Inal Eport: Improvement Demand and Service Level
Inal Eport: Improvement Demand and Service Level
Inal Eport: Improvement Demand and Service Level
FINAL REPORT
Improvement Demand and Service Level
Fabre-Surinon
10
Final Report 2010
Sommaire
I. First mission .......................................................................................................................................... 3
A. Model of the system ......................................................................................................................... 3
B. Improve the production to reach 1100 units per week .................................................................... 7
C. Improve the production to reach 1250 units per week .................................................................... 7
II. Second mission ..................................................................................................................................... 9
A. Model of the demand ....................................................................................................................... 9
B. Results ............................................................................................................................................. 11
C. Improvement .................................................................................................................................. 12
ANNEXE
Assumption report
2
I. First mission
A. Model of the system
i. Flow scheme simplified
- Start the production of the bundle only if size of the finished bundles stock is below 60
units.
- Product the bundles in lot of 300 units, and send those once all are prepared
Create raw cable by 300 units, batch 300 units and then separate in 300 units
Once chips are created, they era part of a KANBAN system. To make a model of that we choose:
1°/ To create single chip, creation follow this expression: 40+UNIF (3, 37)
In order to model the dwarf’s KANBAN, we decided to create only the KANBAN Number and
to make a loop with a separate at the end of the product line. We also had to recalculate the
KANBAN Number (N=Demand*Loop_length/Capacity). So we calculate the length of the loop with a
box readwrite, which gives us the entering time and the living time of the system for each dwarf.
4
Final Report 2010
v. Model of the delay
The only constraint is to create one “delay” at the start of the simulation, so that the first dwarf is
able to go through the match box.
5
Final Report 2010
Average
After having analyzed the data we estimated the warm-up period at 3 hours.
Then we search the number of replication. To do so we made ten simulations, with a number
of replication from one to ten. We then calculated the average of the dwarf created.
Number of Number of
repilcation dwarf
created
20 1023,55
10 1024,1
9 1024,56
8 1023,88
7 1024,86
6 1025,5
5 1024,6
4 1025
3 1025,67
2 1022
1 1033
Average 1024,35
After having analyzed the data we estimated that a good number of replication was 10.
Finally our model is able to produce around 1024 units per week. We believe that the difference
between our model and the real system (average of production per week is 1040 for the real system)
is due to the fact that the real system is launched at its full capacity.
6
Final Report 2010
B. Improve the production to reach 1100 units per week
The first goal of the first mission was to increase the weekly production to 1100 units. So as the
demands arose, we first recalculate the number of KANBAN to fit with the new demand.
Thus we choose to add a resource. But it the production was still not enough (1213.6). As the
demand increase so does the work in each station. So in order to prevent any strike from the labor
force, we choose to add an operator in the assembly lantern station:
Delay between dwarf: 1min 38s and Conveyor belt speed: 0.75 km/h
Finally this last modification enables us to product more than 1250 dwarf per week.
7
Final Report 2010
Objectives Demands Increase N Ressource added Operator added in
Paintshop Assembly lantern
1100 27,5 1102,2
1250 31,25 1117,5 1213,6 1252,7
1040 26
8
Final Report 2010
9
Figure 4 Model of the demand section
F
i. Delayed demand
We create an entity which follows a normal law (1, 1/6). This entity is store in a
“hold” with the following condition: “hold the command while the second hold (c.f
below) is up to 0”. It permits us to know the numbers of delayed command. We also
placed a record in order to know the total number of demand.
10
F
B. Results
We use a read write, which write the number of waiting demand in the
queue of the first hold.
Total demand
delayed 16
Total demand per
replication 69,2
Number of
replication 10
Service level 97,7
Tableau 1First set of results
11
F
C. Improvement
We calculate a new cycle time of 5.75 hours so we calculate a new numbers
of Kanban of 180, and we obtained a service level of 100%.
Number of
demand delayed 0
Average number
of demand
delayed 1073,8
Number of
replication 10
Service level
100
Tableau 2 Second set of results
We are surprised of this value because we excepted some delayed commands but even after
a check of our model everything seems to be normal.
12