An Articulate Virtual Laboratory for Thermodynamic Cycles

CyclePad is the first articulate virtual laboratory the Qualitative Reasoning Group has implemented. CyclePad enables students to construct and analyze a wide variety of thermodynamic cycles. A hypertext explanation facility provides the student with access to the chain of reasoning underlying the derivation of each value. CyclePad is currently being field-tested in undergraduate engineering classes at Northwestern University, The U.S. Naval Academy, and Oxford University.

A thermodynamic cycle is a collection of components which either takes in heat and produces energy, or takes in work and produces some transfer of heat, perhaps as a refrigerator or as a heat pump. Examples of thermodynamic cycles include power plants, refrigerators, propulsion plants, and engines. CyclePad helps you:

  • Specify the structure of your design , in terms of the parts of the cycle and how they are connected together.
  • Analyze your design, by figuring out the consequences of assumptions you make about it. Such assumptions include numerical values, e.g. operating temperatures and pressures, and modeling assumptions, e.g., whether or not to consider a turbine as isentropic.
  • Perform sensitivity analyses to understand how different choices of your design contribute to its performance. For example, CyclePad can figure out how the efficiency of a system changes as a function of other parameters, such as a turbine inlet temperature.

CyclePad performs steady-state analyses of both open and closed cycles. In an open-cycle, it is the components that are open to the passage of mass through them, while in a closed cycle different processes take place within a single component. Gas turbines are therefore open cycles, and piston engines are closed cycles. Note that a closed-loop steam cycle containing a boiler, turbine, condenser, and pump is still considered to be an open cycle.

Steady-state analyses provide the kind of initial guidance needed in conceptual design, because in the conceptual design of thermodynamic cycles the important questions concern the operating conditions and estimates of efficiency and cooling/heating/power produced by the cycle. (Later stages of design concern issues such as the response of the system to transients, developing procedures for safe startup and shutdown, and ensuring that the system is easy to monitor and maintain).

CyclePad works in two phases, build mode and analyze mode. In the first phase (build), you use a graphical editor to place components and connect them with stuffs. Such a structure might look like this:

While you can always quit CyclePad at any time, you can only proceed to the next phase (analysis) when CyclePad is satisfied that your design is fully laid out, that is, when every component is connected via some other component via stuffs, and every stuff has been used as both an input and an output for components in the design. Once your design is laid out, the real fun begins--the analysis phase.

In the analysis phase, you specify:

  • What working fluid you are using.
  • What modeling assumptions you wish to make in analyzing your design.
  • Numerical values for the properties of components and stuffs

As soon as you give CyclePad some information, it draws as many conclusions as it can about your design, based on everything you have told it so far. When you specify a working fluid, for instance, it knows whether to use property tables or an ideal gas approximation. When you specify numerical values, CyclePad sees if it can then calculate other numerical values. It displays the results of its calculations, and you are free to inquire about how values were derived and how one might proceed at any time, using a hypertext query system.

As you provide more information, CyclePad deduces more about the physical system. Eventually, you may have filled in all of the relevant information about the cycle, so that you have numerical values for properties such as the coefficient of performance (if you are designing a refrigerator), the thermal efficiency (if you are designing a heat engine), or other properties of interest such as the total amount of work produced or consumed by the cycle. How far you go is up to you. At any time you can save your design to a file so that you can continue working on it later, and generate reports describing the state of your analysis of the design.

CyclePad also supports sensitivity analyses. For instance, suppose you wanted to understand how the thermal efficiency of the cycle varies as a function of the efficiency of a compressor or some other component. Such analyses are quite tedious to do by hand, but CyclePad makes them quite easy and will generate such information for you in graphical form.

"Common Lisp remains one of the best languages for Artificial Intelligence applications, its flexibility enables rapid experimentation and deployment", said Professor Ken Forbus, Walter P. Murphy Professor of Computer Science at Northwestern University. "Today's Lisp compilers are robust and flexible allowing development entirely within Lisp or in combination with other languages. For example, our CyclePad system is written entirely in Allegro CL. Similarly, our sketch understanding system, CogSketch, which is a novel platform for both cognitive science research and education is primarily written in Allegro CL with two modules in C."

You can download CyclePad for free here. For additional informaiton on the CyclePad Project, see here.

About Ken Forbus

Ken Forbus, the project leader, is the Walter P. Murphy Professor of Computer Science and Professor of Education at Northwestern University. He is notable for his work in qualitative process theory, automated sketch understanding and on automated analogical reasoning. He developed the structure mapping engine based on the structure-mapping theory of Dedre Gentner. He is a Fellow of the Association for the Advancement of Artificial Intelligence. He is the Chair-elect of the Cognitive Science Society and became the chair (president) of the society in August, 2011.

Copyright © 2023 Franz Inc., All Rights Reserved | Privacy Statement Twitter