Middlesex Factory simulations

At Middlesex, we believe that Factory Simulation is a widely accepted way to analyze the performances of the transport system in function of the specific production environment.

To cope with the demanding requirements of the industry, we have developed a Simulation Environment, which has been widely used in the years to improve the quality of our transport systems.

Factory Simulations is a service provided by Middlesex not only to his clean room conveyor technology customers but to every customer that requires an accurate and realistic simulation of his factory.

For any specific inquiries you can contact us, we will be glad to provide you with all the required information.

simulation results


The simulation environment

Middlesex has developed an advanced Factory simulation environment for deploying accurate simulations for the semiconductor, in line assembly, hard disk, flat panel and pharmaceuticals industries.

The principal features of the Middlesex Simulation Environment are:

  • Our Simulation uses a time resolution of .25 s that allows simulating the factory with a high degree of accuracy, without incurring in excessive execution time.
  • The Middlesex Simulation Environment contains detailed models of all the proprietary conveyor-based transport equipments: rails, rotary corners, shuttles and elevators. But it is not limited to this: Our goal is to simulate the overall factory, not just the transport system.
  • We have implemented models for the most widely used tools, single WIP or batch processing; additionally, different transport system models are available, in order to allow the simulation of environments where more than one transport technology is used. For example, Fabs in which co-exist the conveyor and the vehicle transport technology can be easily modeled and simulated.
  • Our models can be easily modified or extended to fit the particular operative conditions of the factory: For example, semiconductor factories adopt production paradigms different from the ones of a disk maker or wafer maker: The number of process steps that the WIP has to complete, recursive or non-recursive processing are all factors that must be accurately modeled.

What we need to know

Aiming at simulating the overall production process not the mere transport system, we need to know some - limited – information from the customer.

Basically, we need to know:

  • the tools layout
  • the process time of the tools
  • the tools operative modality (batch, single WIP, ...)
  • the recipe for each type of WIP produced in the factory

The WIP start rate is another important input for the simulation. Generally, we suggest the customer to run the same simulation scenario with different start rates, in order to better explore the search space for the optimization problem.

The customer might also want to analyze the transport system response with dynamical conditions.

Tool downtime can cause temporary accumulation of carriers on the transport line. Defined the statistical distribution of the tool downtimes, the simulation will show if the intrinsic storage capacity of the conveyors is adequate to temporarily absorb the increase of in-line stored carriers.

Another example of dynamical condition could be the WIP start rate itself. This condition can bring to run more accurate simulations, where the WIP start rate is not fixed but it is varying, depending on the market requirements.

What to expect from the simulation?

At Middlesex, we consider simulations as a key factor in the design of the transport system.

Nearly every transport system that we have designed has been thoroughly in-house simulated. We think of this aspect as one of the ways to implement our total-quality commitment.

When the transport system is fairly simple, of course, there is no need for simulations. But when the dimension of the system begins to increase, we believe that the simulation is the only effective way to test and measure the system performances before installing the real system.

Dealing with complex transportation networks, the level of predictability decreased, mainly for the high interaction between carriers. Simple tool-to-tool traffic tables aren’t enough to analyze in depth the overall system behavior.

In these situations, we believe that only the comprehensive factory simulation can provide the most accurate results about the performance of the transport system.

If the transport system layout is provided by the customer, we can analyze its performances in “real conditions”, so to put in evidence potential bottle necks, possible improvements, online storage capacity and so on.

The following figures show the layout of the system and the related simulation.

simulation layout

Transport System Layout example

simulation GUI

Graphic Interface for the simulation (running off-line)

The results produced by our simulations are in form of charts of tabular data, according to the customer requirements. Typical statistics collected by the simulation are:

  • WIP total cycle time (average, std dev, min, max)
  • WIP transport/storage/process time (average, std dev, min, max)
  • tool utilization
  • online buffer (conveyors) utilization
  • moves (tool-to-tool) statistics
  • the operating curve (OC)

The OC (Operating Curve), i.e. the function that shows how the cycle time (total time to complete the production phase of a WIP) depends on the WIP starting rate, can be defined running simulations for different WIP start rates and then interpolating the result of the single simulations.

Another key point of our simulations is that we have developed all the code in-house, minimizing the dependence on commercial off-the-shelf packages. This allows us the greatest level of flexibility. Usually, we keep track of standard statistics like the ones defined above but the customer can look for particular indicators, related to the specific application field. Acting directly on the simulation code, for us is very straight to provide the customer with these additional data.

The following picture shows the conveyor utilization percentage throughout the 800 hours simulation. This particular indicator was requested by a customer.

simulation utilization percentage
Graph generated by the simulator showing the behavior of an indicator defined by the customer