are between the mean and the value I’m interested in, which in this case is
y = 0 (i.e. I want to know the probability y < 0 or P {y < 0}). So then

0 − 25, 333
X= = −2.78
9, 114

So there are 2.78 standard deviations between the mean and 0. I look up X = 2.78
(since the normalized distribution is symmetric about X = 0, it doesn’t matter if it is

2
Figure 2: Sketch of probability distributions for random variable associated with reliability,
y.

plus or minus, I just need to remember now I’ve reflected the distribution about the
mean since I changed the sign, because the normal distribution is symmetric about the
mean). In Table 2.9, I find the F (X) = 0.9973. Now F (X) is the integral from −∞ to
X on the standard normal distribution. I’m interested in the region in the tail, which
is on the other side of X and is 1 − F (X) = 0.0.0027 = 0.27%. This is the probability
of failure, so out of 1000 parts, 3 are expected to fail.

4. (a) Using Eqn. (2-101), for a system having 3 components in series, each with a relia-
bility of 0.90
Rs = (0.98)3 = 0.9412

Thus the system reliability would be 94.1%.

(b) Adding a duplicate redundant system, then we use Eqn. (2-102) to obtain the
system reliability
Rs = 1 − (1 − 0.9412)2 = 0.9965

Thus, the system reliability would be 99.65% if a redundant system were added. This
gives a significant improvement in reliability (from a failure rate of 1 in 17 to a failure
rate of 1 in 286).
(c) Adding a third redundant system, following the same procedure as in part (b) we
obtain a reliability of 99.98% (or a failure rate of 1 in 5000). While the reliability
can be improved significantly reducing the failure rate, adding redundant subsystems
increases cost, weight, and space. Creating/using more reliable components is likely a
better way to get improvements in reliability and would allow for less redundancy to

3
 get the desired overall reliability.

5. For a heavy drive fit, we need FN 3 class fit (Table 6.6). From Table 6.7, we find for
a nominal size of 1.735 in, the limits on the shaft outer diameter:

ds(max) = 1.7350 + 0.0028 = 1.7378 in

ds(min) = 1.7350 + 0.0022 = 1.7372 in

and on the gear bore diameter are:

dh(min) = 1.7350 + 0.0000 = 1.7350 in

dh(max) = 1.7350 + 0.0010 = 1.7360 in

or they may be written in the following form on an engineering drawing:

1.7378
ds = in (sleeve)
1.7372
1.7350
dh = in (hole)
1.7360

That is, the shaft outer diameter must be between 1.7378 and 1.7372 inches, and the
housing bore diameter must be between 1.7350 and 1.7360 inches. Note, for the shaft,
we put the larger number on top, because in the metal removal process to get the
sleeve outer diameter down to the desired diameter, we go from larger to smaller. For
the hole, the smaller number is on top for a similar reason, during machining, the hole
goes from a smaller to a larger diameter.
Note: The limits of interference are 0.0012 to 0.0028 in.

6. The local stress state on the steel axle of a locomotive is σxx = 40 MPa, σyy = 0,
σzz = −15 MPa, σxy = −20 MPa, and all other stress components are zero.
(a) The local strain state can be found using Hooke’s Law, Eqs. (5-15)-(5-20) in your
book. The material is steel, so the elastic properties are (Table 3.9) E = 207 GPa,

4
G = 79 GPa, and ν = 0.30.

1
εxx = [40 − (0.3)(0 − 15)] = 2.15 × 10−4
207, 000
1
εyy = [0 − (0.3)(40 − 15)] = −3.62 × 10−5
207, 000
1
εzz = [−15 − (0.3)(40 + 0)] = −1.30 × 10−4
207, 000
−20
γxy = 2εxy = = −2.53 × 10−4
79, 000
γxz = 2εxz = 0
γyz = 2εyz = 0

(b) The Mohr’s circle is shown in Fig. 3. From the circle, the principal stresses are
computed as
r
1 40 − 0 2
h = (40) MPa = 20 MPa , R = ( ) + (20)2 = 28.28 MPa
2 2
σ1 = h + R = 48.28 , σ2 = h − R = −8.28 , σ3 = −15 MPa

The maximum shear stress is the radius of the largest circle, or the half the difference
of the maximum and minimum principal stresses. Thus,

1 1
τmax = (σ1 − σ3 ) = (48.28 + 15) = 31.64
2 2

Figure 3: Mohr’s circle diagram. Stress element diagram shown in upper right.

Let’s verify with an online calculator.

5
Figure 4: Using an online tool (https://www.graniteng.com/mohr-3d?lang=en) to estimate
the principal stresses.

Eigenvalue formulation (Use this approach in the exam):

The local stress state on the steel axle of a locomotive is given by:

σxx = 40 MPa, σyy = 0 MPa, σzz = −15 MPa, σxy = −20 MPa

and all other stress components are zero.

The stress tensor in matrix form is:
 
40 −20 0
σ = −20 0 0 
 

0 0 −15

To find the principal stresses, we solve the characteristic equation:

det(σ − λI) = 0

where λ are the eigenvalues of the matrix (principal stresses) and I is the identity
matrix. The characteristic equation is:
 
40 − λ −20 0
det  −20 −λ 0 =0
 

0 0 −15 − λ

6
Expanding this determinant, we get:

−λ 0 −20 0
(40 − λ) − (−20) =0
0 −15 − λ 0 −15 − λ

Simplifying the determinants inside:

(40 − λ)(λ(15 + λ)) − 400(15 + λ) = 0

(40 − λ)(−15λ − λ2 ) − 400(15 + λ) = 0

−600λ − 40λ2 + 15λ2 + λ3 − 6000 − 400λ = 0

λ3 − 25λ2 − 1000λ − 6000 = 0

We solve this cubic equation to find the eigenvalues λ, which are the principal stresses.
To find the maximum shear stress, we use the principal stresses found above. The
maximum shear stress τmax is given by:

σ1 − σ3
τmax =
2

where σ1 and σ3 are the maximum and minimum principal stresses, respectively.
Let’s solve for the eigenvalues using the characteristic equation:

λ3 − 25λ2 − 1000λ − 6000 = 0

This cubic equation can be solved numerically or using software tools to find:

λ1 ≈ −15 MPa, λ2 ≈ −8.2843 MPa, λ3 ≈ 48.2843 MPa

The principal stresses must be :

σ1 = 48.2843 MPa, σ2 = −8.2843 MPa, σ3 = −15 MPa

48.2843−(−15)
Maximum shear stress, τmax = 2
= 31.64 MPa

7
7. (a) At A, we can see that there will be shear stress due to torsion and bending stress.
To compute the stress components, we need to find the torque and bending moment
acting at A. The torque acting on the shaft at this location is T = 600 N-m. To
compute the bending moment, we need to make a cut at A and find the moment that
balances the forces and moments on the remainder of the shaft. Before we do that, we
need to find the force P and the reactions at the bearings, using equilibrium. Balancing
the torque (moment about x), allows us to find P ,

T − P (0.15 m) = 600 N-m − P (0.15 m) = 0 → P = 4000 N

Next, balancing the moments about z (we assume the flywheel is right against the
bearing, so that no moment is generated by its weight and the weight of the flywheel
is entirely supported by the left bearing at the origin),

−P (0.25 m)+RCy (0.5 m) = −(4000 N )(0.25 m)+RCy (0.5 m) = 0 → RCy = 2000 N

Since there is no force acting in the z direction on this system, RCz = 0. Also, we will
assume that the bearings only support a normal load, so there are only force reactions,
no moment reactions at the bearings (i.e. they prevent deflections of the shaft in y
and z, but they don’t prevent rotation of the shaft about y and z). Now we know all
the forces acting on the crankshaft to the right of A. Thus, we can make a free body
diagram of this section and find the bending moment in the shaft at A (Fig. 5). In
addition to the torque due to P , there is a moment about an axis parallel to z at A
(MzA ). Balancing the moments about an axis parallel to the z axis at A yields

MzA − (4000 N )(0.15 m) + (2000 N )(0.4 m) = 0 → MzA = −200 N-m

Since the sign is negative, the moment is acting in the direction opposite to the way
it is drawn in the free body diagram. This will generate a compressive stress on the
element at A along the x direction, as shown on the stress element in Fig. 6. If we look
at the direction the torque is acting to the left of A, balanced by that due to the force
P to the right of A, we can determine the direction of the resulting shear stress, which
is also shown on the stress element in Fig. 6. We can see that the shear stress τxz is
acting in the negative direction (it is acting in the −z direction on the surface whose
unit normal is along +x and in the −x direction on the surface whose unit normal is
along +z).

8
 Figure 5: Free body diagram of crank.

Figure 6: Stress element at A, looking down from the top. Note, the stresses are shown in
the direction in which they are acting; i.e. compressive/negative axial stress σx and negative
shear stress τxz .

The stresses at A due to the torsion and bending moment are then:

MzA c (−200 N-m)(0.0175 m)

σxx = σx = = π = −47.51 MPa
I 64
(0.035 m)4

Tc (600 N-m)(0.0175 m)
σxz = τxz = − =− π = −71.27 MPa
J 32
(0.035 m)4

9
 Figure 7: Mohr’s circle diagram for stress element shown in Fig. 6.

(b) Mohr’s circle based solution: A Mohr’s circle is shown in Fig. 7. This case
is different from what we typically expect. This is because it is with respect to the
z − x axes instead of the x − y axes that we commonly encounter. Note that in this
coordinate frame, the x axis is 90◦ counterclockwise from the z axis. As such, making
the analog to the x − y axes that was discussed in class, here the z axis is like the x
axis in the x − y frame and the x axis is like the y axis in the x − y frame. Thus,
the points we plot on the Mohr’s circle are (σz , −τxz ) and (σx , τxz ). This is important
so that the rotations on the Mohr’s circle will be in the same direction as in physical
space. Another way of thinking of it, that may be easier to remember, is that we plot
clockwise shear stresses as positive and counterclockwise shear stresses as negative.
First look at the face with normal along z in Fig. 6, on this face we have σ = 0 and
τ = 71.27 MPa acting clockwise about the center of element A, and so we plot the point
(0, 71.27). Now look at the face with normal along x, on this face we have σ = −47.51
MPa and τ = 71.27 MPa acting counterclockwise about the center of element A, so we
plot the point (-47.51, -71.27). This gives us the diameter of the Mohr’s circle. From
the Mohr’s circle diagram, we can compute the principal stresses:

1 p
h = (−47.51 MPa) = −23.76 MPa , R= (23.76)2 + (71.27)2 = 75.13 MPa
2

σ1 = h + R = 51.37 MPa , σ3 = h − R = −98.89 MPa , σ2 = 0

Note that σ2 is zero because the stresses on the shaft surface at A axis are zero, i.e.,

10
σyy = σxy = σyz = 0. We can also get the angle of rotation to the principal stress
directions ϕp from this diagram. From geometry, we see that

1 2 × (−71.27)
ϕp = tan−1 ( ) → ϕp = −35.8◦ = 35.8◦ (CW w.r.t. z-axis)
2 0 − (−47.21)

See part (c) and the stress element orientation. See the online Mohr’s circle calculator
results in Figs. 8 and 9.

Figure 8: Results obtained from online Mohr circle principle stress calculator.

Figure 9: Mohr’s circles from online Mohr circle principle stress calculator.

11
Eigenvalue formulation: The stresses at A due to the torsion and bending moment
are:
M Ac (−200 N-m)(0.0175 m)
σxx = σx = z = π = −47.51 MPa
I 64
(0.035 m)4
Tc (600 N-m)(0.0175 m)
σxz = τxz = − =− π 4
= −71.27 MPa
J 32
(0.035 m)

The stress tensor for this the plane stress condition σyy = 0, is:
 
−47.51 0 −71.27
σ= 0 0 0 
 

−71.27 0 0

To find the principal stresses, we solve the characteristic equation:

det(σ − λI) = 0

 
−47.51 − λ 0 −71.27
det  0 −λ 0 =0
 

−71.27 0 −λ

Expanding the determinant:

−λ 0 0 0
(−47.51 − λ) − 0 + (−71.27) =0
0 −λ −71.27 −λ

λ3 + 47.51λ2 − 5079.4λ = 0

This cubic equation can be solved numerically to obtain the eigenvalues approximately
as follows:

λ1 ≈ −98.87956 MPa, λ2 ≈ 0 MPa, λ3 ≈ 51.36956 MPa

Thus, the principal stresses are:

σ1 ≈ 51.3696 MPa, σ2 ≈ 0 MPa, σ3 ≈ −98.8796 MPa

12
 Figure 10: Direction of eigenvectors.

Figure 10 shows eigenvectors (obtained using MATLAB/Python) and the direction of

the principal stresses with respect to the z and x axes.

13
 (c) Figure 11 shows the principal stress element oriented with respect to the z and x
axes. Note that we rotate the stress element in the same direction as on the Mohr’s
circle (i.e. note that 2ϕp is going clockwise on the Mohr’s circle taking us from the
stress state in the z − x coordinates to the principal stresses, so we rotate the the stress
element by ϕp in the clockwise direction). We can also see on the Mohr’s circle that the
point (σz , −τxz ) rotates to σ1 . Thus, this face on the element rotated to the principal
stresses will have σ1 acting on it.

Figure 11: Stress elements oriented with the principal stresses.

8. Using Hooke’s law:

1 6000
εx = σx = = 2.0 × 10−4
E 30 × 106
ν −0.3 × 6000
εy = − σx = = −6.0 × 10−5
E 30 × 106
ν −0.3 × 6000
εz = − σx = = −6.0 × 10−5
E 30 × 106
εx + εy + εz = (2.0 − 0.6 − 0.6) × 10−4 = 8 × 10−5

14