www.getmyuni.

com

UNIT 3:

ANALYSIS OF AUTOMATED FLOW LINE & LINE BALANCING

General Terminology & Analysis:

There are two problem areas in analysis of automated flow lines which must be addressed:
1. Process Technology
2. Systems Technology
Process Technology refers to the body of knowledge about the theory & principles of the
particular manufacturing process used on the production line. E.g. in the manufacturing
process, process technology includes the metallurgy & machinability of the work material,
the correct applications of the cutting tools, chip control, economics of machining, machine
tools alterations & a host of other problems. Many problems encountered in machining can
be overcome by application of good machining principles. In each process, a technology is
developed by many years of research & practice.

Terminology & Analysis of transfer lines with no Internal storage:

There are a few assumptions that we will have to make about the operation of the Transfer
line & rotary indexing machines:
1. The workstations perform operations such as machining & not assembly.
2. Processing times at each station are constant though they may not be equal.
3. There is synchronous transfer of parts.
4. No internal storage of buffers.
In the operation of an automated production line, parts are introduced into the first
workstation & are processed and transported at regular intervals to the succeeding stations.
This interval defines the ideal cycle time, Tc of the production line. Tc is the processing
time for the slowest station of the line plus the transfer time; i.e. :

Tc = max (Tsi) + Tr ---------------- (1)

Tc = ideal cycle on the line (min)

Tsi = processing time at station (min)
Tr = repositioning time, called the transfer time (min)

In equation 1, we use the max (Tsi) because the longest service time establishes the pace of
the production line. The remaining stations with smaller service times will have to wait for
the slowest station. The other stations will be idle.

In the operation of a transfer line, random breakdowns & planned stoppages cause
downtime on the line.

Common reasons for downtime on an Automated Production line:

1. Tool failures at workstations.
2. Tool adjustments at workstations
3. Scheduled tool charges
4. Limit switch or other electrical malfunctions.

1
www.getmyuni.com

5. Mechanical failure of a workstation.

6. Mechanical failure of a transfer line.
7. Stock outs of starting work parts.
8. Insufficient space for completed parts.
9. Preventive maintenance on the line worker breaks.

The frequency of the breakdowns & line stoppages can be measured even though they occur
randomly when the line stops, it is down for a certain average time for each downtime
occurrence. These downtime occurrences cause the actual average production cycle time of
the line to be longer than the ideal cycle time.

The actual average production time Tp:

Tp = Tc + FTd ------------------------ 2
F = downtime frequency, line stops / cycle
Td = downtime per line stop in minutes
The downtime Td includes the time for the repair crew to swing back into action, diagnose
the cause of failure, fix it & restart the drive.

FTd = downtime averaged on a per cycle basis

Production can be computed as a reciprocal of Tp

Rp = 1 ----------------------------- 3
Tp
Where, Rp = actual average production rate (pc / min)
Tp = the actual average production time
The ideal production rate is given by

Rc = 1 ------------------------------ 4
Tc
Where Rc = ideal production rate (pc / min)

Production rates must be expressed on an hourly basis on automated production lines.

The machine tool builder uses the ideal production rate, Rc, in the proposal for the
automated transfer line & calls it as the production rate at 100% efficiency because of
downtime. The machine tool builder may ignore the effect of downtime on production rate
but it should be stated that the amount of downtime experienced on the line is the
responsibility of the company using the production line.
Line efficiency refers to the proportion of uptime on the line & is a measure of reliability
more than efficiency.
Line efficiency can be calculated as follows:

E = Tc = Tc + FTd ----------------- 5
Tp Tc

2
www.getmyuni.com

Where E = the proportion of uptime on the production line.

An alternative measure of the performance is the proportion of downtime on the line which
is given by:

D = FTd = FTd + FTd ----------------- 6

Tp Tc
Where D = proportion of downtime on the line

E + D = 1.0

An important economic measure of the performance of an automated production line is the

cost of the unit produced. The cost of 1 piece includes the cost of the starting blank that is to
processed, the cost of time on the production line & the cost of the tool consumed. The cost
per unit can be expressed as the sum of three factors:

Cpc = Cm + CoTp + Ct ------------------------ 7

Where Cpc = cost per piece (Rs / pc)
Cm = cost per minute to operate the time (Rs / min)
Tp = average production time per piece (min / pc)
Ct = cost of tooling per piece (Rs / pc)
Co = the allocation of capital cost of the equipment over the service life, labour to
operate the line, applicable overheads, maintenance, & other relevant costs all reduced to
cost per min.

Problem on Transfer line performance:

A 30 station Transfer line is being proposed to machine a certain component currently

produced by conventional methods. The proposal received from the machine tool builder
states that the line will operate at a production rate of 100 pc / hr at 100% efficiency. From a
similar transfer line it is estimated that breakdowns of all types will occur at a frequency of
F = 0.20 breakdowns per cycle & that the average downtime per line stop will be 8.0
minutes. The starting blank that is machined on the line costs Rs. 5.00 per part. The line
operates at a cost for 100 parts each & the average cost per tool = Rs. 20 per cutting edge.
Compute the following:
1. Production rate
2. Line efficiency
3. Cost per unit piece produced on the line

Solution:

1. At 100% efficiency, the line produces 100 pc/hr. The reciprocal gives the unit time
or ideal cycle time per piece.

Tc = 1 = 0.010hr / pc = 0.6 mins

100

The average production time per piece is given by:

3
www.getmyuni.com

Tp = Tc + FTd
= 0.60 + 0.20 (8.0)
= 0.60 + 1.60
= 2.2 mins / piece
Rp = 1 / 2.2m = 0.45 pc / min = 27 pc / hr

Efficiency is the ratio of the ideal cycle time to actual production time

E = 0.6 / 2.2
= 27 %

Tooling cost per piece

Ct = (30 tools) (Rs 20 / tool)

100 parts
= Rs. 6 / piece

The hourly ratio of Rs 100 / hr to operate the line is equivalent to Rs. 1.66 / min.

Cpc = 5 + 1.66 (2.2) + 6

= 5 + 3.65 + 6
= Rs 14.65 / piece

Upper Bound Approach:

The upper bound approach provides an upper limit on the frequency on the line stops per
cycle. In this approach we assume that the part remains on the line for further processing. It
is possible that there will be more than one line stop associated with a given part during its
sequence of processing operations. Let
Pr = probability or frequency of a failure at station i where i = 1, 2,………. η
Station i where i = 1, 2, ……………. η

Since a part is not removed from the line when a station jam occurs it is possible that the
part will be associated with a station breakdown at every station. The expected number of
lines stops per part passing through the line is obtained by summing the frequencies Pi over
the n stations. Since each of the n stations is processing a part of each cycle, then the
expected frequency of line stops per cycle is equal to the expected frequency of line stops
per part i.e.
η
F = ∑ Pi ----------------------------- 8
i=1
where F = expected frequency of line stops per cycle
Pi = frequency of station break down per cycle, causing a line stop
η = number of workstations on the line
If all the Pi are assumed equal, which is unlikely but useful for computation purposes, then

F = η.p where all the Pi are equal ---------------- 9

p = p = ………….. p = p
1 2 η

4
www.getmyuni.com

Lower Bound Approach:

The lower bound approach gives an estimate of the lower limit on the expected frequency of
line stops per cycle. Here we assume that a station breakdown results in destruction of the
part, resulting in removal of the part from the line & preventing its subsequent processing at
the remaining workstations.

Let Pi = the probability that the workpiece will jam at a particular station i.

Then considering a given part as it proceeds through the line, Pi = probability that the part
will jam at station 1
(1 - Pi) = probability that the part will not jam station 1 & thus will available
for processing at subsequent stations. A jam at station 2 is contingent on successfully
making it through station 1 & therefore the probability that the same part will jam at station
2 is given by
P (1 – P )
1 2

Generalising the quantity

P (1 – Pi - 1) (1 – Pi - 2) = (1 – P2) (1 – P1)
Where i = 1,,2, ………. η
is the probability that a given part will jam at any station i. Summing all these probabilities
from i = 1 through i = η gives the probability or frequency of line stops per cycle.

Probability that the given part will pass through all η stations without a line stop is
η
π (1 - Pi)
i=1

Therefore the frequency of line stops per cycle is:

η
F = 1 – π (1 - Pi) --------------------------------- 10
i=1
If all the probabilities, Pi, are equal, Pi = P, then
η
F = 1 - (1 - P)

Because of parts removal in the lower bound approach, the number of parts coming of the
line is less than the number launched onto the front of the line.
If F= frequency of line stops & a part is removed for every line stop, then the proportion of
parts produced is (1 - F). This is the yield of the production line. The production rate
equation then becomes:

Rap = 1 – F --------------------------------------- 11
Tp
where Rap = average actual production rate of acceptable parts from the line
Tp = average cycle rate of the transfer machine
Rp = 1 = average cycle rate of the system
Tp

5
www.getmyuni.com

Example 2 Upper Bound v/s Lower Bound Approach

A 2 station transfer line has an ideal cycle time of Tc = 1.2 mins. The probability of station
breakdown per cycle is equal for all stations & P = 0.005 breakdowns / cycle. For each of
the upper bound & lower bound determine:
a) frequency of line stops per cycle
b) average actual production rate
c) line efficiency

a) For the Upper bound approach

F = 20 (0.005) = 0.10 lines per cycle
20 20
F = 1 - (1 – 0.005) = 1 – (0.995)
= 1 – 0.0946
= 0.0954 line stops per cycle
For the Upper bound approach the production rate,

Rp = 1
20
= 0.500 pc / min
= 30 pc /hr
For the lower bound approach the production time we calculate by using the formula for F
Tp = Tc + F (Td)
= 1.2 + 0.0954 (0.8)
= 1.9631 mins

Production rate = 0.9046

1.9631
= 0.4608 pc / min
= 27.65 pc / hr
The production rate is about 8% lower than that we computed by the upper bound approach.
We should note that:
Rp = 1
0.9631
= 0.5094 cycles / min
= 30.56 cycles / hr
which is slightly higher than in the upper bound case.

c) For the upper bound the line efficiency will be

E = 1.2
2.0
= 0.6
= 60 %
For the lower bound approach we have

E = 1.2
1.9631
= 0.6113
= 61.13 %

6
www.getmyuni.com

Line efficiency is greater with lower bound approach even though production rate is lower.
This is because lower bound approach leaves fewer parts remaining on the line to jam.

Analysis of Transfer Lines with Storage Buffers:

In an automated production line with no internal storage of parts, the workstations

are interdependent. When one station breaks down all other stations on the line are affected
either immediately or by the end of a few cycles of operation. The other stations will be
forced to stop for one or two reasons 1) starving of stations 2) Blocking of stations
Starving on an automated production line means that a workstation is prevented from
performing its cycle because it has no part to work on. When a breakdown occurs at any
workstation on the line, the stations downstream from the affected station will either
immediately or eventually become starved for parts.

Blocking means that a station is prevented from performing its work cycle because it
cannot pass the part it just completed to the neighbouring downstream station. When a break
down occurs at a station on the line, the stations upstreams from the affected station become
blocked because the broken down station cannot accept the next part for processing from the
neighbouring upstream station. Therefore none of the upstream stations can pass their just
completed parts for work.

By Adding one or more parts storage buffers between workstations production lines
can be designed to operate more efficiently. The storage buffer divides the line into stages
that can operate independently for a number of cycles.
The number depending on the storage capacity of the buffer
If one storage buffer is used, the line is divided into two stages.
If two storage buffers are used at two different locations along the line, then a three
stage line is formed.
The upper limit on the number of storage buffers is to have a storage between every pair of
adjacent stations.
The number of stages will then be equal to the number of workstations.
For an η stage line, there will be η – 1 storage buffers. This obviously will not include the
raw parts inventory at the front of the line or the finished parts inventory that accumulates at
the end of the line.
Consider a two – stage transfer line, with a storage buffer separating the stages. If we
assume that the storage buffer is half full. If the first stage breaks down, the second stage
can continue to operate using parts that are in the buffer. And if the second stage breaks
down, the first stage can continue to operate because it has the buffer to receive its output.
The reasoning for a two stage line can be extended to production lines with more than two
stages.

Limit of Storage Buffer Effectiveness:

Two extreme cases of storage buffer effectiveness can be identified:

1. No buffer storage capacity at all.
2. Infinite capacity storage buffers
If we assume in our Analysis that the ideal cycle time Tc is the same for all stages
considered.

7
www.getmyuni.com

In the case of no storage capacity, the production line acts as one stage when a station
breaks down the entire line stops. This is the case of a production line with no internal
storage.
The line efficiency of a zero capacity storage buffer:

Eo = Tc --------------------------------- 12
Tc + FTd
The opposite extreme is the case where buffer zones of infinite capacity are installed
between every pair of stages. If we assume that each storage buffer is half full, then each
stage is independent of the next. The presence of the internal storage buffer means that then
no stage will ever be blocked or starved because of a breakdown at some other stage.
An infinite capacity storage buffer cannot be realized in practice. If it could then the
overall line efficiency will be limited by the bottleneck stage.
i.e. production in all other stages would ultimately be restricted by the slowest stage. The
downstream stages could only process parts at the output rate of the bottleneck stage.
Given that the cycle time Tc is the same for all the stages the efficiency for any stage
k is given by:

Ek = Tc
Tc + F Td
k k

where k is used to identify the stage.

The overall line efficiency would be given by:

E = Minimum (E )
∞ k

where the subscript ∞ identifies E as the efficiency of a line whose storage buffers have
∞
infinite capacity.
By including one or more storage buffers in an automated production line, we expect to
improve the line efficiency above E , but we cannot expect to achieve E .
0 ∞
The actual value of line efficiency will fall somewhere between these extremes for a given
buffer capacity

E <E <E
0 b ∞
Analysis of a Two stage transfer line:

The two stage line is divided by a storage buffer of capacity is expressed in terms of
the number of work parts that it can store. The buffer receives the output of stage 1 &
forwards it to stage 2, temporarily storing any parts not immediately needed by stage 2 upto
its capacity b. The ideal cycle time Tc is the same for both stages. We assume the downtime
distributions of each stage to be the same with mean downtime = Td, let F 1 & F2
be the breakdown rates of stages 1 & 2 respectively.
F1 & F2 are not necessarily equal.

8
www.getmyuni.com

Over the long run both stages must have equal efficiencies. If the efficiency of stage 1 is
greater than the efficiency of stage 2 then inventory would build up on the storage buffer
until its capacity is reached.
Thereafter stage 1 would eventually be blocked when it outproduced stage 2.
Similarly if the efficiency of stage 2 is greater than the efficiency of stage 1 the inventory
would get depleted thus stage 2 would be starved.
Accordingly the efficiencies would tend to equalize overtime in the two stages.
The overall efficiency for the two stage line can be expressed as:

1
E = E + { D η (b) } E ------------------------- 13
b 0 1 2

where Eb = overall efficiency for a two stage line with a buffer capacity

E = line efficiency for the same line with no internal storage buffer
0
1
{ D η (b) } E represents the improvement in efficiency that results from having a
1
storage buffer with b > 0
when b = 0

E = Tc ----------------------------------14
0 Tc + (F + F ) Td
1 2
1
The term D can be thought of as the proportion of total time that stage 1 is down
1

1
D = F Td
1 1 ----------------------------------------- 15
Tc + (F + F ) Td
1 2

The term h (b) is the proportion of the downtime D'1 (when the stage 1 is down) that stage 2
could be up & operating within the limits of storage buffer capacity b. The equations cover
several different downtime distributions based on the assumption that both stages are never
down at the same time. Four of these equations are presented below:

Assumptions & definitions: Assume that the two stages have equal downtime distributions
(Td1 = Td2 = Td) &
equal cycle times (Tc1 = Tc2 = Tc).
Let F1 = downtime frequency for stage 1, & F2 = downtime frequency for stage 2. Define r
to be the ration of breakdown frequencies as follows:

r = F1 ------------------- 16 F2

Equations for h(b) :

With these definitions & assumptions, we can express the relationships for h(b)for two
theoretical downtime distributions :

9
www.getmyuni.com

Constant downtime:
Each downtime occurrence is assumed to be of constant duration Td. this is a case of no
downtime variation. Given buffer capacity b, define B & L as follows:

b = B Td + L --------------- 17
Tc

Where B is the largest integer satisfying the relation : b Tc ≥ B,

Td
& L represents the leftover units, the amount by which b exceeds B Td .
Tc
There are two cases:

Case 1: r=1.0.h(b)
= B + LTc 1 --------18
B+1 Td (B+1)(B+2)

Case 2: r≠1.0.h(b)
=r1 - rB+L Tc rB(1 - r)2 -----19
Td (1 – r ) (1 – rB+2)
B+1

Geometric downtime distribution:

In this downtime distribution, the probability that repairs are completed during cycle
duration Tc, is independent of the time since repairs began. This a case of maximum
downtime variation. There are two cases:
Case 1: r = 1.0.h(b)
B Tc
= Td ------------------- 20
2 + (b – 1) Tc
Td
Case 2: r ≠ 1.0.
1 + r - Tc
Define K = Td --------------- 21
1 + r – r Tc
Td

Then h(b) = r (1 - Kb) --------------22 1-

rKb

Finally, E2 corrects for the assumption in the calculation of h (b) that both stages are
never down at the same time. This assumption is unrealistic. What is more realistic is that
when stage 1 is down but stage 2 could be producing because of parts stored in the buffer,
there will be times when stage 2 itself breaks down. Therefore E2 provides an estimate of
the proportion of stage 2 uptime when it could be otherwise be operating even with stage 1
being down. E2 is calculated as:

E2 = Tc --------------- 23 Tc + F2Td

10
www.getmyuni.com

Two-Stage Automated Production Line:

A 20-station transfer line is divided into two stages of 10 stations each. The ideal cycle time
of each stage is Tc = 1.2 min. All of the stations in the line have the same probability of
stopping, p = 0.005. We assume that the downtime is constant when a breakdown occurs, Td
= 8.0 min. Using the upper-bound approach, compute the line efficiency for the following
buffer capacities: (a) b = 0, (b) b = ∞, (c) b = 10, (d) b = 100

Solution:

F = np = 20(0.005) = 0.10

E0 = 1.2 = 0.60
1.2 + 0.1(8)

(a) For a two stage line with 20 stations (each stage = 10 stations) & b = ∞, we first
compute F:

F1 = F2 = 10(0.005) = 0.05

E∞ = E1 = E2 = 1.2 = 0.75
1.2 + 0.05(8)

(b) For a two stage line with b = 10, we must determine each of the items in equation 13.
We have E0 from part (a). E0 = 0.60. And we have E2 from part (b). E2 = 0.75

D'1 = 0.05 (8) = 0.40 = 0.20

1.2 + (0.05 + 0.05) (8) 2.0

Evaluation of h(b) is from equation 18 for a constant repair distribution. In equation 17, the
ratio

Td = 8.0 = 6.667.
Tc 1.2
For b = 10, B = 1 & L = 3.333.

Thus,
h(b) = h(10)
= 1 + 3.333 (1.2) 1
1+1 (8.0) (1 + 1)(1 + 2)
= 0.50 + 0.8333 = 0.5833
We can now use equation 13:

E10 = 0.600 + 0.20 (0.5833) (0.75)

= 0.600 + 0.0875 = 0.6875

11
www.getmyuni.com

(c) For b = 100, the only parameter in equation 13 that is different from part (c) is h(b).
for b = 100, B = 15 & L = 0 in equation 18. Thus, we have:

h(b ) = h(100) = 15
15 + 1
= 0.9375
Using this value,

E100 = 0.600 + 0.20 (0.9375) (0.75)

= 0.600 + 0.1406 = 0.7406

The value of h(b) not only serves its role in equation 13 but also provides information on
how much improvement in efficiency we get from any given value of b. note in example 15
that the difference between E∞ & E0 = 0.75 – 0.60 = 0.15.
For b = 10, h(b) = h(10) = 0.58333, which means we get 58.33% of the maximum possible
improvement in line efficiency using a buffer capacity of
10 {E10 = 0.6875 = 0.60 + 0.5833)(0.75 – 0.60)}.
For b = 100, h(b) = h(100) = 0.9375, which means we get 93.75% of the maximum
improvement with b = 100 {E100 = 0.7406 = 0.60 + 0.9375 (0.75 – 0.60)}
We are not only interested in the line efficiencies of a two stage production line. We
also want to know the corresponding production rates. These can be evaluated based on
knowledge of the ideal cycle time Tc & the definition of line efficiency. According to
equation 5, E = Tc / Tp. Since Rp = the reciprocal of Tp, then E = TcRp. Rearranging this we
have:

Rp = E ------------------- 24
Tc

Production Rates on the Two-Stage Line of the example above:

Compute the production rates for the 4 cases in the above example. The value of
Tc = 1.2 min is as before.

Solution:

(a) For b = 0, E0 = 0.60. Applying equation 23, we have

Rp = 0.60/ 1.2 = 0.5 pc/min = 30 pc /hr.

(b) For b = ∞, E∞ = 0.75.

Rp = 0.75 / 1.2 = 0.625 pc / min = 37.5 pc /hr

(c) For b = 10, E10 = 0.6875.

Rp = 0.6875 / 1.2 = 0.5729 pc / min = 34.375 pc /hr.

(d) For b = 100, E100 = 0.7406

12
www.getmyuni.com

Rp = 0.7406 / 1.2 = 0.6172 pc / min = 37.03 pc / hr

Effect of High Variability in Downtimes:

Evaluate the line efficiencies for the two-stage line in above example, except that the
geometric repair distribution is used instead of the constant downtime distribution.

Solution:
For parts (a) & (b), the values of E0 & E∞ will be the same as in the previous
example. E0 = 0.600 & E∞ = 0.750.

(c) For b = 10, all of the parameters in equation 13 remain the same except h(b).

Using equation 20, we have:

h(b) = h(10) = 10(1.2/8.0) = = 0.4478

2 + (10 - 1)(1.2/8.0)

Now using equation 13, we have

E10 = 0.600 + 0.20 (0.4478)(0.75)

= 0.6672

(d)For b = 100, it will be:

h(b) = h(100) = 100(1.2/8.0) = 0.8902

2 + (100 - 1)(1.2/8.0)

E100 = 0.600 + 0.20 (0.8902)(0.75)

= 0.7333

Transfer Lines with More thanTwo Stages:

If the line efficiency of an automated production line can be increased by dividing it into
two stages with a storage buffer between, then one might infer that further improvements in
performance can be achieved by adding additional storage buffers. Although we do not
exact formulas for computing line efficiencies for the general case of any capacity b for
multiple storage buffers, efficiency improvements can readily be determined for the case of
infinite buffer capacity.

Transfer Lines with more than One Storage Buffer:

For the same 20-station transfer line we have been considering in the previous examples,
compare the line efficiencies & production rates for the following cases, where in each case
the buffer capacity is infinite: (a) no storage buffers, (b) one buffer, (c) three buffers, &

13
www.getmyuni.com

19 buffers. Assume in cases (b) & (c) that the buffers are located in the line to equalise the
downtime frequencies; i.e. all Fi are equal. As before, the computations are based on the
upper-bound approach.

Solution:
(a) For the case of no storage buffer, E∞ = 0.60

Rp = 0.60/1.2 = 0.50 pc/min = 30 pc/hr

(b) For the case of one storage buffer

(a two stage line), E∞ = 0.75

Rp = 0.75/1.2 = 0.625 pc/min = 37.5 pc/hr

(c) For the case of three storage buffers (a four stage line), we have

F1 = F2 = F3 = F4 = 5(0.005) = 0.025

Tp = 1.2 + 0.025(8) = 1.4 min / pc.

E∞ = 1.2 / 1.4 = 0.8571

Rp = 0.8571/1.2 = 0.7143 pc/min

= 42.86 pc/hr.

(d) For the case of 19 storage buffers (a 20 stage line, where each stage is one station), we
have

F1 = F2 = ….. = F20 = 1(0.005) = 0.005

Tp = 1.2 + 0.005(8) = 1.24 min / pc.

E∞ = 1.2 / 1.24 = 0.9677

Rp = 0.9677/1.2 = 0.8065 pc/min

= 48.39 pc/hr.

This last value is very close to the ideal production rate of Rc = 50 pc/hr

Problem:

Suppose that a 10 station transfer machine is under consideration to produce a component

used in a pump. The item is currently produced by mass conventional means but demand for
the item cannot be met. The manufacturing engineering department has estimated that the
ideal cycle time will be
Tc = 1.0 min. From similar transfer lines & that the average downtime for line stop occur
with a frequency;

14
www.getmyuni.com

F = 0.10 breakdown/cycle & the average downtime per line stop will be 6.0 min. The scrap
rate for the current conventional processing method is 5% & this is considered a good
estimate for a transfer line. The starting costing for the component costs Rs. 1.50 each & it
will cost Rs 60.00 / hr or Rs 1 / min to operate the transfer line. Cutting tools are estimated
to cost Rs 0.15/ work part. Compute the following measures of line performance given the
foregoing data.

(a) Production rate

(b) Number of hours required to meet a demand of 1500 units/week.
(c) Line efficiency
(d) Cost per unit produced.

Problem:

If a line has 20 work stations each with a probability of breakdown of 0.02, the cycle time of
the line is 1 min & each time a breakdown occurs, it takes exactly 5 minutes to repair. The
line is to be divided into two stages by a storage buffer so that each stage will consist of 10
stations. Compute the efficiency of the two stage line for various buffer capacities.

Solution:

Let us compute the efficiency of the line with no buffer

F = np = 20(0.02) =0.4
Eo = 1.0 = 0.20
1.0+0.4(10)

Next dividing the line into equal stages by a buffer zone of infinite capacity each stage
would have an efficiency given by
F1 = F2 = 10 (0.02) = 0.2
E1 = E2 = Tc = 1.0 = 0.333
Tc + (F1+F2)Td 1.0+ 0.2(10)

d) The cost per product can be computed except that we must account for the scrap rate.
Cpc = 1 (1.50+1.00 x 1.60 + 0.15) = Rs.3.42/good unit
0.95

The Rs.3.42 represents the average cost per acceptable product under the assumption that
we are discarding the 5% bad units with no salvage value and no disposal cost. Suppose that
we could repair these parts at a cost of Rs.5.00/unit. To compute the cost per piece the repair
cost would be added to other components.
Cpc = 1.50 + 1.00 x 1.60 + 0.15 + 0.05 (5.00) = Rs. 3.50/unit.
The policy of scrapping the 5% defects ,yields a lower cost per unit rather than repairing
them.

15
www.getmyuni.com

Problem:

An eight station rotary indexing machine operates with an ideal cycle time of 20 secs. The
frequency of line stop occurrences is 0.06 stop / cycle on the average. When a stop occurs it
takes an average of 3 min to make repairs. Determine the following:

1. Average production time 3. Line efficiency

2. Proportion of downtime 4. Average production rate
Solution
Tp = Tc + F ( Td)
= 0.33 + 0.06(3)
= 0.5133 minutes.
Rp = 1 = 1.94 pieces /minutes
0.5133
Line efficiency = Tc = 0.333 = 0.491
Tp 0.51
Proportion of downtime can be calculated by D = F Td = 0.06(3) = 0.35
Tp 0.5133

Partial Automation:

Many assembly lines in industry contain a combination of automated & manual work
stations. These cases of partially automated production lines occur for two main reasons:

1. Automation is introduced gradually on an existing manual line. Suppose that

demand for the product made on a manually operated line increases, & it is desired
to increase production & reduce labour costs by automating some or all of the
stations. The simpler operations are automated first, & the transition toward a fully
automated line is accomplished over a long period of time. Meanwhile, the line
operates as a partially automated system.
2. Certain manual operations are too difficult or too costly to automate. Therefore,
when the sequence of workstations is planned for the line, certain stations are
designed to be automated, whereas the others are designed as manual stations.

Examples of operations that might be too difficult to automate are assembly

procedures or processing steps involving alignment, adjustment, or fine-tuning of the
work unit. These operations often require special human skills and/or senses to carry
out. Many inspection procedures also fall into this category. Defects in a product or a
part that can be easily perceived by a human inspector are sometimes extremely
difficult to identify by an automated inspection device. Another problem is that the
automated inspection device can only check for the defects for which it was
designed, whereas a human inspector is capable of sensing a variety of unanticipated
imperfections & problems.

To analyze the performance of a partially automated production line, we

build on our previous analysis & make the following assumptions:
1. Workstations perform either processing or assembly operations;

16
www.getmyuni.com

2. Processing & assembly times at automated stations are constant, though not
necessarily equal at all stations;
3. Synchronous transfer of parts;
4. No internal buffer storage ;
5. The upper bound approach is applicable &
6. Station breakdowns occur only at automated stations.
Breakdowns do not occur at manual workstations because the human workers are
flexible enough, we assume, to adapt to the kinds of disruptions & malfunctions that
would interrupt the operation of an automated workstation. For example, if a human
operator were to retrieve a defective part from the parts bin at the station, the part
would immediately be discarded & replaced by another without much lost time. Of
course, this assumption of human adaptability is not always correct, but our analysis
is based on it.
The ideal cycle time Tc is determined by the slowest stations on the line,
which is generally one of the manual stations. If the cycle time is in fact determined
by a manual station, then Tc will exhibit a certain degree of variability simply
because there is a random variation in any repetitive human activity. However, we
assume that the average Tc remains constant over time. Given our assumption that
breakdowns occur only at automated stations, let na = the number of automated
stations & Td = average downtime per occurrence. For the automated stations that
perform processing operations, let pi = the probability (frequency) of breakdowns
per cycle; & for automated stations that perform assembly operations, let qi & mi
equal, respectively, the defect rate & probability that the defect will cause station i to
stop. We are now in a position to define the average actual production time:

Tp = Tc + ∑ piTd --------- 25
iєna
where the summation applies to the na automated stations only. For those automated
stations that perform assembly operations in which a part is added,

pi = miqi

If all pi, mi, & qi are equal, respectively to p, m, & q, then the preceding equations
reduce to the following:

Tp = Tc +napTd ---------- 26

and p = mq for those stations that perform assembly consisting of the addition of a
part.
Given that na is the number of automated stations, then nw = the number of
stations operated by manual workers, & na + nw = n, where n = the total station
count. Let Casi = cost to operate the automatic workstation i ($ / min), Cwi = cost to
operated manual workstation
i ($ / min), Cat = cost to operate the automatic transfer mechanism. Then the total
cost to operate the line is given by:

Co = Cat + ∑Casi + ∑Cwi -------- 27

iєna iєnw

17
www.getmyuni.com

where Co = cost or operating the partially automated production system ($ / min).

For all Casi = Cas & all
Cwi = Cw, then
Co = Cat + naCas + nwCw -------- 28

Now the total cost per unit produced on the line can be calculated as follows:

Cpc = Cm + CoTp + Ct ---------- 29

Pap

Where Cpc = cost per good assembly ($ / pc), Cm = cost of materials & components
being processed & assembled on the line ($ / pc),
Co = cost of operating the partially automated production system by either of the
equations 27 or 28
($ / min), Tp = average actual production time (min / pc), Ct = any cost of disposable
tooling ($ / pc), & Pap = proportion of good assemblies.

Problem On Partial Automation:

It has been proposed to replace one of the current manual workstations with
an automatic work head on a ten-station production line. The current line has six
automatic stations & four manual stations. Current cycle time is 30 sec. The limiting
process time is at the manual station that is proposed for replacement. Implementing
the proposal would allow the cycle time to be reduced to 24 sec. The new station
would cost $0.20/min. Other cost data: Cw = $0.15/min,
Cas = $0.10/min, & Cat = $0.12/min. Breakdowns occur at each automated station
with a probability p = 0.01. The new automated station is expected to have the same
frequency of breakdowns. Average downtime per occurrence Td = 3.0min, which
will be unaffected by the new station. Material costs & tooling costs will be
neglected in the analysis. It is desired to compare the current line with the proposed
change on the basis of production rate & cost per piece. Assume a yield of 100%
good product.

Solution:

For the current line,

Tc = 30 sec = 0.50min.
Tp = 0.50 + 6(0.01)(3.0) = 0.68 min.
Rp = 1/0.68 = 1.47 pc/min = 88.2pc/hr
Co = 0.12 + 4(0.15) + 6 (0.10)
= $1.32 / min
Cpc = 1.32 (0.68) = $0.898 / pc

For the proposed line,

Tc = 24 sec = 0.4 min.
Rp = 1/0.61 = 1.64 pc/min = 98.4pc/hr
Co = 0.12 + 3(0.15)+ 6(0.10)+ 1(0.20)

18
www.getmyuni.com

= $1.37/min
Cpc = 1.67(0.61) = $0.836 / pc

Even though the line would be more expensive to operate per unit time, the proposed
chage would increase production rate & reduced piece cost.

Storage Buffers:

The preceding analysis assumes no buffer storage between stations. When

the automated portion of the line breaks down, the manual stations must also stop for
lack of work parts (either due to starving or blocking, depending on where the
manual stations are located relative to the automated stations). Performance would
be improved if the manual stations could continue to operate even when the
automated stations stop for a temporary downtime incident. Storage buffers located
before & after the manual stations would reduce forced downtime at these stations.

Problem On Storage Buffers on a Partially Automated Line:

Considering the current line in the above example, suppose that the ideal cycle time
for the automated stations on the current line Tc = 18 sec. The longest manual time is
30 sec. Under the method of operation assumed in the above example both manual &
automated stations are out of action when a breakdown occurs at an automated
station. Suppose that storage buffers could be provided for each operator to insulate
them from breakdowns at automated stations. What effect would this have on
production rate & cost per piece?

Solution:

Given Tc = 18sec = 0.3min, the average actual production time on the automated
stations is computed as follows:

Tp = 0.30 + 6(0.01)(3.0) = 0.48min

Since this is less than the longest manual time of 0.50, the manual operation could
work independently of the automated stations if storage buffers of sufficient capacity
were placed before & after each manual station. Thus, the limiting cycle time on the
line would be
Tc = 30sec = 0.50 min, & the corresponding production rate would be:

Rp = Rc = 1/0.50 = 2.0pc/min
= 120.0 pc/hr

Using the line operating cost from the previous example, Co = $1.32/min, we have a
piece cost of

Cpc = 1.32 (0.50) = $0.66 / pc

Comparing with the previous example, we can see that a dramatic improvement in
production rate & unit cost is achieved through the use of storage buffers.

19
www.getmyuni.com

Problem On Partial Automation:

A partially automated production line has a mixture of three mechanized & three
manual workstations. There are a total of six stations, & the ideal cycle time Tc = 1.0
min, which includes a transfer time Tr = 6 sec. Data on the six stations are listed in
the following table. Cost of the transfer mechanism Cat = $0.10/min, cost to run each
automated station Cas = $ 0.12/min, & labour cost to operate each manual station
Cw = $ 0.17 /min. It has been proposed to substitute an automated station in place of
station 5. The cost of this station is estimated at
Cas5 = $ 0.25 / min, & its breakdown rate P5 = 0.02, but its process time would be
only 30 sec, thus reducing the overall cycle time of the line from 1.0 min to 36 sec.
Average downtime per breakdown of the current line as well as the proposed
configuration is Td = 3.5 min. Determine the following for the current line & the
proposed line: (a) production rate, (b) proportion uptime, & (c) cost per unit. Assume
the line operates without storage buffers, so when an automated station stops, the
whole line stops, including the manual stations. Also, in computing costs, neglect
material & tooling costs.

Station Type Process Time (sec) pi

1 Manual 36 0
2 Automatic 15 0.01
3 Automatic 20 0.02
4 Automatic 25 0.01
5 Manual 54 0
6 Manual 33 0

Solution : Tc = 1.0 min

Tp = 1.0 + 2(0.01)x3.5 + 1( 0.02) x 3.5 = 1.14 mins
Rp = 1 = 0.877 pcs /min x 60 = 52.65 pcs / hr
1.14
Cp = 0.12 + 3(0.17) + 3(0.10) = $ 0.93/mins
Cpc = 0.93x 1.14 = $1.062/piece
For the proposed line Tc = 36 secs = 0.6 mins
Tp = 0.6 + 2(0.01)3.5 + 2(0.02)3.5 = 0.81 mins
Rp = 1.234 pieces / min = 74.07 pieces/hr
Cp = 0.012 + 2(0.17) + 3(0.10) + 1(0.25) = $ 0.902/min
Cpc = 0.90 x 0.81 mins = $ 0.73062/piece

Transfer of Work between Work Stations:

There are two basic ways in which the work (the subassembly that is being built up) is
moved on the line between operator workstations.
1. Nonmechanical Lines. In this arrangement, no belt or conveyor is used to move the
parts between operator workstations. Instead, the parts are passed from station to
station by hand. Several problems result from this mode of operation:

Starving at stations, where the operator has completed his or her work
but must wait for parts from the preceding station.

20
www.getmyuni.com

Blocking of stations, where the operator has completed his or her

work but must wait for the next operator to finish the task before
passing along the part.

As a result of these problems, the flow of work on a nonmechanical line

is usually uneven. The cycle times vary, and this contributes to the
overall irregularity. Buffer stocks of parts between workstations are often
used to smooth out the production flow.
• Moving conveyor lines. These flow lines use a conveyor (e.g., a moving belt,
conveyor, chain-in-the-floor, etc.) to move the subassemblies between workstations.
The transport system can be continuous, intermittent (synchronous), or
asynchronous. Continuous transfer is most common in assembly lines, although
asynchronous transfer is becoming more popular. With the continuously moving
conveyor, the following problems can arise:

Starving can occur as with non mechanical lines.

Incomplete items are sometimes produced when the operator is
unable to finish the current part & the next part travels right by on the
conveyor. Blocking does not occur.

Again, buffer stocks are sometimes used to overcome these problems. Also stations
overlaps can sometimes be allowed, where the worker is permitted to travel beyond the
normal boundaries of the station in order to complete work.

In the moving belt line, it is possible to achieve a higher level of control over the
production rate of the line. This is accomplished by means of the feed rate, which refers to
the reciprocal of the time interval between work parts on the moving belt. Let fp denote this
feed rate. It is measured in work pieces per time & depends on two factors: the speed with
which the conveyor moves, & the spacing of work parts along the belt. Let Vc equal the
conveyor speed (feet per minute or meters per second) & sp equal the spacing between parts
on the moving conveyor (feet or meters per work piece). Then the feed rate is determined by
fp = Vc -------------- 1
Sp

To control the feed rate of the line, raw work parts are launched onto the line at
regular intervals. As the parts flow along the line, the operator has a certain time period
during which he or she begin work on each piece. Otherwise, the part will flow past the
station. This time period is called the tolerance time Tt. It is determined by the conveyor
speed & the length of the workstation. This length we will symbolize by Ls, & it is largely
determined by the operator’s reach at the workstation. The tolerance time is therefore
defined by
Tt = Ls ----------- 2
Vc

For example, suppose that the desired production rate on a manual flow line with
moving conveyor were 60 units/h. this would necessitate a feed rate of 1 part/min. This
could be achieved by a conveyor speed of 0.6m/min & a part spacing of 0.5m. (Other
combinations of Vc & sp would also provide the same feed rate.) If the length of each
workstation were 1.5m. the tolerance time available to the operators for each work piece

21
www.getmyuni.com

would be 3 min. It is generally desirable to make the tolerance time large to compensate for
worker process time variability.

Mode Variations:

In both nonmechanical lines & moving conveyor lines it is highly desirable to assign
work to the stations so as to equalize the process or assembly times at the workstations. The
problem is sometimes complicated by the fact that the same production line may be called
upon to process more than one type of product. This complication gives rise to the
identification of three flow line cases (and therefore three different types of line balancing
problems).
The three production situations on flow lines are defined according to the product or
products to be made on the line. Will the flow line be used exclusively to produce one
particular model? Or, will it be used to produce several different models, & if so how will
they be scheduled on line? There are three cases that can be defined in response to these
questions:

1. Single-model line. This is a specialized line dedicated to the production of a single

model or product. The demand rate for the product is great enough that the line is
devoted 100% of the time to the production of that product.
2. Batch-model line. This line is used for the production of two or more models. Each
model is produced in batches on the line. The models or products are usually
similar in the sense of requiring a similar sequence of processing or assembly
operations. It is for this reason that the same line can be used to produce the
various models.
3. Mixed-model lines. This line is also used for the production of two or more models,
but the various models are intermixed on the line so that several different models
are being produced simultaneously rather than in batches. Automobile & truck
assembly lines are examples of this case.
To gain a better perspective of the three cases, the reader might consider the
following. In the case of the batch-model line, if the batch sizes are very large, the batch-
model line approaches the case of the single-model line. If the batch sizes become very
small (approaching a batch size of 1), the batch-model line approximates to the case of the
mixed-model line.
In principle, the three cases can be applied in both manual flow lines & automated
flow lines. However, in practice, the flexibility of human operators makes the latter two
cases more feasible on the manual assembly line. It is anticipated that future automated lines
will incorporate quick changeover & programming capabilities within their designs to
permit the batch-model, & eventually the mixed-model, concepts to become practicable.
Achieving a balanced allocation of work load among stations of the line is a problem
in all three cases. The problem is least formidable for the single-model case. For the batch-
model line, the balancing problem becomes more difficult; & for the mixed-model case, the
problem of line balancing becomes quite complicated.
In this chapter we consider only the single-model line balancing problem, although
the same concepts & similar terminology & methodology apply for the batch & mixed
model cases.

22
www.getmyuni.com

The Line Balancing Problem:

In flow line production there are many separate & distinct processing & assembly
operations to be performed on the product. Invariably, the sequence of processing or
assembly steps is restricted, at least to some extent, in terms of the order in which the
operations can be carried out. For example, a threaded hole must be drilled before it can be
tapped. In mechanical fastening, the washer must be placed over the bolt before the nut can
be turned & tightened. These restrictions are called precedence constraints in the language
of line balancing. It is generally the case that the product must be manufactured at some
specified production rate in order to satisfy demand for the product. Whether we are
concerned with performing these processes & assembly operations on automatic machines
or manual flow lines, it is desirable to design the line so as to satisfy all of the foregoing
specifications as efficiently as possible.
The line balancing problem is to arrange the individual processing & assembly tasks
at the workstations so that the total time required at each workstation is approximately the
same. If the work elements can be grouped so that all the station times are exactly equal, we
have perfect balance on the line & we can expect the production to flow smoothly. In most
practical situations it is very difficult to achieve perfect balance. When the workstations
times are unequal, the slowest station determines the overall production rate of the line.
In order to discuss the terminology & relationships in line balancing, we shall refer
to the following example. Later, when discussing the various solution techniques, we shall
apply the techniques to this problem.

23