While computer simulations (such as Finite Element Analysis) are valuable tools, the results of these analyses should be viewed with caution – especially when these analyses are not carried out by specialists (i.e. untrained/inexperienced engineers). 

Colourful models can be convincing, but they are often sensitive to small changes including i) how, where and what loads are modelled, ii) the magnitude of these loads, iii) how accurately these loads represent the actual, in service loads, iv) the boundary conditions (are ‘simply supported’, ‘built in’ or ‘point load’ assumptions valid), v) the extent of modelling (which areas of the component are modelled and in what detail), vi) how joints/structural interactions are modelled (contact, friction), and vii) differences between modelled and as-manufactured geometry.

Without validation, these models may overlook critical factors, leading to inaccuracies and in order to gain true confidence in a design, physical verification is essential.  Physical measurement serves as the definitive test for the accuracy of a numerical model.  By measuring actual stresses on manufactured components, engineers can bridge the gap between theoretical design and physical reality.

These assertions are illustrated by many real-world engineering case studies that include:

Definition of Dynamic Loading Cases Environmental Interactions

Although computer models can be used to predict the effects of dynamic loading, the nature of the input loading is often difficult to assess accurately.  Strain gauging can capture the complex, fluctuating stresses that are difficult to model.

• Case Study: Offshore Drill Pipes
  While models of an offshore drill rig predicted high bending stresses in the pipes closest to the seabed, they failed to account for the full dynamic loading.  Strain gauging revealed that engagement with the seabed induced dynamic stress spikes with magnitudes of up to 40% of the yield strength of the material that travelled up through the string and were then reflected from the top of the drill string.  Furthermore, the strain gauging exercise identified highly cyclic ‘stick-slip’ torque reversals,  a phenomenon the original models had completely overlooked.

Identifying Unanticipated Loading Scenarios

Designers often model ideal scenarios (e.g. pure axial, bending or torsional loads are applied to the structure), but real-world operation and manufacturing tolerance often introduce unexpected additional loads.  Strain gauging can help identify these loads.

• Case Study: Underwater Diving Chamber
  Measurement of strain in the shell of an underwater diving chamber during acceptance testing (undertaken in accordance with the requirements of the applicable design code) highlighted that the stresses in one area of the shell of the chamber were notably higher than those predicted by the FEA modelling undertaken during the design.  Detailed inspection and further FEA analysis highlighted that small variances in the alignment of the plates used to manufacture the shell of the chamber made a significant difference to the local stresses induced by the high pressure loading experienced by the diving chamber.

• Case Study: Valve Stems
  Despite calculations showing surprisingly low stresses in the stems of the valves used in a power plant, ongoing fatigue failures of these stems were experienced.  Strain gauging and careful inspection revealed that an eccentricity of the seating of a lock-nut caused large bending stresses (150MPa) to be induced into the stem, which was only designed for axial and torsional loads.  Once identified, a simple increase in the length of the lock-nut reduced the stress to 25MPa, which was well below the fatigue threshold.

Validating Fatigue and Lifecycle Predictions

In critical applications involving wind or cyclic loading, fatigue (failure over time due to cyclic loading) is a major concern.  Strain gauging can provide the data necessary to give accurate predictions of fatigue life that aren’t related to assumptions regarding the frequency and magnitude of the induced loads.

• Case Study: Excavators and Cranes
  In order to increase productivity, operators of excavators often slew the boom as fast as possible.  While the boom is optimised for bending, the rapid deceleration of a heavily loaded bucket introduces significant torsion (twisting) at the hinge joints between the boom and the stick as well as the joint between the boom and the body of the excavator.  In a particular case of an excavator that had been fitted with a long-reach boom, strain gauging highlighted that these torsional forces were causing premature fatigue failure of the boom/stick hinge.  This fundamental understanding of the cause of cracking allowed the operational limits on slew speed to be reduced, which in turn reduced the stresses sufficiently to prevent further fatigue cracking.

• Case Study: Stadium Roofing
  Wind loading on structures is often difficult to analyse, even when CFD analyses are employed.  In one unique case involving an annular roof on a sports stadium bolted to a prestressed cantilevered framework, repeated fatigue failures of the connecting bolts were experienced.  Measurement of wind loading and the strain induced in the bolts showed that there was a direct correlation between wind speed/direction and high magnitude, fluctuating stresses in the bolts, which, in the absence of a flexible joint that allowed movement, were being forced to carry the load.  Once understood, the problem could be rectified by relatively simple modifications to the support structure.

These selected examples illustrate the symbiosis between stress measurement/strain gauging and design/FEA.  While computer modelling is an efficient and indispensable tool for design, strain gauging remains an essential step for validation of numeric analyses, identification of loading, and preventing catastrophic failure.  Ideally, both strain gauging and numerical modelling (CFD and FEA) should be employed to gain full understanding of the loading and to optimise the design.