Throughput and cost of a new flexible manufacturing system

Flexible manufacturing systems (FMS) can produce parts with low cost, small in-process-inventory, high quality, minimal non-value-added work, agility, consistency, and smooth production. When these objectives are achieved, FMSs are the best way of manufacturing parts. In order to start from a blank slate and design an effective FMS, many questions must be answered. Instead of attempting to produce a design all at once, we have developed a method which solves these questions one by one in sequence. When following our template, only a couple of related questions are investigated at once and any engineer can therefore design an FMS.

Stage 1: Project Goals

What is the projected order demand for one year? What are the estimated cycle times? The first step in designing a new system is to assemble an initial best guess for the goals for the system. This initial estimate becomes the starting point for sketching out the design and even if the guesses are inaccurate, part of the analysis will be investigating how the system handles variations in demand.

The next component of the project goals is the flow route, which is the sequence of workcenters that parts pass through to be produced. Each workcenter is either a labor step or a non-labor step. This sometimes means splitting what you might initially consider a workcenter in two. For example, a part might visit a manually-loaded machine. This should become three workcenters: a loading step where the operator loads the part into the machine, the machining step where the machine runs without the operator's attention, and an unload step where the operator unloads the part from the machine. The reason to split workcenters in this way is to better understand the labor requirements and the trade-offs between capital and labor: the load and unload workcenters can be studied and analyzed separately from the machining workcenter.

The final component of the project goals is the time that each part will spend at each workcenter. These numbers can initially be an approximation; as the project progresses these timings can be adjusted. For example, you might just initially estimate that the time the part spends at the loading workcenter is 5 minutes, but later measure the average loading time.

Stage 2: Manufacturing Strategy

How many machines to purchase? Should the system use an automated handling system, robots, manual labor, or a mix of all three? There are two challenges with selecting a manufacturing strategy: machines come in large chunks of capacity and the choice of handling system has a large impact on traffic jams and efficiency. It is easy using a spreadsheet and Little's Law to calculate the number of machines required to meet the part demand from Stage 1, but this assumes that machines operate at 100% efficiency. We have been involved on many systems over the decades and consistently see that in a flexible and lean process, manually operated machines operate at about 60% to 65% efficiency (that is, 60%-65% of the time the machines are cutting material). The remaining time is used for setups, loading, tool replacements, maintenance, labor breaks, and other tasks required to keep the machines operating properly. Therefore, we shouldn't view 65% efficiency as bad or the fault of the operator, but just the reality of manually-loaded machines. (The efficiency can be increased slightly with batching, but batching brings its own costs and prevents lean manufacturing.)

When designing a manually-operated flexible and lean system, we therefore use 60% to 65% machine efficiency as the baseline for the design. We use Little's Law to calculate how many machines are required to keep efficiency under 65% and also use Little's Law to estimate total manual labor hours required. Since machines come in large chunks, adding an extra machine might reduce the efficiency far below 65%. In this case, we might also vary the number of operating days per year or hours per day. For example, we might consider a design with overtime and fewer machines and compare it to a design with only 2 shifts and more machines. This is the fundamental capital versus labor trade-off.

The choice of machine quantities, operating days, and labor assignments are then used to calculate an ideal manufacturing cost. This cost focuses only on the cost of the capital and labor and ignores overhead costs, material costs, and many other costs. The ideal manufacturing cost is simply the price of a machine times the number of machines, plus the average labor wage rate times the number of hours of labor. While this can be done in a spreadsheet, our SeedTactic: Planning and SeedTactic: Designer tools are built exactly for this calculation.

Next, we want to consider adding an automated handling system and see how that impacts the ideal manufacturing cost. Automated handling systems can reduce the amount of labor (while adding capital expense) and therefore alter the ideal manufacturing cost. But a much larger effect of an automated handling system is its impact on efficiency. In our experience, a properly-managed automation system can reach 80%-85% efficiency with manual loading and 90+% efficiency with robotic loading. With larger efficiency, we might be able to purchase fewer machines, run fewer shifts, or add additional work to the system (lowering the part cost per piece). Similar to before, we therefore adjust the machine quantities, labor assignment, and operating hours to target 80%-90% machine efficiency and then compute an ideal manufacturing cost (adding the cost of the automation system into the ideal manufacturing cost).

This focus on efficiency via an automated handling system is a key advantage of our SeedTactic: Planning and SeedTactic: Designer tools. The labor savings from automation are typically easily visible and calculated on spreadsheets, but efficiency is typically hidden in these simple analyses. Our SeedTactic: Planning tool provides key visibility of the connection between efficiency and cost. In our experience, the cost savings from higher efficiency has a huge impact on the final cost of the system.

The efficiency increases from automation require some operational management techniques to prevent traffic jams. An improperly managed automation system typically runs around 60%-65% efficiency, which is the about the same as a manually operated system. Thus an improperly managed automation system is typically worse than manual handling; the lower labor costs are typically canceled by the cost of the automation system and no efficiency changes take place. Since the initial design probably targeted higher efficiency, additional overtime labor is usually required to keep up with order demand.

The good news is that properly managing an automated handling system is not hard and not that costly, it just requires a few simple tools and techniques. The main challenge is that while not costly, these tools and techniques should be budgeted and planned for ahead of time. The guide on preventing traffic jams discusses these tools and tactics.

The final result of Stage 2 is a variety of manufacturing scenarios, each with an ideal manufacturing cost. This comparison allows us to understand the capital versus labor trade-off and choose a manufacturing strategy. For the future stages, a specific manufacturing strategy will be studied in detail.

Stage 3: Flexibility Plan

How to implement lean manufacturing? Can we reduce in-process inventory and maintain a smooth flow? How to minimize non-value-added work? The calculations using Little's Law in Stage 2 assumed a fully flexible system where any part can go to any machine (one reason we call it an ideal manufacturing cost). A fully flexible system is highly prone to traffic jams, has a difficult prove-out process, and provides challenges for quality control. On the other hand, the system needs some flexibility to be robust to changing daily order mixes and resilient to traffic jams. One of the main challenges is that too much flexibility leads to traffic jams and problems but too little flexibility also leads to traffic jams and loss of throughput.

In our experience on a range of systems, the sweet spot is when about 20% to 25% of the work is flexible. This provides enough flexibility to maintain lean manufacturing, reduces in-process inventory, and keeps the machines busy. The part prove-out process is manageable and an inspection strategy can make sure that all possible combinations of flexibility are periodically examined. This is a main strength of our SeedTactic: Planning and SeedTactic: Designer tools: you specify a candidate flexibility plan and the tool uses simulation to present detailed projections of the operation of the system under this flexibility plan. The underlying calculations have been tried and tested on systems large and small for over 30 years, and have been refined to very accurately predict the operation of a candidate flexibility plan.

Stage 4: Cost Analysis

What is the projected cost/piece? Is the system justified from a business standpoint? Once the flexibility plan has been developed, the SeedTactic: Planning tool produces a detailed projection of the operation of the system. This projection can be used to refine the ideal manufacturing cost from Stage 2. Here, we add in material costs, overhead costs, and can use the projections to determine how to divide the capital and labor costs up among all the parts produced by the system. The SeedTactic: Planning tool performs these calculations and the resulting cost/piece can then be used to create a business justification for the project.

Sensitivity Analysis

Will the system still perform well in a variety of different daily order mixes? How will flexibility changes impact the performance? Systems typically have day-to-day or week-to-week variations in part demand. The goal of flexibility is to be able to adapt to these variations in demand, but is this true for the specific flexibility plan developed in Stage 3? Once the initial design has been reasonably finalized, a variety of order mixes over various time frames are entered and compared in the SeedTactic: Planning tool.


Our SeedTactic: Planning and SeedTactic: Designer tools are targeted at designing a flexible, lean manufacturing system through a series of simple stages: start with the goals for the system, decide on the manufacturing strategy, discover the correct amount of flexibility, and finally study and report on the system design. The SeedTactic: Planning tool walks you through this process, so we believe any engineer can design and implement a great, low-cost flexible manufacturing system.

Have more questions?

Contact Us