Philosophy of Computer Modeling

Some might say that a "good" computer model is one that accurately predicts the response of a system, but this is only partly true. The practicing engineer generally must weigh the accuracy of the model against the risk, schedule, and/or budget involved. Through experience, an engineer learns where to focus resources and when an answer is good enough to base important decisions on.

Although AFT Fathom is designed to speed up the modeling process, it should never be used as a "black box." Your model input and output should always be carefully reviewed.

To make the most effective use of AFT Fathom, you should follow these guidelines:

  • Be aware of what AFT Fathom can and cannot do and the assumptions it makes.

  • Learn and use the tools in AFT Fathom that are appropriate for the specific analysis you want to perform. Detailed and accurate answers demand detailed and accurate input. Quick and dirty answers require less detailed input.

  • Learn and use the tools in AFT Fathom that allow you to process information in the formats that mean the most to you. Because pipe flow applications vary so much within different industries, AFT Fathom is customizable for the way you want to do your pipe flow work. Understanding the tools will help you gain greater conceptual control over your pipe flow analyses and reduce modeling errors.

  • Check your input data. Use the features available in AFT Fathom to review your results for input errors. Pay attention to the information in the output. Don't neglect warnings in the output. Perform hand analysis checks of results.

  • Bound your problem. Instead of trying to calculate the exact resistance in a network, or the right loss factor for an orifice, use your judgment, guided by data, on how much variation might occur. By running your model at the upper and lower limits, you can obtain answers, for example, on the maximum and minimum possible pressure drops in the system. If the results at the maximum and minimum are acceptable, your analysis is probably finished. If they are not acceptable, you can further refine the input model.

  • Recognize whether your problem is one of maximization or minimization. That is, recognize whether the design goal is to minimize the flow rate, for example, or maximize the flow rate. Then make your assumptions conservative enough to provide high confidence that your design requirements are being satisfied.