What is Flow Simulation?
Flow simulation, or Discrete Event Simulation (DES), is a powerful technique used to analyze the flow within a system. This flow could represent various scenarios such as a manufacturing process in a factory, logistics within a warehouse, or even the movement of people in a hospital.
Flow simulation serves multiple purposes. It’s often used to calculate or evaluate specific parameters, but it’s also effective for gaining insights into a system or for project guidance.
CALCULATIONS IN DYNAMIC SYSTEMS
One of the unique aspects of flow simulation is its ability to perform calculations on dynamic systems. For instance, it can answer questions like:
- What will the throughput/h be in the new production system?
- How many storage spaces are required for each variant?
- How many nurses are needed per shift?
FLOW SIMULATION MODEL
To perform these calculations, a Flow Simulation model is built. This model is essentially a simplified digital replica of reality. The Flow Simulation model is usually limited to covering only the area needed to answer the current questions.
THE FLOW SIMULATION MODEL USUALLY CONSISTS OF:
- Static objects that usually represent some form of equipment or process.
- Moving objects that usually symbolize a product, resource, or work order.
- Flows that outline how the moving objects can move between the different static objects.
- Logic that dictates exactly how the moving objects move in different situations or how equipment or resources make decisions.
DYNAMIC INPUT DATA AND DISTRIBUTIONS
Input data from either an existing or future reality is fed into the Flow Simulation model. This input data is often dynamic – for example a manual cycle time in an assembly station or the time interval between two patients arriving at a hospital. The dynamic input data is not described with constant values but with the help of a distribution (e.g., A normal distribution with an average value and a standard deviation).
RANDOM REPLICATIONS
When the simulation model runs, values are randomly generated from the distribution which creates a potential scenario and gives a result for each unique run (often called replication or observation). To get a reliable (sufficiently accurate) value, a replication analysis is done to find out how many replications need to be run.
STATISTICAL ANSWERS
The answers from all replications together provide a statistical answer to the question. For example:
- With 95% certainty, the true mean value lies between 10.3 parts/h and 10.7 parts/h. So there is a confidence level in the answer (95%) and a confidence interval (10.3 – 10.7).”