An Auto Export Center Simulation is being used to restructure staffing, change line scheduling and determine future investments to meet the requirements of US auto exporters. The margin for vehicle export preparation is tight and requires great control of the processes to maintain a profit. Process simulation was used to find hidden dollars and prepare for export growth in the near future.
Cars such as the Ford Mustang are a huge hit in Europe, China and Australia, achieving best-seller status in some markets. The volume is in the thousands. Car companies use vehicle processing centers to prepare inventory for overseas markets. This allows the car manufacturer to standardize manufacturing and leave export details to a flexible outside group. Some of the possible final export details could include:
- Final Inspection
- Washing and detailing
- Installing protection for shipping
- Language specific decals and user’s manuals
- Accessory installation
- Staging for export
As we started to build the Auto Export Center Simulation, we learned many new things about what we thought was a simple process. To start off with, the combination of car models and existing markets means there are over 500 unique car combinations to be represented. For example, a car takes less time to undercoat than a truck. Every auto model has differences that alter the time to perform all process steps. The result of the above means that for every step in the process, each model in each market has different requirements and takes different times.
Differing times for each model/market was easy to represent by using attributes to carry the time values for each of the process steps. This was made easier using the Excel entry spreadsheet provided in ProcessModel Model objects library. We stored different model/market combinations on a separate back sheet. Each row of the back sheet is identified by a model (ex. Mustang) and a market (ex. China) and then the attributes that describe how long each process step takes. When we wanted to schedule a specific car we just copied the associated row from the back sheet with the car model and its pre-defined attributes. After inserting the row (or rows to schedule several different type of car in the same day) into the scheduling sheet, the update button is pressed and the arrivals and attributes for that car are read into the model.
We also learned there are some subtle rules about how some resources are required to finish an activity while other resources are allowed to stop in the middle of a task. For example, an undercoating resource needs to finish the spraying task before taking a break or leaving for the day. A person installing a smokers’ kit can walk away at any point during the step and finish after a break or the next day. A driver may be able to get out of the car at some activities and not at others, such as undercoating. The detailed rules were easy to set up with the built-in priorities in ProcessModel.
The production process is labor intensive, requiring many drivers and workers at each step of the process. The number of workers varies as the car model and markets change. A major problem is figuring out the correct staffing to get maximum throughput for the system. To compound the problem, the calculation is difficult because each step requires a variable amount of processing time. You can’t just put the numbers into a spreadsheet and make a quick and accurate calculation. Process simulation allowed the complex calculations to be automatically performed for each new scenario.
As we started the project we had three major questions that needed to be answered:
- What would be needed to increase production by 20% (over max production) using only the current single shift?
- How much overtime would be needed to achieve the 20% if no process or staffing changes were implemented?
- What are the bottlenecks of the current system?
Auto Export Center Simulation Provides Immediate Savings
The Auto Export Center Simulation was built, tested and initial results identified in less than 2 weeks. This was a huge benefit. Another vendor quoted many weeks of development and a huge price tag. With consulting from the ProcessModel team we were able to make a useful model quickly. One of the immediate benefits of using the model was the ability to see what was causing our existing bottlenecks.
With some simple resource reallocation, we increased production by 19% for the highest volume products (other products have increased by as much as 37%). This was a pleasant surprise from the Auto Export Center Simulation — we didn’t expect this. We had felt like the resources were reasonably balanced in real life. However, the model helped us to see the impact of moving work elements to other stations and the potential resulting increase in production. We implemented the change in real life and it worked just like the model had predicted. Our cost to get this 19% production increase was literally pennies.
The management team has now changed the goal from a 20% to a 50% increase in production. Two months ago I would not have thought an increase of 50% was possible. With what we have learned from the simulation, I am now confident we can achieve this goal and more.
The model is setup to optimize resources to match any schedule defined. Any schedule can be entered and the model will automatically readjust the quantity of resources needed at all of the process steps to maximize the number of cars processed in a day. A new schedule optimization with replications takes a couple of minutes to get the results. This is an enormous time saver in planning the daily resource assignments.
This was one of many facilities that can benefit from simulation. Hopefully, we will never make a process change again without using process simulation. The Auto Export Center Simulation has helped us to show management the benefit of process simulation and has provided confidence to make investments to increase production in the future.