Reliability

The world after the Product is made

 Reliability
The ability of a product, part, or system to perform its
intended function under a prescribed set of conditions

Reliability management involves establishing,

achieving, and maintaining reliability objectives for
products (e.g., the expected life of a particular make of
light bulb may be specified to be 5,000 hours).
 Reliability: How to increase overall reliability

Enhance design to avoid failure

Overdesign (e.g, use more durable material)

Design Reduce # of components

simplification,

Use
redundancy in By providing backup components
the design
 Reliability: Rule 1

Rule 1: If two or more events are independent and

“success” = occurrence of all of the events, then
Then Probability of success Ps = product of the probabilities
of the events,
In reliability terms, multi-
components are in series (if
one fails, whole system fails)

Ps = P1 × P2 ×….. Pn
 Reliability: Rule 1. Example 1
Suppose a room has two lamps, but to have adequate light, both
lamps must work (success) when turned on. One lamp has a
probability of working of 0.90, and the other has a probability of
working of 0.80. What’s the probability that both will work?

Ps = P1 × P2 ×….. Pn

Lamp 1 Lamp 2

.90 .80 .90 x .80 = .72

 Reliability: Rule 2
Rule 2: If 2 or more events are independent and “success”
is defined as occurrence of at least one of the events,
then the probability of success Ps is equal to 1 – probability that
none of the events will occur, i.e., 1- (1-P1)(1-P2)(1- P3)….
In reliability terms, multi-components are
in parallel (if one fails, the other
component still function and overall
system still functions )

Ps = P1 + (1- P1) P2+ (1- P1) (1- P2) P3+…..

 Reliability: Rule 2. Example 2
There are two lamps in a room. When turned on, one has probability of working of
0.90 and the other has probability of working of 0.80. Only a single lamp is needed to
light the room for success. What’s the probability of success?
Note that the threshold for success is different in this example

.80 Lamp 2 (backup)

.90 .90 + (1-.90)*.80 = .98

Lamp 1
 Reliability: Failure Rate as a Function of Time

number of
products fail Product
Random
shortly after wears out
Failures
they are put
ng rate
into service lo l u re es
sts e Fai creas
Failure rate

a
L tim
u se in
e ca are to
b ey ve h
th ecti wit
f
de egin
b
0
Burn-in Steady state (random failures) Wear-out

Rate of failure Time, T

decrease rapidly
as defective
items are
weeded out
Reliability: Finding Probability of Functioning for a
Given Length of Time

Failure rate = # of failures

per hour Total operating hours
 Reliability: Finding Probability of Functioning for a
Given Length of Time. Example 3
Two hundred units of a particular component were subjected to accelerated life testing
equivalent to 2,500 hours of normal use. One unit failed after 1,000 hours and another
after 2,000 hours. All other units were still working at the conclusion of the test. What is
the failure rate per hour?
Failure rate = # of failures
2 units failed overall per hour Total operating hours
1 unit failed after 1000 hours Remaining units (200 – 2 = 198)
failed after 2500 hours
1 unit failed after 2000 hours

The failure rate per hour = 2/[198(2,500) + 1,000 + 2,000] =

0.000004016 per hour
This assumes constant failure rate
Reliability: Finding Probability of Functioning for a Given
Length of Time
• Mean time to failure (MTTF)
• Average length of time before failure of a product or
component
• Mean time between failures (MTBF)
• Average time from the up time after a repair following a
failure to the next failure
• For repairable items.

MTTF  1 / Failure rate per hour ds

n e
ta
T s r tim
Reliability = P(no failure before T ) = e-T/MTTF fo

P(failure before T ) = 1 − e−T/MTTF

Reliability: Exponential Distribution Curve

Re liability  Pno failure before T   e T MTTF

12
Reliability: Inverse of the failure rate

The inverse of failure rate per hour is mean time to failure (MTTF), the
average length of time (in hours) before failure.
Reliability: Finding Probability of Functioning for a Given
Length of Time. Example 4
For example 3, where we found the failure rate to be 0.000004016 per hour. Find the
Mean time to failure

MTTF  1 / Failure rate per hour

MTTF = 1/Failure rate per hour = 1/0.000004016 =

249,000 hours
Reliability: Finding Probability of Functioning for a Given
Length of Time. Example 5a
By means of extensive testing and data collection, a manufacturer has determined that a
particular model of its vacuum cleaners has an expected life that is exponential with a
mean of four years and insignificant burn-in phase. Find the probability that one of these
vacuum cleaners will have a life that ends: a. After the initial four years of service;

MTTF= 4 years  
Re liability  P no failure before T  e T MTTF

T = 4 years

T/MTTF = 4/4 = 1.00

Reliability = e-1.00 = 0.3679.

Reliability: Finding Probability of Functioning for a Given
Length of Time. Example 5b

Find the probability that one of these vacuum cleaners will have a life that ends: b. Before
four years of service are completed;

T = 4 years
P(failure before T ) = 1 − e−T/MTTF
MTTF= 4 years

T/MTTF = 4/4 = 1.00

P(failure before T ) = 1-e-1.00 = 1 - 0.3679. = 0.6321

Reliability: Finding Probability of Functioning for a Given
Length of Time. Example 5c

Find the probability that one of these vacuum cleaners will have a life that ends: c. Not
before six years of service.

T = 6 years
Re liability  Pno failure before T   e T MTTF

MTTF= 4 years

T/MTTF = 6/4 = 1.5

Reliability = e-1.50 = 0.2231

Reliability: Normal Curve and wear-out time

T  Mean wear - out time

z
Standard deviation of wear - out time

Figure 4S-5
18
Reliability: Wear-out time Example 6
The mean life of a certain ball bearing can be modelled using a Normal distribution with a
mean of six years and a standard deviation of one year. Determine each of the following:
a. The probability that a ball bearing will fail before seven years of service.
b. The probability that a ball bearing will fail after seven years of service (i.e., find its
reliability).
c. The service life that will provide a failure probability of 10 percent.

T  Mean wear - out time

z
Standard deviation of wear - out time
Reliability: Wear-out time Example 6a
The mean life of a certain ball bearing can be modelled using a Normal distribution with a
mean of six years and a standard deviation of one year. Determine each of the following:
a. The probability that a ball bearing will fail before seven years of service.

T  Mean wear - out time

z= (7-6)/1 = +1.00 z
Standard deviation of wear - out time
Reliability: Wear-out time Example 6b
The mean life of a certain ball bearing can be modelled using a Normal distribution with a
mean of six years and a standard deviation of one year. Determine each of the following:
b. The probability that a ball bearing will fail after seven years of service (i.e., find its
reliability).
T  Mean wear - out time
z= (7-6)/1 = +1.00 z
Standard deviation of wear - out time
Reliability: Wear-out time Example 6c
The mean life of a certain ball bearing can be modelled using a Normal distribution with a
mean of six years and a standard deviation of one year. Determine each of the following:
c. The service life that will provide a failure probability of 10 percent
We don’t have a z, but we have a T  Mean wear - out time
probability (so we go in reverse) z
Standard deviation of wear - out time

-1.28 = T - 6 T =
4.72 years
 Reliability: Availability
Availability is the fraction of time a piece of equipment is
expected to be available for operation.

MTTF
Availabili ty 
MTTF  MTTR
MTTF = mean time to failure
MTTR = mean time to repair

Implications:
 Availability increases as the mean time to failure increases
 Availability increases as the mean repair time decreases
Reliability: Availability. Example 7
A copier is expected to operate for 200 hours after repair, and the mean repair time is
expected to be two hours. Determine the availability of the copier.

MTTF
MTBF = 200 hours Availabili ty 
MTTF  MTTR
MTTR = 2 hours

Availability = MTTF/(MTTF + MTTR) = 200/(200 + 2) =

0.990099
Reliability: Example 8a
The guidance system of a ship is controlled by a computer that has three major modules.
In order for the computer to function properly, all three modules must function. Two of
the modules have reliability of 0.97, and the other has reliability of 0.99.
a) What’s the reliability of the computer?
All three must function,
so modules are in series
Ps = P1 × P2 ×….. Pn

Ps = 0.97 × 0.97 × 0.99 = 0.931

Reliability: Example 8b
The guidance system of a ship is controlled by a computer that has three major modules.
In order for the computer to function properly, all three modules must function. Two of
the modules have reliability of 0.97, and the other has reliability of 0.99.
b) A backup computer identical to the one being used can be installed to improve overall
reliability. Assuming that the new computer can automatically function if the first
computer fails, determine the resulting reliability.
Prob. (works) = 0.931
Ps = P1 + (1- P1) P2
Prob. Original (works) + Prob. Original (not work)* Prob. backup(works)

Ps = 0.931 + (1-0.931)(0.931) = 0.995

Reliability: Example 8c
The guidance system of a ship is controlled by a computer that has three major modules.
In order for the computer to function properly, all three modules must function. Two of
the modules have reliability of 0.97, and the other has reliability of 0.99.
c) If the backup computer must be activated by a switch in the event that the first
computer fails, and the switch has a reliability of 0.98, what is the overall reliability of the
system? (Both the switch and the backup computer must function in order for the backup
system to function.)
Prob. (works) = 0.931
Ps = P1 + (1- P1) P2 P3
Prob(works) + [Prob(not work) × Prob(switch works) × Prob(backup works)]

Ps = 0.9315 + [(1 -0 .9315) × 0.98 × 0.9315] = .994

Reliability: Example 9a
One of the industrial robots designed by a leading producer has four major components.
Components’ reliability are 0.98, 0.95, 0.94, and 0.90. All of the components must
function in order for the robot to operate effectively.
a. Calculate the reliability of the robot.

Ps = P1 × P2 ×….. Pn

Ps =.98 × .95 × .94 × .90 = 0.7876

Reliability: Example 9b
One of the industrial robots designed by a leading producer has four major components. Components’
reliability are 0.98, 0.95, 0.94, and 0.90. All of the components must function in order for the robot to
operate effectively.
b. Designers want to improve the reliability of the robot by adding a backup component. Due to space
limitations, only one backup can be added. The backup for any component will have the same reliability as
the unit for which it is the backup. Which component should get the backup in order to achieve the highest
reliability of the robot?
Ps = P1 + (1- P1) P2 P3 Ps = P1 × P2 ×….. Pn
If 1st: [.98 + (1 - .98) × .98] × .95 × .94 × .90 = 0.8034 The fourth
component should
If 2nd: .98 × [.95 + (1 - .95) × .95] × .94 × .90 = 0.8270 be backed up
because it
If 3rd: .98 × .95 × [.94 + (1 – .94) × .94] × .90 = 0.8349 produces the
highest overall
If 4th: .98 × .95 × .94 × [.90 + (1 - .90) × .90] = 0.8664 system reliability
Reliability: Example 9c
One of the industrial robots designed by a leading producer has four major components. Components’
reliability are 0.98, 0.95, 0.94, and 0.90. All of the components must function in order for the robot to
operate effectively.
c. If one backup with a reliability of 0.92 can be added to any one of the main components, which
component should get it to achieve the highest overall reliability?

Ps = P1 + (1- P1) P2 P3 Ps = P1 × P2 ×….. Pn

We should add it to the one with a reliability of .90 since it poses the
greatest risk of failure

System overall reliability:

0.98 × 0.95 × 0.94 × [0.90 + (1 -0.90) × 0.92] =0.86814
Reliability: Example 10a
A production line has three machines A, B, and C, with reliabilities of 0.99, 0.96, and 0.93, respectively. The
machines are arranged so that if one breaks down, the others must shut down. Operations managers are
weighing two alternative designs for increasing the line’s reliability. Plan 1 involves adding an identical
backup line (i.e., a series backup), and Plan 2 involves providing a backup for each machine (i.e., a parallel
backup). In either case, three additional machines (A, B, and C) would be used with reliabilities equal to the
original three.
a. Which plan will provide higher reliability?

Let’s draw the options

Reliability: Example 10a
A production line has three machines A, B, and C, with reliabilities of 0.99, 0.96, and 0.93, respectively. The
machines are arranged so that if one breaks down, the others must shut down. Operations managers are
weighing two alternative designs for increasing the line’s reliability. Plan 1 involves adding an identical
backup line (i.e., a series backup), and Plan 2 involves providing a backup for each machine (i.e., a parallel
backup). In either case, three additional machines (A, B, and C) would be used with reliabilities equal to the
original three.
a. Which plan will provide higher reliability?

Calculate the probability that the line works for option 1

Prob.(line works) = .99 × .96 × .93 = 0.8839

Prob. of overall system works: Prob.(line works) + Prob.(line fails) × Prob.(backup works)
Prob. of overall system works (option 1) = .8839 + [(1 - .8839) × (.8839)]
= 0.9865
Reliability: Example 10a
A production line has three machines A, B, and C, with reliabilities of 0.99, 0.96, and 0.93, respectively. The
machines are arranged so that if one breaks down, the others must shut down. Operations managers are
weighing two alternative designs for increasing the line’s reliability. Plan 1 involves adding an identical
backup line (i.e., a series backup), and Plan 2 involves providing a backup for each machine (i.e., a parallel
backup). In either case, three additional machines (A, B, and C) would be used with reliabilities equal to the
original three.
a. Which plan will provide higher reliability?
Calculate the probability that the line works for option 2

Prob. (Comp. A) = Prob. (Comp. B) = Prob. (Comp. C) =

.99 + [(1 - .99) × .99] .96 + [(1 - .96) × .96] .93 + [(1 - .93) × .93]=
=0.9999 =0.9984 0.9951
Prob. of overall system works (option 2) = .9999 × .9984 × .9951 = 0.9934
Reliability: Example 10a
A production line has three machines A, B, and C, with reliabilities of 0.99, 0.96, and 0.93, respectively. The
machines are arranged so that if one breaks down, the others must shut down. Operations managers are
weighing two alternative designs for increasing the line’s reliability. Plan 1 involves adding an identical
backup line (i.e., a series backup), and Plan 2 involves providing a backup for each machine (i.e., a parallel
backup). In either case, three additional machines (A, B, and C) would be used with reliabilities equal to the
original three.
a. Which plan will provide higher reliability?

Prob. of overall system works (option 1) = 0.9865

Prob. of overall system works (option 2) = 0.9934

Plan 2 is better (.9934 > .9865)

Reliability: Example 10b
A production line has three machines A, B, and C, with reliabilities of 0.99, 0.96, and 0.93, respectively. The
machines are arranged so that if one breaks down, the others must shut down. Operations managers are
weighing two alternative designs for increasing the line’s reliability. Plan 1 involves adding an identical
backup line (i.e., a series backup), and Plan 2 involves providing a backup for each machine (i.e., a parallel
backup). In either case, three additional machines (A, B, and C) would be used with reliabilities equal to the
original three.
b. why are the two reliabilities not the same, which other factors might enter into the decision of
which plan to adopt?

In Plan 1 the system will fail if any one of original and any one of backup
machines fail.
In Plan 2 the system will fail if any one of original and its backup machine both
fail. So, overall failing is more likely for #1.

Other factors: Availability of the layout space for each plan & Ease of shifting to
backup when needed
Reliability: Example 11a
An electronic chess game has a useful life that is exponential with a mean of 30 months.
Determine each of the following:
a. The probability that any given unit will operate for at least (1) 39 months, (2) 60
months.

MTTF= 30 months
 
Re liability  P no failure before T  e T MTTF

T/MTTF = 39/30 = 1.3

T= 39
Reliability = e-1.30 = 0.2725.

T/MTTF = 60/30 = 2
T= 60
Reliability = e-2 = 0.1353
Reliability: Example 11b
An electronic chess game has a useful life that is exponential with a mean of 30 months.
Determine each of the following:
b. The probability that any given unit will fail sooner than (1) 33 months, and (2) 6
months.

MTTF= 30 months P(failure before T ) = 1 − e−T/MTTF

T/MTTF = 33/30 = 1.1
T= 33
P(failure before T ) = 1 − e−1.1 = 0.6671
T/MTTF = 6/30 = 0.2
T= 6
P(failure before T ) = 1 − e−0.2 = 0.1813
Reliability: Example 11c
An electronic chess game has a useful life that is exponential with a mean of 30 months.
Determine each of the following: c. The length of service time after which the percentage
of failed units will approximately equal (1) 50 percent, (2) 85 percent, (3) 95 percent, (4)
99 percent.

MTTF= 30 months P(failure before T ) = 1 − e−T/MTTF

Prob.= 1 − e−T/MTTF ln(1-Prob)=-T/MTTF
Prob.- 1= − e−T/MTTF T =- ln(1-Prob)*MTTF
1-Prob.= e−T/MTTF
Prob. = 0.5 T =- ln(1-0.5)*30 = 20.79
Reliability: Example 11c
An electronic chess game has a useful life that is exponential with a mean of 30 months.
Determine each of the following: c. The length of service time after which the percentage
of failed units will approximately equal (1) 50 percent, (2) 85 percent, (3) 95 percent, (4)
99 percent.

MTTF= 30 months
T =- ln(1-Prob)*MTTF

Prob. = 0.85 T =- ln(1-0.85)*30 = 56.91

Prob. = 0.95 T =- ln(1-0.95)*30 = 89.97
Prob. = 0.99 T =- ln(1-0.99)*30 = 138.16
Reliability: Example 12a
Lucky Lumen light bulbs have an expected life that is exponentially distributed with a
mean of 20,000 hours. Determine the probability that one of these light bulbs will last:
a. At least 24,000 hours.
No Std. Dev. is given & Exponentially dist. (not normally dist.)
MTTF= 20,000 hrs
 
Re liability  P no failure before T  e T MTTF
T= 24,000 hrs

T/MTTF = 24000/20000 = 1.2

Prob. that one lightbulb will last = e-1.2 = 0.3012

Reliability: Example 12b
Lucky Lumen light bulbs have an expected life that is exponentially distributed with a
mean of 20,000 hours. Determine the probability that one of these light bulbs will last:
b. No longer than 4,000 hours.

MTTF= 20,000 hrs

P(failure before T ) = 1 − e−T/MTTF
T= 4,000 hrs

T/MTTF = 4000/20000 = 0.2

Prob. lightbulb fails before 4000 hrs = 1- e-0.2 = 0.1813

 Reliability: Example 12c
Lucky Lumen light bulbs have an expected life that is exponentially distributed with a
mean of 20,000 hours. Determine the probability that one of these light bulbs will last:
c. Between 4,000 hours and 24,000 hours.

MTTF= 20,000 hrs P(failure before T ) = 1 − e−T/MTTF

T= Between 4,000 and 24,000 hrs

Prob. lightbulb fails before 4000 hrs = 0.1813

Prob. that one lightbulb will fail atleast 24000 hrs = 0.3012
Prob. that one lightbulb will fail before 24000 hrs = 1 - 0.3012 = 0.6988

Prob. (4000 ≤ x ≤ 24000) = 0.6988 – 0.1813 = 0.5175

 Reliability: Example 13a
A copier manufacturer has determined that its major product has a service life that can
be modelled by Normal distribution with mean of six years and standard deviation of
one-half year. a. What probability can you assign to service lives of at least five years?

μ= 6 yrs σ= 0.5 yrs Std. Dev. is given & Normally dist.

Based on z curve to the right, shading of

prob. Starts from halfway  -z so left side
of curve  At least (more than), which means
everything right to z  so 0.5 + 0.4772 =
0.9772
 Reliability: Example 13b
A copier manufacturer has determined that its major product has a service life that can
be modelled by Normal distribution with mean of six years and standard deviation of
one-half year. b. What probability can you assign to service lives of at most seven and a
half years?
μ= 6 yrs σ= 0.5 yrs

Based on z curve to the right, shading

of prob. Starts from halfway  +z so
right side of curve  At most, which
means everything left to z  so 0.5 +
0.4987 = 0.9987
 Reliability: Example 13c
A copier manufacturer has determined that its major product has a service life that can
be modelled by Normal distribution with mean of six years and standard deviation of
one-half year. c. If the manufacturer offers a service warranty of four years on these
copiers, what percentage can be expected to fail during the warranty period?
μ= 6 yrs σ= 0.5 yrs

Based on z curve to the right, shading

of prob. Starts from halfway  -z so
right side of curve  4 years or less,
which means everything left to z 
so 0.5 - 0.5 = approximately zero
Reliability: Example 13d
A copier manufacturer has determined that its major product has a service life that can
be modelled by Normal distribution with mean of six years and standard deviation of
one-half year. d) What probability can you assign to service lives of at least six years?

μ= 6 yrs σ= 0.5 yrs

At least means shading is to the

right  Which means everything
right to z (0.5)  At z = 0, prop. =
0 0 + 0.5 = 0.5
More Exercises
 Example A
Determine the reliability of the system shown below.

We need to first reduce the series into three Ps = P1 + (1- P1) P2+ (1-
components using rule 2
P1) (1- P2) P3+…..

Next, we compute the system reliability

 Example B
Due to the extreme cost of interrupting production, a manufacturer has two standby
machines available in case a particular machine breaks down. The machine in use has a
reliability of 0.94, and the backups have reliabilities of 0.90 and 0.80. In the event of a
failure, a backup machine is brought into service. If this machine also fails, the other backup
is used. Calculate the system reliability.
 Example C
A weather satellite has expected life of 10 years from the time it is placed into Earth’s orbit.
Determine its reliability for each of the following lengths of service (assume that
exponential distribution is appropriate). a. 5 years; b. 12 years; c. 20 years; d. 30 years

Calculate the ratio T/MTTF for T = 5, 12, 20, and 30

 Example D
A weather satellite has expected life of 10 years from the time it is placed into Earth’s orbit.
What’s the probability that the satellite will fail between 5 and 12 years.

P(5 years < failure < 12 years) = P(failure after 5 years) − P(failure after 12 years)
 Example E
One line of specialty tires has a wear-out life that can be modelled using Normal
distribution with a mean of 25,000 km and a standard deviation of 2,000 km. Determine
each of the following: a. The percentage of tires that can be expected to wear out within
±2,000 km of the average (i.e., between 23,000 km and 27,000 km); b. The percentage of
tires that can be expected to fail between 26,000 km and 29,000 km.; c. For what tire life
would you expect 4 percent of the tires to have worn out?

Here instead
T  Mean wear - out time of time, we
z have
Standard deviation of wear - out time
kilometres
 Example E part a
One line of specialty tires has a wear-out life that can be modelled using Normal
distribution with a mean of 25,000 km and a standard deviation of 2,000 km. Determine
each of the following: a. The percentage of tires that can be expected to wear out within
±2,000 km of the average (i.e., between 23,000 km and 27,000 km);
The phrase “within ±2,000 km of the average” translates to within one standard
deviation of the mean because the standard deviation equals 2,000 km.

Therefore, the range of z is z = −1.00 to z = +1.00,

the area under the curve between those points is found
as the difference between P(z < +1.00) and P(z < −1.00)

P(−1.00 < z < +1.00) = 0.6826

which means 68.26 percent of tires will wear out

between 23,000 km and 27,000 km
 Example E part b
One line of specialty tires has a wear-out life that can be modelled using Normal
distribution with a mean of 25,000 km and a standard deviation of 2,000 km. Determine
each of the following: b. The percentage of tires that can be expected to fail between
26,000 km and 29,000 km
 Example E part c
One line of specialty tires has a wear-out life that can be modelled using Normal
distribution with a mean of 25,000 km and a standard deviation of 2,000 km. Determine
each of the following: c. For what tire life would you expect 4 percent of the tires to have
worn out?

Find z for 4 percent: z = −1.75. Find tire life using μ + zσ: 25,000
− 1.75(2,000) = 21,500 km.
 Example F
Five assembly units were tested for failure time. The observed failure times (in days) were
550, 680, 820, 910, and 1,110. The units go on a product called DreamXT. Calculate the
MTTF and the failure rate for DreamXT.
Failure rate = 1 / MTTF

MTTF = (550 + 680 + 820 + 910 + 1,110) / 5 = 814 days

Failure rate = 1 / MTTF = 1 / 814 = .0012285 per day

 Example G
Auto batteries have an average life of 2.7 years. Battery life is normally distributed with a
mean of 2.7 years and a standard deviation of 0.3 year. The batteries are warranted to
operate for a minimum of two years. If a battery fails within the warranty period, it will be
replaced with a new battery at no charge. The company sells and installs the batteries. Also,
the usual $5 installation charge will be waived. What percentage of batteries would you
expect to fail before the warranty period expires?
 = 2.7 years  = .3 year x = 2 years
 Example G

P(-2.33) = 0.4901

We want the
probability left of the
curve

0.5 – 0.4901 = 0.0099

 Example G
Auto batteries have an average life of 2.7 years. Battery life is normally distributed with a
mean of 2.7 years and a standard deviation of 0.3 year. The batteries are warranted to
operate for a minimum of two years. If a battery fails within the warranty period, it will be
replaced with a new battery at no charge. The company sells and installs the batteries. Also,
the usual $5 installation charge will be waived. What percentage of batteries would you
expect to fail before the warranty period expires?
 = 2.7 years  = .3 year x = 2 years

We would expect approximately (0.0099) 1% of the

batteries to fail before the warranty period ends.