Simulation is a key part of modern engineering. Our device circuits, our bridges, our aircraft, our ships, our cars, our chemical plant are all simulated before they are built. We cannot nowadays afford to do otherwise . At one level you could argue that one of the important roles for an education in engineering is to enable our graduates to understand what lies behind these simulations, so that all of them understand their limitations, and a few of them can make better simulations in the future. This implies that we should stick to teaching the principles, not the use of the simulators. However there are many circumstances in which a simulation can be a very powerful learning tool.
Many engineers will be familiar with finite element analysis (FE), computational fluid dynamics (CFD), computer-aided design and manufacture (CAD/CAM) and with the use of tools such as ProEngineer or AutoCAD. You may choose to teach your students how to use one or more of these in a rudimentary manner, although they are very unlikely to become truly expert (see Chapter 4b). You are likely to teach the finite element method, rather than how to use ABAQUS, Nastran or ANSYS, and you want your graduates to be aware of the capabilities and potential of computer-aided methods. However I do not consider these tools, important though they are, to be genuine teaching simulations. Neither does MATLAB – often called a simulation environment – meet the full requirements. It is often used to illustrate the effect of variables on engineering behaviour but usually in essentially mathematical or graphical terms. This is only partial simulation.
The full potential of simulations to support teaching and learning is realised when difficult and/or complex situations are simulated in order to clarify and explain – to accelerate understanding, not the immediate solution of problems. Simulations of this type are surprisingly rare, probably because they are time-consuming to write and the commercial return on specialised teaching materials is relatively low. (As low as on books like this!) However there are some very good examples out there.
Flight simulation is an obvious place to start. There are many simulators in university engineering Schools, ranging from PC-based versions with a single screen to research simulators with vast capabilities and high realism. Their appeal to students is obvious, although their attraction usually results from the student’s enthusiasm for flying, not for engineering. Indeed it seems to me that few engineering principles are clarified by a flight simulator, although the scope for learning about control theory is significant. In my own university we use sophisticated simulators to help students learn about flight handling qualities, but not about many other aspects of aerospace or aeronautical engineering. They are principally a motivational tool.
Flight simulators do illustrate one of the key educational points about a simulation: In order to learn, it is essential that the student can determine and change the key parameters and thus interact fully with the simulator. A simulation is not the same as an animation.
Good examples of educational simulations include several examples of truss bridge simulators (e.g. DrFrame2D or West Point Bridge Design) which can be used in conjunction with practical exercises such as building cardboard bridges. They are configured as teaching tools in that they produce real-time responses as loads or structures are changed. A large-scale extension of this is the Constructionarium, at which students of civil or structural engineering build one-tenth scale models of real bridges, buildings and other large engineering structures. In this case simulation on the computer precedes, and is eventually validated by, the building of a real model.
Another very sophisticated and wide-ranging simulation tool is at www.steeluniversity.org This package of simulations of steelmaking processes was developed for use by both university students and steel industry employees and it illustrates many of the best features of an educational simulation, including:
- It deals with processes which, because of geography and danger, cannot be seen by most students;
- It was developed in close collaboration with steel company employees who are experts in their subject, so it is highly realistic;
- It offers the student the opportunity to control many parameters and to observe their effect, not just on the equations but on the product;
- It is supported by many (web) pages of background material, so that students can drill down for more detail or further explanation;
- It is associated with an international competition which increases student engagement.
Within steeluniversity, and the related MATTER site (www.matter.org.uk) are simulations of tests such as the Charpy and Jominy tests.
There are probably other good examples, but in my experience educational simulations meeting the criteria listed above for steeluniversity are rather rare. Electrical and electronic engineering is an area ripe with potential for simulations but a number of nominal ‘simulations’ have disappointing interfaces and outputs. For example SPICE (Simulation Program with Integrated Circuit Emphasis) originally required non-intuitive stacks of data, although more recent versions (e.g. 5SPICE) are somewhat more user-friendly and provide graphical representations of circuits. It would be good in many areas of engineering study to have more software which provided realistic, interactive and intuitive simulation.
There are also a number of simulations and games designed for school-level use, several of which can be found by searching the web sites of the former Engineering Subject Centre and UKCME. Racing Academy is a good example of a competitive approach to teaching mechanics and dynamics (Darling, 2008). At this level a simulation does not need to be computer-based. Magill and Roy  describe an exercise to simulate the fabrication of silicon devices using red ink, sticky labels and felt pens – low tech works well here and is very cheap.
When planning to use a simulation you need to think carefully how to ensure that the students learn rather than just play. Random clicking and changing of parameters might just reveal something interesting but this is frankly unlikely. Students need to be directed towards a simulation with a clear task in mind. You should at the very least ask them a question to which they can find the answer using the simulator. You also need to tell them that, if the simulation itself does not provide a facility for taking notes or recording their actions, it will be essential to record (possibly even on paper!) what they are doing as they proceed. In these circumstances a simulation can be extremely useful and thought-provoking, and the more inquisitive student will explore further and learn more (having taken ownership of this part of his learning).
Although it might seem that games for educational use will inevitably be computer based, this is not necessarily the most educationally-effective approach, nor the cheapest. Let me give one example of a simple cheap game which has been shown to make key issues in the optimisation of production clear and relevant to students. I owe this example to Laurence Legg . The game involves only paper, scissors and a stapler, and the artefact which is produced by the participating teams is simply two pieces of folded paper stapled together. This product could be referred to as a plane, a bridge, a chair or a Christmas card, depending on the class and the season. The manufacturing process is broken down into six or seven very simple operations, one of which takes much longer than the others. (When I did it the slow process was writing out the conference name on both sides of the ‘fuselage’.) There are a few rules but the key learning outcome is the importance of a bottleneck in production (the ‘drum’ whose beat controls the output of the process). The teams can be taught the exercise in about 15 minutes and can then be given 15 minutes to plan their production line and 15 minutes to produce as many products, or make as much money, as possible. The whole exercise can thus be completed within an hour. The consensus among the teams of academics when I took part was that this exercise is best carried out before the key concepts are introduced to the class, so that they realise the significance of the dry stuff about scheduling. I could however also see value in doing it after some of the topic had been covered. There are surely many other examples of this type of approach in use – ask around.
Read on … (but first please add a comment)
 Our climate is of course being simulated too, for predictive purposes, although I am not fully convinced that we know the parameters accurately enough to make this quantitatively reliable.