Created analysis model Validated analysis model Verified analysis

ü Created analysis model ü Validated analysis model ü Verified analysis model But do you really understand how your structure is behaving? ? ?

Sensitivity Analysis

Sensitivity Analysis • Investigate the effects of different values for features/parameters. Why? • Better understanding of the behaviour of system being modelled. How much is necessary? • Depends on how certain you are of the model behaviour.

Reference model results Reference model Change 1 feature/ parameter New revised model New results Change a different feature/ parameter New revised model New results • Change variables one at a time, go back to the reference model after each change.

• Consider the parameters most indicative of the model behaviour: – Maximum deflection? – Maximum bending moment/shear force/axial force? – Lowest natural frequency? • Compare and report results without dimensions. – Use percentage difference +/- % %diff = (value – reference value)/reference value 100 – Or ratios For more detail see: http: //www. imacleod. com/Using. Struct. M/Sensitivity-Analysis. pdf

Example 1. 5 m The Model Geometry: The model is created in 2 D, in the x-y plane. 1. 3 m 1. 4 m 1. 3 m 100 x 3. 6 SHS – for all top and bottom chord members and vertical members. 100 x 8 EA – for tension diagonals.

Model 1 – pin joints • Axial load • Displacement

Model 2 – fixed/rigid joints • Axial load • Bending moment

Model 3 – no bracing • Axial load • Bending moment

Comparison of results Max Displacement at node 4 % (mm) difference ratio Max Axial Bending % (k. N) % (k. Nm) difference compression difference Reference - Model 1 Pinned truss -0. 5920 - Model 2 Fixed truss -0. 5840 -1. 4% 1. 0 -0. 1163 Model 3 Vierendeel, no bracing -6. 1200 933. 8% 10. 5 -2. 294 - - -13. 67 - - -13. 52 -1. 1% -10. 62 -22. 3% 1872. 5% Structural system comparison

Consider truss chord depth v diagonal depth Reference model - Model 1 Pinned truss 100 x 3. 6 SHS – for all top and bottom chord members and vertical members. 100 x 8 EA – for tension diagonals.

Model 4 – Equal sections 100 x 3. 6 SHS – for all top and bottom chord members and vertical members. 100 x 3. 6 SHS – for tension diagonals. • Axial load • Displacement

Model 5 – section d, Chords–Ties 2: 1 100 x 3. 6 SHS – for all top and bottom chord members and vertical members. 50 x 3. 6 SHS – for tension diagonals. • Axial load • Displacement

Model 6 – section d, Chords–Ties 1: 2 100 x 3. 6 SHS – for all top and bottom chord members and vertical members. 200 x 5 SHS – for tension diagonals. • Axial load • Displacement

Comparison of results Max Displacement at node 4 % (mm) difference Reference Model 1 Pinned SHS & EA ratio x Max Axial (k. N) % compression difference -0. 5920 - -13. 67 Model 4 Equal sections -0. 6072 +2. 6% 1. 02 -13. 67 0% Model 5 d, Chords–Ties 2: 1 -0. 7572 +27. 9% 1. 28 -13. 67 0% Model 6 d, Chords–Ties 1: 2 -0. 5193 -12. 3% 0. 88 -13. 67 0% Truss chord depth v Diagonal depth -