Tips for Validating the Results of Structural Engineering Software: International Steel Construction Conference Recap

Last month, The Harman Group Vice President and Principal, Clifford Schwinger, P.E., was invited to to Medellín, Colombia to present a seminar at the Colombian Institute of Steel Construction’s 7th International Steel Construction Conference. Cliff was honored to be selected as one of the keynote speakers. He spoke on the topic, “Tips for Validating the Results of Structural Engineering Software.” With over 400 people in attendance, Cliff’s session reviewed the need for designers to validate structural engineering analysis and design models. The following points were discussed:

  • Use the computer as a tool, not a crutch. Do not assume a computer analysis is correct until the results are manually validated.
  • Validation of computer models requires an understanding of the minutia of the building code and design standards. Many equations and code provisions are programmed into structural engineering design software. Designers must be aware of these equations and code provisions in order to properly monitor the design and validate the results. Simple computations can be used to validate otherwise complex models.
  • Never stop thinking about structural engineering fundamentals: Strength, Stability, Load Paths, Bracing, Connections, Serviceability, Redundancy, and Constructability. There are important aspects of each that computer models will not consider.
  • Never get complacent, and always be wary of the results. If member sizes in the model do not look right, there is likely a problem. Do not rely on the computer to tell you that a suspect design works.
  • Question everything. Be cautious. Be skeptical. Is everything in the model? Perform in-house quality assurance reviews to get a second set of eyes on all designs.
  • All software has limitations. Be aware of and understand those limitations. What are the defaults? What does the software not check? How does the model differ from the real structure? Are lateral loads properly distributed through the model? Are any unrealistic results occurring due to rigid diaphragm assumptions? Are drag strut forces properly considered in the model? Are diaphragms properly connected to the lateral load resisting system? Do floor diaphragms have sufficient strength and stiffness? Did the software design the diaphragms?
  • Pay attention to unusual framing configurations that cannot be precisely modeled.
  • Look for unusual connections that are not analyzed by the software. Just because the software can model unusual or complex framing does not mean that the framing can be easily connected. Are the connections constructable? Are the connections designable? Constructability and connection designability are not considered by the software.

Computers and structural engineering software are wonderful productivity tools; however, as with all tools and technology, they must be used intelligently. If you have any questions, reach out to get in touch with Cliff at cschwinger@harmangroup.com.