Examples of manned single station cells include:
·
A CNC centre producing identical parts,
with an operator required to load and unload parts as the processing
programme is completed by the machine.
·
A CNC centre producing non-identical parts,
with an operator required to load new part programmes as necessary.
·
A cluster of two CNC turning centres, each
producing the same part, but operating independently. A single worker
attends to both machines.
·
An assembly workstation where an operator
performs mechanical assembly of a simple product (or sub-assembly) from
components located in tote bins at the station ENDLIST.
Examples of single-station automated cells include:
·
A CNC machining centre complete with parts
carousel and automatic pallet changer producing identical parts.
·
A CNC machining centre complete with parts
carousel and automatic pallet changer producing non-identical parts. In this
case the appropriate part programme is downloaded automatically as
necessary.
·
An automated insertion machine assembling
electronic components onto printed circuit boards in a batch operation.
·
A robotic assembly cell with one robot
assembling a simple product or subassembly.
Numerous examples of both manned and automated single-station cells can be
given, right across the manufacturing environmental spectrum.
Analysis of Single-Station Systems
In our analysis here we can determine:
·
The number of single stations required to
satisfy specified production requirements.
·
The number of machines to assign to
a worker in a machine cluster.
Number of Workstations Required
We must determine how many workstations are required, given a certain
production rate, or a given quantity of work units. This is generally done
by determining the total workload that must be accomplished over a certain
period, and dividing that by the hours available on one workstation during
the same period. Workload is determined thus:
where WL is the workload scheduled for a given period; Q is
the quantity to be produced during the same period; and Tc
is the cycle time required per piece. If the workload includes multiple part
or product styles that can all be produced on the same workstation, then:
where Qj is the quantity of part or product style j produced
during the period; and Tcj is the cycle time of part or
product style j.
We must now divide the result by the number of hours
available on one workstation, thus:
where n is the number of workstations; and AT is the available
time on one station in the period under consideration. These equations do
not take into account a number of potential complicating factors, which
makes it more difficult to assess the number of workstations required; these
include:
·
Set-up time in batch production.
·
Availability of machines.
·
Utilization of machines.
·
Defect rates from various machines.
Example:
800 parts are to be produced. Cycle time is
11.5min. Determine number of machines given 40hrs availability.
WL = 800(11.5) = 9,200min = 153.33hrs AT = 40hrs
n = 153.33/40 = 3.83 or 4 machines
Availability
time may be measured as follows, with the available time becoming the actual
shift time in the period multiplied by availability and utilization:
AT = HshAU
where AT is available time; Hsh
is the shift hours during the period; A is availability; and U
is utilization. The defect rate—that is, the fraction of parts produced that
are defective—must be assessed so that it can be factored-in to the starting
batch size, so that the output can compensate for defective parts produced.
Example:
800 shafts are in 20 different types. Average batch size is 40. Set-up time
between batches is 3.5hr.
WL = 20(3.5) + 20(40)(11.5/60)
= 70 + 153.33 = 223.33hrs n = 223.33/40 = 5.58 or 6 machines
The relationship between starting quantity and actual quantity produced is:
Q = Qo
(1 - q)
where Q is the quantity of good units made in
the process; Qo is the original or starting quantity; and
q is the fraction defect rate. This formula can be rearranged to give
us the amount of starting units we require, thus:
Taking these factors
into consideration, we can now amend our original formula, thus:
Example:
Using previous data and Availability is
100% during set-up and 92% during running. Utilisation is 100%. Fraction
defect rate is 5%. Determine number of machines.
For set-up:
WL = 20(3.5) = 70.0hrs AT = 40 (1.0)(1.0) =
40 nsu = 70/40 = 1.75 machines.
For production runs:
WL = ((20)(40)(11.5/60)) / (1-0.05) =
161.4hrs AT = 40(0.92) = 36.8hrs per machine npr = 161.4 / 36.8 =
4.39 machines Total Machines = 1.75 + 4.39 = 6.14 or 7 machines.
The number of
workstations that are required is determined by the total workload that must
be accomplished over a certain period, divided that by the hours available
on one workstation during the same period; together with a consideration of
any mitigating factors.
Analysis of Machine Clusters
Sometimes
opportunities exist to allow a worker to oversee more than one machine at a
time, owing to the semi-automatic machine cycle of individual machines.
Worker attention remains important; however not as regularly as every work
cycle; instead, the worker will be required on the basis of a manning level
of less than one for each workstation. This type of organisation is referred
to as a machine cell, or machine cluster, which is defined as a collection
of two or more machines producing parts or products with identical cycle
times and serviced by one worker. A machine cluster must satisfy several
conditions in order to exist:
·
The semi-automatic machine cycle must be
long relative to the service portion of the cycle that requires the worker’s
attention.
·
The semi-automatic machine cycle time must
be the same for all machines.
·
The machines that the worker would service
must be located in close enough proximity to allow time to walk between
them.
·
The work rules of the plant must permit a
worker to service more than one machine.
If we consider
a situation where we have a collection of single workstations, all producing
the same parts and operating under the conditions outlined above, then we
can say:
Each machine
operates for a certain portion of the total cycle under its own control Tm
(machine time), before requiring servicing from an operator Ts
(service time)
The total cycle time (Tc) of the machine is therefore
machine time plus service time; or, in equation form:
Tc = Tm + Ts
If we add
another machine to the operator’s purview, then the operator will lose some
time walking to this machine, called the repositioning time (Tr)
The total time
that an operator needs to service one machine must be adjusted from simply
Ts, to: Ts + Tr
We must also
factor-in the time to service n machines, which is: n(Ts
+ Tr)
Thus, the
original equation takes the form:
Further, we can
determine the number of machines that should be assigned to one worker by
solving:
Note that the
result for n is unlikely to be an integer, so worker time in the
cycle cannot be perfectly balanced with the cycle time of the machines. In
reality, either the machines or the workers will experience some idle time.
There arise two potential cases from this consideration:
Case 1: the
number of machines will be an integer greater than the value of n
(call this n1); or
Case 2: the
number of machines will be an integer less than the value of n (call
this n2).
Example:
Machine shop
contains number of machines with cycle time for one particular part of
2.75min. One worker can load and unload machines in 25sec. It takes 20sec to
walk between machines. How many machines can one worker manage if no machine
idle time is allowed.
Tm = 2.75min; Ts =
25sec = 0.4167min; Tr = 20sec=0.33min
The most
favourable case is dependent on other factors, such as the labour cost rate
(CL) and the machine cost rate (Cm). For case 1 (where
n1 < n) the worker will have idle time, and the cycle time
of the machine cluster will be:
If
we assume that one work unit is produced by each machine during a cycle, we
have:
For case 2
(where n2 > n) the machines will have idle time, and the
cycle time of the machine cluster will be the time it takes for the worker
to service the n2 machines, which is:
The corresponding cost per piece is given by:
The selection of either n1 or n2
is based upon whichever results in the lowest cost per work unit. In the
absence of precise cost data, the layout should be arranged so that any idle
time is taken by the workers, and not the machines. This is because the
total hourly cost rate f operating n machines is larger than the
labour rate of one worker. The corresponding number of machines to assign
the worker is given by:
A machine cluster is defined as a collection of two or more machines producing parts or products with identical cycle times and serviced by one worker. |