VSM
- Project name: VSM
- Class: 4IT495 Simulation of Systems (WS 2012/2013)
- Author: Jonas De Mets
- Model type: Discrete simulation
- Software used: SIMPROCESS
Contents
Introduction
The company TWI industries produces different components for tractors. In this study we only consider one product family “steering arms”. There is a wide variety of product specifications and configurations and these customer requirements are different from order to order. Therefore TWI industries uses a “make-to-order” system for its business.
Problem definition
The biggest problem for TWI industries is the long lead time. In order to reduce this, I am asked to make a simulation of the production process for the steering arms. Another problem is the occupied time of the employees. Some employees are overworked while other workers have a lot of idle time. Thirdly, to be as lean as possible the company needs to decrease inventories. The goal is to avoid as good as all inventories.
Method
Because my problem is a discrete simulation of different processes I will use Simprocess for simulating the problem. It’s a tool that’s easy to use and useful for visualizing following processes. Because I could not use a full version of Simprocess, there was a problem that occurred during my work on the project. It was impossible to add more delays and transforms, to solve this problem I had to make a simplification in the model. I decided to neglect the initial inventories before and after the processes. This simplification doesn’t have a great influence on the result of the model because it’s our goal to decrease the inventories to almost zero. Therefore balancing the processes is the most important target of the simulation. Another unimportant simplification is the neglecting of some operators because their work is linked to machine operations
Model
The sequence of processes is the following:
I will describe every step in the production process separately to explain the model.
General information about the model:
- Two shifts in all production departments
- 8 hours shifts
- Production from 8h – 24h
Receive rods
I will use a Poisson distribution that makes sure that the “CUT RODS” process has enough “RAW RODS” to cut, because I am only asked to look at the internal processes. But based on the conclusion I will also give a recommendation about inventories and ordering of “RAW RODS” in the conclusion.
Receive forgings
I will use a Poisson distribution that makes sure that the “MACHINING FORGINGS” process has enough “RAW FORGINGS” to operate on, because I am only asked to look at the internal processes. But based on the conclusion I will also give a recommendation about inventories and the ordering of “RAW FORGINGS” in the conclusion.
Cut rods
- Manual process with 1 operator
- Cycle Time: 15 seconds
- Reliability: 100%
We use a transform operation “CUT RODS” that needs 1 resource “Cut operator”. The entity “RAW ROD” is transformed in a “CUT ROD”. The resource works two shifts. Because the reliability is 100% we don’t need an extra branch for defaults.
Machining of Forgings
- Automatic process with one machine
- Cycle Time: 30 seconds
- Reliability: 100%
We use a transform operation “MACHINING OF FORGINGS” that needs 1 resource “Forging machine”. The entity “RAW FORGING” is transformed in a “MACHINED FORGING”. The resource works two shifts. Because the reliability is 100% we don’t need an extra branch for defaults. But we do need a branch because some of the “MACHINED FORGINGS” go to “WELD one end” and others go to “WELD other end”.
Weld one end
- This operation welds a machined forging to the rod
- Automatic process with one machine
- Cycle Time: 30 seconds
- Reliability: 90%
We use an assembly operation: “WELDING 1” to make a new entity “STEER ARM” from the two entering entities “CUTTED RODS” and “MACHINED FORGINGS”. The resource needed for this process is “Welding 1 machine”. Because the reliability is 90% we need an extra “DEFECTS WELD 1” branch for defaults. The 10% defects go out of the model through “DISPOSE DEFECTS WELD 1”.
Weld other end
- This operation welds the second machined forging to the rod
- Automatic process with one machine
- Cycle Time: 30 seconds
- Reliability: 80%
We use an assembly operation: “WELDING 2” to adjust the entity “STEER ARM” from the two entering entities “STEER ARM” and “MACHINED FORGINGS”. The resource needed for this process is “Welding 2 machine”. Because the reliability is 80% we need an extra “DEFECTS WELD 2” branch for defaults. The 20% defects go out of the model through “DISPOSE DEFECTS WELD 2”.
Deflash
- This operation removes the flash of the steer arm
- Automatic process with one machine
- Cycle Time: 30 seconds
- Reliability: 100%
We use a delay operation “DEFLASH” that needs 1 resource “Deflash machine”. The entity “STEER ARM” stays in the process for 30 seconds. The resource works two shifts. Because the reliability is 100% we don’t need an extra branch for defaults.
Painting
- Painting is done by an outside vendor
- Painting lead time: 2 days
End-Assembly
- Manual process with six operators
- Total Work Time Per Piece: 195 seconds divided by 6 operators.
- Average Cycle Time = 32,5 seconds
- Reliability: 100%
Shipping
- Shipping to customers
Results
When we run the model in the setup I explained earlier we can see several problems that occur in the process. This table shows the biggest problem in the simulation.
The “Cut operator” and “Forging machine” are always busy because of the huge inventory created by “RECEIVE ROD” and “RECEIVE FORGINGS”. But the other machines and operators are “Idle” a lot of the time. When we visualize the simulation and look at the data we can see the reasons for this problem.
1) First of all the “THE MACHINING OF FORGINGS” is too slow so it’s not possible to balance the system. The resource is 100% busy but the resources that follow in the other processes. Such as “Welding 1 machine”, “Welding 2 machine” and “Deflash machine” are idle for about 50% of the time. To solve this problem we should add extra activities in the “MACHINING OF FORGINGS” process in parallel. (It’s impossible to add extra transform operations so I divide the 30 seconds cycle time by the number of extra operations). The best solution is to use the “CUT RODS” as a pacemaker for the whole process. That means we need 2 “MACHINED FORGINGs” for every “CUT ROD”. This results in a theoretical desired cycle time of 7.5 seconds, this equals to 4 parallel processes.
2) Secondly when we use the “CUT RODS” as a pacemaker we need to adjust “WELD one end” and “WELD other end”. Their cycle time should be about 15 seconds. This means doubling the capacity. So install 2 machines for each process.
3) The process “DEFLASH” should be optimized but not doubled because only 72% of the “CUT RODS” reach this process, due to defaults. So small improvements here might be useful but big changes are not really necessary.
4) The “PAINTING” process makes the lead time much longer. The best solution is to do the painting in the own company. Because I don’t have any detailed data about the painting and drying time I will estimate the cycle time as 1 day instead of 2 days when the company does the painting itself.
5) The “END ASSEMBLY” is already quite balanced according to the input of the painting process.
The result of implementing this gives us the following data:
Conclusion
Short:
1) Implementing this solutions decreases the lead time by more than 1 day.
2) Balancing the processes leads to low Idle time
3) Balancing processes leads to almost no internal inventories
Explanation:
A good solution for the problem is to batch the first part of the internal process, and use the “CUT RODS” as a pacemaker. This balanced batch reduces inventories and lead time. The idle time is reduced to a minimum so there are no big losses. The second part of the solution in to include the painting to the own processes. The batches are handled by the painting department in for example 1 day. A good suggestion might be to do the painting or drying overnight or when the other processes are down. The ordering and receiving of inventories should be adjusted to the pacemaker, “CUT RODS” in order to avoid excessive inventories of raw material. This report and recommendations will strongly improve the overall efficiency of the company.
