simulation
The simulation or simulation is a procedure for analyzing systems that are too complex for theoretical or formula-based treatment . This is predominantly the case with dynamic system behavior. In the simulation, experiments are carried out on a model in order to gain knowledge about the real system. In connection with simulation, one speaks of the system to be simulated and of a simulator as an implementation or realization of a simulation model . The latter represents an abstraction of the system to be simulated (structure, function, behavior). The process of the simulator with concrete values ( parameterization ) is called a simulation experiment. Its results can then be interpreted and transferred to the system to be simulated.
That is why the first step in a simulation is always finding a model. If a new model is developed, one speaks of modeling. If an existing model is suitable for making statements about the problem to be solved, only the parameters of the model have to be set. The model or the simulation results can then be used to draw conclusions about the problem and its solution. This can be followed by statistical evaluations - provided stochastic processes have been simulated .
The simulation method is used for many practical problems. Well-known fields for the use of simulations are flow, traffic, weather and climate simulation.
Subdivision
A distinction can be made between simulations with and without a computer. A simulation is an “as if” running through of processes; you can do that without a computer.
Without a computer
Physical experiments are also known as simulations: a car crash test, for example, is a simulation for a real traffic situation in which a car is involved in a traffic accident . The previous history of the accident, the traffic situation and the exact nature of the other party involved in the accident are greatly simplified. Also, no people are involved in the simulated accident; instead, crash test dummies are used, which have certain mechanical properties in common with real people . A simulation model only has very specific aspects in common with a real accident. Which aspects these are depends largely on the question that is to be answered with the simulation.
Tests in flow wind tunnels also fall into this category. Here, for example, statements about aerodynamic drag and lift of aircraft can be made on a scaled-down model. The same applies to fire simulations: Dangerous situations such as fires in closed rooms or vehicles are simulated and trained with real personnel for training purposes in rescue or extinguishing or new materials are tested for their fire protection properties.
With a computer
When we talk about “simulation” today, we almost always mean computer simulations. Basically, the simulation can be divided into static vs. dynamic and stochastic vs. classify deterministic simulation. In the static simulation, time as a dynamic variable does not play a role and is not part of the system. The deterministic simulation excludes random (stochastic) events.
Calls
There can be several reasons for using simulations:
- An investigation on the real system would be too time-consuming, too expensive, ethically unjustifiable or too dangerous. Examples:
- Driving simulator (too dangerous in reality).
- Flight simulator for pilot training, re-enactment of critical scenarios (engine failure, emergency landing).
- A power plant simulator , in which the operating teams of nuclear power plants in particular train the control of incidents up to and including a meltdown (too dangerous in reality).
- Crash test (too dangerous or too complex in reality).
- Simulation of production systems before a conversion (multiple conversion of the system in reality would be too time-consuming and expensive).
- Simulators in medical training (training on patients is not ethical in some areas).
- The real system does not (yet) exist. Example: Wind tunnel experiment with aircraft models before the aircraft is manufactured
- The real system cannot be observed directly.
- Due to the system. Example: Simulation of individual molecules in a liquid , astrophysical processes.
- The real system works too fast. Example: simulation of circuits .
- The real system is working too slowly. Example: simulation of geological processes.
- The real system is not understood or very complex. Example: Big Bang .
- The real system is understood in its elementary dynamics, but the development over time is too complex, or an exact solution of the equation of motion is not (yet) possible. Examples: three-body problem , double pendulum , molecular dynamics , generally non-linear systems .
- Simulation models can be modified much more easily than the real system. Example: biosimulation .
- Simulations are reproducible.
Areas of application
Different simulation types can be distinguished from the application point of view:
- Technical simulations, for example for strength calculations (FEM) , flow simulation, of factory processes and complex logistic systems , for virtual commissioning or circuit simulation
- Scientific computing , with applications in physics , chemistry , biology , meteorology
- Simulations for education and training, for example business simulation games or medical simulations
- Game simulations , for example flight simulations , racing simulations , economic simulations
Limits
There are also limits to any form of simulation that must always be observed. The first limit follows from the limited nature of the means, that is, the finiteness of energy (for example also computing capacity), time and not least money. A simulation must therefore also make sense from an economic point of view. Because of these limitations, a model must be as simple as possible. This in turn means that the models used often represent a gross simplification of reality . These simplifications naturally also affect the accuracy of the simulation results.
The second limit follows from this: A model only delivers results in a certain context that can be transferred to reality. In other parameter areas the results can simply be wrong. Therefore, the validation of the models for the respective application is an important part of the simulation technology. Other possible limits are inaccuracies in the output data (e.g. measurement errors) and subjective obstacles (e.g. insufficient flow of information about production errors).
See also
- emulator
- Hardware in the loop
- Monte Carlo simulation
- Numerical simulation
- Physics engine
- For a philosophical - aesthetic concept of simulation cf. the media theory of the French philosopher Jean Baudrillard and the related term of the simulacrum .
literature
- FE Cellier: Continuous System Modeling. Springer, New York 1991 ISBN 0-387-97502-0
- RM Fujimoto: Parallel and Distributed Simulation Systems. Wiley-Interscience, New York 1999, ISBN 0-471-18383-0
- BP Zeigler, H. Praehofer, TG Kim: Theory of Modeling and Simulation. 2nd edition. Academic Press, San Diego 2000, ISBN 0-12-778455-1
Web links
- Literature on simulation in the catalog of the German National Library
- Working group simulation of the society for informatics
- Link catalog on simulation at curlie.org (formerly DMOZ )