Thermal Residual Stresses in Bonded Repairs

More Info
expand_more

Abstract

The number of ageing aircraft is increasing and will further increase considerably in the near future. The control of damage by inspection and repair becomes therefore more and more important. The bonded repair method proved to be a cost-effective way to repair damage. However, this method has some disadvantages. One of them is the introduction of thermal residual stresses in the repair, due to the mismatch in coefficients of thermal expansion between the skin material and the repair material. Two analytical models are available to determine these residual stresses in the repairs, i.e. the Wang-Rose model and the Van Barneveld-Fredell model. A new computer program CalcuRep2000 [1] is being developed. This computer program will be used to calculate the stresses in bonded repairs, and is developed by the US Air Force in co-operation with the faculty of Aerospace Engineering of Delft University of Technology in The Netherlands. For this program an analytical model, which is able to calculate the thermal stresses in a bonded rapair accurately, is essential. This paper describes how the most adequate analytical model was identified by comparison with Finite Element (FE) analysis. Two repair configuratios were distinguished: The test specimen, i.e., bonded repair specimen used for fatigue tests in laboratories, and the in-field specimen, i.e., a realistic repair. The difference between both repair configurations is that a test specimen is free to expand during curing while in a realistic repair configuration expansion is partly prevented by the cooler surroundings of the heated area. The analytical model of Van Barneveld and Fredell proved to be the most accurate one. The results of this model agree with the results of the FE model of the test specimen and the in-field specimen. The Wang-Rose model calculates the thermal residual stresses accurately in the case of a test specimen. but the stresses calculated in the case of an in-field specimen are less accurate. No useful experimental data are available in literature from measurements of the thermal residual strains on bonded repairs. Therefore experiments were conducted to measure thermal residual strains in bonded repairs. The same repair configurations were used for these experimentss. The results obtained with the experiments are compared with FE results. Finally, the temperature distribution calculated by the analytical models is compared with the temperature distribution measured during the experiments of the in-field specimen. The influence on the thermal residual stresses of a different temperature field is discussed here, using the FE results