System Background


Flexible manufacturing system is the philosophy of producing different part types within the same production system setting. Multiple parts of different processes can be produced at the simultaneously. This flexibility in manufacturing generates a complexity in term of prediction, system measuring, and performance that are affected by the variability of the parts feed into the production line. Even though the production line is automated using CNC machines, the material handling, setup, and teardown are still managed by operators.

Challenges


A company seeking help to find the most optimal and efficient process as possible. the production line produces four different parts brackets, Flanges, Cover plates, and Hex bearing broaching. Setup, process, and teardown operations are common between all parts but the sequential and number of processes is different among all.
The following table shows the parts and the processes.




Approach

To maximize the throughput and maintain best optimal stable system, a simulation model will be created represent the sequential processes in the above sequential table. The following diagram illustrate the sequential model that will be used to verify and optimize the system.



The model will have two decision variables, the number of the operator and the arrival rate and the stability of the system will be measured by the of departure part over arrival part. The system will have analyzed utilizing steady state approach.

Verification

Utilization will be calculated base on the system model diagram, the part arrivals and processes time. The data table bellow shows the processes rate, Arrival time, and percentage of mix in addition to the calculated utilization.



Expected Values based on 4 operators and 8 hour schedules




Simio Model

The following is the Simio model with five stations, four parts entities and a worker’s resource to allow operators to navigate between station to setup/teardown parts when needed based on FIFO system.





Verify Simio Model


Run a 10 replicant with 8000 hours and compare the value to the calculated utilization with one worker to test the model



Comparing the results with utilization table, we can determine that the model is working and able to produce a result that represent the system.

Data Tables


Creating schedule, part, and sequential table





Variable for optimization

Output statistis

Weekly Schedule

Experiment


We will replicate four scenarios with different number of workers and analyses the system output.



The results show that 1 and 4 workers in the system will produce and optimal performance and utilization. The Two workers in the system seems to produce the best solution to maximize throughput with minimum works.

Optimization


Let’s test the model using optQuest with different arrivals intervals and two to three workers to find most efficient solution.



Looking at the results of the optimization using a mean of 0.1 increment of arrival time shows that, arrivals of four per hour with three workers will provide and efficient and stable system.