Modeling and Simulation

Modeling and simulation (M&S) refers to the use of models (e.g., physical, mathematical, or logical representation of a system, thing, phenomenon, or process) as the basis for simulations to generate data that can be used for management or engineering decision-making.

Digital Application of Modelling and Simulation

Modeling and Simulation’s digital application uses computer software to create a mathematical model that contains the key parameters of the physical model. The mathematical model represents the physical model in virtual form, and conditions are applied that initiate the intended digital experiment. The simulation starts—that is, the computer calculates the results of these conditions on the mathematical model—and outputs the results in a format that is either machine-readable or human-readable, depending on the implementation.

From a technical point of view, the simulation is well accepted nationally and internationally. The reasons for the ever-growing interest in simulation applications include the following:

1. Using simulations is generally cheaper, safer, and sometimes more ethical than conducting real-world experiments. For example, supercomputers are used in climate analysis to simulate weather phenomena or, in extreme cases, to simulate hurricanes and other natural disasters.
2. Simulations can often be performed faster than real-time. As a result, they can be used for efficient if-then analysis of various alternatives, especially when the data required to initialize the simulation can be easily obtained from operational data. This use of simulations expands the toolbox of traditional decision support systems to include decision support simulation systems.
3. Simulations enable the construction of a coherent synthetic environment that enables the integration of simulated systems in the early analysis phase, mixed virtual systems with initial prototypical components, and a virtual test environment for the final system. When handled correctly, the environment can be moved from the development and test domain to the training and education domain in subsequent lifecycle phases of the systems (including the ability to train and optimize a digital twin of the real system under realistic conditions even before the first components are built).

Advantages and Disadvantages of Modelling and Simulation

The use of such mathematical models and simulations avoids actual experiments, which can be costly and time-consuming. Instead, mathematical knowledge and computing power are used to solve real-world problems in a cost-effective and time-saving manner. In this way, Modelling and Simulation can help to understand the behavior of a system without testing the system in the real world. Simulation can analyze systems under different (real or unreal) conditions. Furthermore, modifications to the system are easy to carry out and thus experiments can be easily controlled and executed in the simulation.

For example, when designing a racing car, to find out what type of spoiler would improve traction the most, a computer simulation of the car could be used to estimate the effects of different spoiler shapes on the coefficient of friction in a corner. Useful insights into various design decisions could be gained without actually building the car.

In addition, simulation can support experiments that take place entirely in software or in human-in-the-loop environments (model that requires human interaction) where the simulation represents systems or generates data needed to achieve experimental goals. Thanks to today’s computer capacity, many experiments can be carried out in parallel without much effort. Simulation also allows systems to be analyzed in very short or very long time intervals that are not observable in reality. Of course, simulation can also be used to train people in a digital environment that would otherwise be difficult or expensive to physically produce.

However, there are also limitations to modeling and simulation:

Simulation models require a lot of data and are complex and expensive to develop. Simulation models provide extreme amounts of data, so that the accuracy and validity of the results are often not sufficiently verified or overestimated. Typical errors in the application of simulation can be observed when the model is created without a specific objective. Also, an incorrect level of detail of the model with unnecessary details can lead to problems. The effort required for data collection and validation is underestimated, or excessive and expensive requirements are placed on the animation. Also, no or too little use of statistical methods or ignorance of the dependencies of the simulation variables or the lack of estimation of the scope of validity can lead to completely nonsensical statements.

Areas of Application

Image source: https://www.researchgate.net/figure/Modeling-and-simulation-driven-engineering-The-methodology-diagram-based-on-the-DEVS_fig2_303533042

M&S are important in scientific research. This often involves scientific computing, with applications and analyses in physics, chemistry, biology or meteorology. Representing real-world systems, either through small-scale physical replicas or through mathematical models that represent the dynamics of the system through simulation, allows for the study of system behavior in a representable way that is often either difficult, impossible, or too risky in the real world.

The use of M&S in engineering is widely recognised. Simulation technology is one of the system-technical tools used by engineers of all disciplines. M&S helps to reduce costs, increase the quality of products and systems, and document and archive experience gained. Examples of technical simulations include strength calculation (FEM), flow simulation, strength calculation etc.

Since the results of a simulation are only as good as the underlying model(s), engineers, users, and analysts must pay special attention to the design of the simulation. To ensure that the results of the simulation are transferable to the real world, the user needs to understand the assumptions, conceptualizations, and limitations of the implementation. In addition, the models can be updated and improved based on the results of actual experiments. M&S is a discipline in its own right. The many areas of application often lead to the assumption that M&S is limited to the mere application of a computer program. This is not the case, but in each specific case, when using M&S, it must be recognized whether the model and the simulator fit together.