The theory of constraints states that every process has a bottleneck and that this bottleneck is the weak link in the process. Determining the weak spot or bottleneck is vital for any company since it influences the maximum throughput of the process. To determine the maximum throughput as a function of various components, Enterprise Dynamics was used to simulate the impact on this constraint in the process. Determining the impact made it possible to change the process to remove the constraint and increase throughput.

EPCOR is a subsidiary of Air France Industries & KLM Maintenance and Engineering and focuses on the repair and overhaul of Auxiliary Power Units and pneumatic components for various aircraft and companies. Pneumatic components that are repaired or overhauled are tested to determine the faults and are tested for final inspection when the components are repaired. Testing is performed on a test bench (UPV-1) which is specially designed for these components. EPCOR requested two students from the Hogeschool van Amsterdam, University of Applied Sciences to perform an analysis of the UPV-1 and determine whether the UPV-1 is currently the bottleneck and what actions should be taken to improve this.

EPCOR has a future prognosis for the amount of components for the pneumatic high flow line and would like to know if the test bench (UPV-1) which is used to test these components has enough capacity. EPCOR is also interested in the optimization of the planning process which feeds components into the UPV-1.

The UPV-1 is a revolver bench which is constructed by Hamilton Sundstrand (the manufacturer of most of the components which are tested) for these pneumatic high flow components. The UPV-1 is able to recreate the conditions in which these components operate when they are installed in the aircraft or on the engines. This includes the pressure, temperature and flow of the air through the component. The revolver construction is required to ensure that all the components can be tested on this machine because the diameter of the components vary. This means that changing to a different component the UPV-1 must be reconfigured for that specific component (diameter) which introduces a changeover or setup time. When the component is installed the test is performed which vary in type but also in length. The following tests can be performed:

  • Pre-test: Performed when components are received by EPCOR to determine the complaint.
  • Final-test: Performed when the components are repaired for testing and adjustment.
  • Pre-Final: Performed when no failures where found.

The time required for the test depends on the type of test and on the part number of the component. Different part numbers indicate a different type of components and thereby a different testing time. The testing time varies per part number but also per performed test (Pre/Final etc.).

The components which are tested on the UPV-1 are sorted and prioritized using the promised delivery date. The components are tested in the following order:

  • Final testing the components that need to be finished that day.
  • Pre-testing components to ensure a sufficient supply of components for the repair department.
  • Final testing the components in order of promised delivery date

The data which is required for the input and creation of the model was supplied from the ERP system EPCOR uses. This created a realistic situation ideal to determine the utilization of the UPV-1. The model was created only for that purpose and not, for example to determine the turnaround time of components. The model was thereafter validated for this specific purpose using the model validation paradigm designed by Robert G. Sargent. This improved the acceptance of the model for the people it would have influence on.


The model which is constructed in Enterprise Dynamics reconstructs the specific features mentioned earlier to match the real process. Components are generated using sources with labels to indicate the part number, test time and priority which is generated by an empirical distribution. The components are sorted per priority which enables the server (UPV-1) to select the parts with the highest priority first. The server uses the test time label to determine the testing time and the product type label to determine the setup time for each component. This entire process is controlled by availability controls to recreate the working hours of a month.

The analysis, construction and the simulation of the UPV-1 process using Enterprise Dynamics resulted in insight how the current process functions and how it will be able to cope with future changes. This enables EPCOR to make decisions based on concrete data which will strengthen their business case for future development.