CALIFORNIA STATE SCIENCE FAIR
2002 PROJECT SUMMARY
Ap2/02
Name(s) Project Number
Project Title
Abstract
Summary Statement
Help Received
Gregory P. Schuster
Truss Analysis: A Study in Engineering Practices
J1809
Objectives/Goals
"Can you combine member strength analysis and static analysis to form a predictive model for craft stick
truss bridge failure?" My project applies engineering practices to a craft stick truss bridge by examining
materials and loads. My predictions are based on data I collect through my own member tension,
compression, and buckling tests and which then is combined with my static analysis of a particular truss.
Then I will build the truss and apply a mid-span point load until the truss fails. My Variable: Truss Type,
will have three study values to be sure that my predictive model works with more than one truss type and I
will build three of each type.
Methods/Materials
My trusses were constructed out of craft sticks, wire pins, and joint plates. I prototyped the joint plates, to
asuure stability, without providing structural support. Using my tension, compression, and buckling
fixtures, I tested the craft stick member's strengths, for both long and short sticks. Combining the data
collected from these member tests with the static analysis of each truss type, I made predictions on the
failures of the three trusses. I set Variable: Truss Type, with three study values: #1: Warren
Through-Truss; #2: Howe Through-Truss; #3: Pratt Through-Truss. Building and testing conditions for
all trusses were the same. On each truss, I applied a mid-span point load using a bucket of weights.
When the truss failed, the weight was recorded with the member name where the truss failed. I compared
the actual failure weight and member location with my predictions.
Results
The Howe Truss was the most accurate prediction, breaking within 2 lbs. of my prediction. The Pratt
Truss also had good results. I estimated a small range (71.7 lbs +/- 6.3 lbs) of failure. The prediction was
104% of the average. The Warren Truss prediction was incorrect. I attribute this to using only the
compression test. Since I performed a buckling test later, the Warren truss prediction would be
somewhere between the buckling and compression averages, which would have been correct. In every
case, I successfully predicted which member would fail.
Conclusions/Discussion
My hypothesis was correct; combining static analysis and member strength testing, I was able to
accurately predict when the truss would fail and which member would fail. The member strength test
results for compression and tension were very accurate.
Combining member strength tests and static analysis, I found that I could make a predictive model of truss
bridge failure by testing three truss types.
My father taught static analysis, aided with experimentation, & building apparatuses.