Stress-wave force balances are important tools for testing of hypersonic models. By measuring the stress waves propagating within the model/balance, and hence the dynamic response of the apparatus to a load, these balances can be used to determine the forces acting on a model. They are of use even if the forces act only for a fraction of a second. Such short-duration loads are encountered during hypersonic testing in free-piston shock tunnels.
Finite element methods are commonly used as aids in the process of designing the stress-wave force balances. Finite element models are used to predict the dynamic response of the force balance system to an applied load. In the past, significant discrepancies have been noted between results gathered through experimental testing of models and results predicted using finite element methods.
This work investigates the use of a commercial finite element package, MSC Nastran, in modelling the propagation of stress waves within two simple wind-tunnel models. Results from finite element analyses are compared to results gathered experimentally. The natures of any deviations between the predicted and measured results are identified. Aspects of the finite element modelling process are altered one at a time to determine which aspects the modelling process the results are sensitive to, and their effects on the accuracy of predictions.
Comparison of data from repeated experiments revealed the possible presence of a systematic error in either the experiment or the subsequent data analysis. Results that showed good initial agreement slowly diverged over time.
The validation process determined that Nastran’s transient analysis solution process could accurately predict the speed of stress wave propagation, and could accurately predict the magnitude of stress waves even after numerous reflections and refractions.
By examining the results of finite element analyses of simple stress wave balances, it was found that Nastran was capable of generating accurate predictions of stress wave propagation within a more involved model. The speed of propagation and amplitude of predicted stress waves showed good agreement with results determined experimentally.
The results suggest that the assumption of unbroken, uniform contact between contacting faces at a component interface in an assembly is valid when constructing a finite element model, at least for the simple cases considered. The assumption that the internal damping properties of common metals can be ignored was also supported by the results, although this part of the investigation was based on incomplete data.
On the basis of this work, it is recommended that investigation into the cause of the divergence of experimental results be initiated. Also, the work conducted in this thesis should be extended to comparing experimental results with finite element predictions for wind-tunnel models of greater complexity. Research into the area of damping properties of commonly used materials is also justified.