4.2 Introduction to Physics Simulation

4.2.1 Physics Engines

A physics engine is a piece of computer software that simulates the physical world [33]. A physics simulation is a recreation of a real-world occurrence. These simulations could be in the real world, in a virtual environment by using a physics engine, or partially in both. The existence of virtual simulation brought about a field of simulation-based physics. This field, also referred to as physics simulation, deals with many topics that are studied within physics, but specifically how they are adapted to be simulated within a physics engine. Each of these terms such as physics engine, physics simulation and simulation-based physics have different meanings. The first is a type of software, the second is what occurs when you run a physics engine, and the third is a field of study.

An example of a physics engine would be something like the Havok physics engine [1]. The Havok engine is a physics engine specifically made for video games and not scientific research. However, it is still a good example of an accessible physics engine. There is a video of the Havok Physics engine running on the front page of their website. This video is showcasing physics simulations. Physics simulation is also synonymous with physics model. This specific physics engine is utilized in many games including the Age of Empires series, the Battlefield series, and some Call of Duty games [2]. An example of simulation-based physics is easy to find within the PhET simulation website [3]. These simulations will be discussed more in Section 4.2.4.

Although it may seem that simulation-based physics is far away from everyday life, it is closer than you may think. Have you ever been in a car or a plane? Have you experienced the weather? For one, simulation-based physics is used to test the safety and effectiveness of cars and planes. This is done through a field of simulation-based physics known as crash simulation [4]. This technique is used by automakers during the design phase of a car to determine if the chassis should perform adequately when tested in the real world. Physics engines can also be used to predict the weather quite effectively. This is done with numerical weather prediction [5]. The NOAA (National Centers for Environmental Information) oversees the standardization of numerical weather prediction. More specific information and what physics engines the NOAA uses are available on their website (Here is a good example).

4.2.2 What to Simulate with Physics Engines

Almost anything can be simulated in a physics engine. Very few problems are not possible to be simulated. An example of such an intractable problem would be simulating the entire universe. According to some theories the universe is of infinite size. This would be a problem for computers, which can only simulate a finite area. Then, if the universe is infinite, computers can not and will not be able to simulate the entire universe. Another example of something which cannot be simulated with conventional computers is some areas of quantum physics. Many phenomena, when simulated with conventional computers, would experience an exponential decrease in performance as the number of particles in the system increases. Thus, quantum computers must be used to simulate these problem, and that is an entirely different topic [6]. Then, what can be simulated?

  • Simulatable physics concepts include:
  • Elementary Newtonian Physics
  • Aerodynamics
  • Fluid Dynamics
  • Climate Science
  • Soft-body Physics
  • Thermodynamics

4.2.3  Examples

In this section, we introduce some examples of problems that can be solved with Big Data Science and Physics simulation.

Aerodynamics Modeling

figure4-1

Figure 1: simulation of aerodynamics around a car [7]

Aerodynamics is a commonly modeled field of physics. For cars, modelling aerodynamics is important to calculate fuel economy, stability, the noisiness of the cabin, the cooling of the engine, and the performance of the engine [8]. These functions are all vital in the design of a car. Fuel economy is one of the most important aspects of modern cars. If a car is not aerodynamically stable, it could behave erratically at high speeds. Discovering the aerodynamic properties of a car is a very important task in designing a car.

Soft-Body Physics

figure4-2

Figure 2: simulated crash running in CryEngine 3 [9][10]

Soft-body Physics is often used in video games, but it is also used in the real world to simulate vehicle collisions [11]. A company called CrashTeams can recreate a real world accident using industry-accepted software. This type of recreation can be important when applying for insurance claims or if the accident case must go to court. Car manufacturers like GM are also using simulated crashes to help streamline the design process and create safer cars [12]. Soft-body Physics has also been used to recreate entire plane crashes and helped determine the causes of those crashes [13]. Soft-body physics is a widely used and integral simulatable field.

Climate Science

figure4-3

Figure 3: basic model of ocean currents around the globe [14]

EPA as well as other similar organizations extensively simulate the climate [15]. They use these simulations to perform many analyses including CO2 projections, forestry land-use, and global change assessments. Climate science models are also used to predict the weather, which may have scientific ramifications. The prediction could also be displayed on your local news channel [16]. Regardless of how climate science models are used, they play an important role in everyday life.

4.2.4 Tools for Interactive Physics Simulations

Accessible simulation engines are available for public consumption. One well-known physics engine is a Falling Sand Game which is available here. This game is useful in demonstrating why simulations need Big Data. This small game can easily be bogged down if too many particles are added, so imagine simulating a real world problem. Many other more simple simulations can be found here. This website has many more useful simulations. There are simulations of almost everything from static electricity to Newtonian mechanics to the photoelectric effect. Companies like Design Simulation Technologies combine these separate simulations into some actually useful simulations [17].

4.2.5 Important Issues and Challenges in Physics Engines

Physics engines have some issues. Modern engines can only simulate a small portion of data when compared to the real world. This error propagates to very large amounts of error if a scientific experiment were to be done within a physics engine. For example, the SuperComputational Life Science team used multiple supercomputers and was only able to accurately simulate the functions of the human heart, a fairly simple organ [18]. Another interesting example is N-body simulation. This type of simulation is used mostly by scientists trying to create models of the universe. Different complex methods of optimization must be employed like tree methods and particle-mesh methods [19]. Big data science must be used to further the capabilities of physics engines. Big data is needed when larger or more complex problems arise and need solved.