What is modeling and simulation (M&S)?

Benefits of modeling and simulation

Creating models to simulate phenomena is extremely valuable. By understanding how something will behave before it is made, organizations can save time and effort by avoiding prototype creation or costly missteps in production. Certain aspects can also be optimized during design, such as whether a manufacturer wants to optimize strength, material use, or a specific balance of the two.

Historically, this might have been done with physical scale models. For example, imagine putting a clay model car in a wind tunnel. This would often be supplemented by calculations of stress or load on the object.

Today, most M&S is done on a computer using mathematical models to simulate physical phenomena.

Digital twins seek to blur the line between simulation and reality. In a digital twin, the digital model and reality are kept in sync. This way, the model is kept accurate by reality, and potential changes can be tried digitally before physical changes are made.

Creating a model and simulation

Our world is complex. Innumerable factors influence the outcome of anything that can be modeled. The accuracy and efficacy of M&S are highly dependent on the level of detail included in the model. Therefore, creating a model is often a balancing act between what details to include and what to exclude. Carefully selecting what the M&S is trying to accomplish can help guide these decisions.

For example, imagine someone wanted to model how long it takes a dropped ball to hit the ground. An extremely simple model would use the height and the gravitational constant. A more complex one will include the drag coefficient of the air on the falling ball. An extremely complex computer model can perform a fluid simulation of the air moving around the ball and account for the gravity changing with distance. They might use a simple model to determine where to put a net to catch dropped golf balls, while a more complex one is used to determine the characteristics of a satellite entering the Earth's atmosphere.

Before beginning an M&S, it is important to understand the system in which it will operate. What is the goal of the simulation? What are the important aspects? All models are limited, so teams need to understand the areas for which they will not account.

Math equations underpin the model. These equations describe how the elements in the model behave. Carefully tuning the equation will ensure that it matches reality as closely as possible.

Once the model is prepared, a computer typically plays it out over time. The simulation can be faster or slower than real time depending on the speed of the computer and the detail of the simulation. For example, a weather simulation using predictive modeling would need to run faster than real time to have any value, while a detailed simulation of an explosion might take hours to run.

With the model prepared, the simulation is often run many times. Small changes can be tried to see how they affect the outcome.

Modeling and simulation with supercomputers

The availability of high-performance computing, often using supercomputers, has now made common the things that were once impossible to simulate. Complex models can now simulate extremely detailed models. These have been applied to areas such as global weather patterns, nuclear engineering or simulating biological processes.

A recent example is NASA using the Oak Ridge supercomputer to simulate a propulsive landing on Mars. This simulation generated petabytes of data each time it ran and used tens of thousands of GPUs. It modeled things like the rocket plume interacting with the Martian atmosphere. Even though running these simulations were expensive, it is still much cheaper and easier than sending an actual rocket to Mars.

Applications of modeling and simulation

Modeling and simulation are common in almost every field of science and engineering. It is now customary to design and model things in computer simulations long before a physical item is made.

Some common real-world applications include the following:

• Product design. A product can be fully modeled before any part of it is built. This model can creator they physical characteristics of the item, such as it strength or performance. Models are used to create the supply chain, the tooling used in product creation, and how to make molds for parts of it.
• Architectural models. For many years, three-dimensional (3D) modeling software could tell you how strong a build is. Newer software can now tell you much more, including accurate simulations of heating and cooling behavior or how people will move throughout a building, which makes these models more accurate.
• Social sciences. M&S can be used to create models of human behavior. This can tell us how people or economies react to changing circumstances.
• Models of weather systems. These simulate behavior based on available data to generate predictive information for forecasts. A hurricane forecast model, for example, is designed to predict a given storm's track and intensity, as well as related events such as storm surges. Simulating the effect of severe weather events, like hurricanes and storm surges, on infrastructure can guide the design of more resilient systems.

Companies across industries use predictive analytics to forecast future trends and actions. Learn about the most common use cases for predictive modeling. Also, learn about the mathematical technique called the Monte Carlo simulation, which simulates the range of possible outcomes for an uncertain event.