|
The paper presents the method for determination the thermal properties of composite material by means of active thermography and 3D numerical non-stationary heat conduction model. The method is tested on the samples made of composite materials having known subsurface defects. The results obtained are discussed from the point of view to predict the behavior of the composite material by estimation of their thermal properties. The software developed enables the calculation of the temperature distribution through the sample versus time and location of the defect by means of inverse method involving experimentally obtained thermogram.
The temperatures and their change in time on the surface of the observed object exposed to heat stimulation are information, which indicate the conditions in the object structure. Surface temperature distribution is in that case also the reflection of the thermal properties of the object material. In the case of material non-homogeneity or presence of another material in the base structure the temperature differences in the surface temperature distribution can be noticed.
Necessity to know the structure of the material in various technical testing without its destruction seeks for adequate non-destructive method. Combination of thermography and 3D numerical model suit well for such assignment.
The paper presents results of the research obtained on the samples having known subsurface defect. Two samples were tested by means of thermography and analysed by numerical method. The samples were made of phenoxy resin and phenoxy resin containing cooper fibres. The defects were in the form of grooves having various dimensions.
|
3. The goals of the research |
Because of a complex relations between defects and process parameters as well as thermal properties of the primary material and material of the defects the research was addressed to particular relations.
The samples made of phenoxy resin with and without cooper fibres were tested. For the sample containing 15% of cooper fibres thermal properties were unknown. The goal of the research was to find them on the base of estimated properties of the material. The method applied gives a good response for defect determination.
In the both experiments the experimental rig was the same one. The IR heater (0.5 kW) thermally stimulates the sample with variable duration of heating. The surface temperature distribution on the sample was recorded by IR camera AGEMA 570 PRO.
Figure 1. The experimental rig, 1 – thermographic equipment, 2 – infrared camera, 3 – lamp, 4 – canal, 5 – isolation, 6 – model, 7 – thermocouple on depth 12 mm, 8 – thermocouple on surface
|
5. Software and numerical analysis |
The 3D numerical model has been developed on the base of control volume approach and it describes the non-stationary heat conduction in the object having known geometry and thermal characteristic. The model enables that the properties of the material can be changed as well as the type of thermal stimulation and start and boundary conditions. The control volume net can be changed too which enable the setting of higher density net in the region of higher temperature gradients. For simulation the start and boundary conditions were the same as in experiments.
Material characteristic:
Figure 2. Cross section of the sample, a) phenoxy resin, b) phenoxy resin containing cooper fibres
Thermal properties of the sample containing cooper fibres are calculated using the following relations:
where j is volume portion of cooper fibres in phenoxy resin.
Number of control volumes: 37 x 32 x 14.
Dimensions of control volumes: from 1 mm to 10 mm.
Time step: from 0.02 s to 0.2 s, depending on the sample material.
Start temperature: uniform 24o C.
The boundary conditions are shown on figure 3. For both simulations the input parameter was heat stimulation calculated as power of the
IR heater reduced for the losses to the surrounding.
|
6.
Comparison the results |
On the following figures the results of the experiments and simulation are compared.
Each of them is obtained for certain time indicated on figure.
Figure 4. Left - experimentally obtained temperature distribution on the surface of phenoxy resin plate after 300 s,
Right - numerically obtained results
Figure 5. Left - experimentally obtained temperature distribution on the surface of phenoxy resin plate with cooper fibres after 600 s,
Right - numerically obtained results
|
7. Analysis of the results |
The results show how the procedure of calculating the heat conduction coefficient influences the line temperature distribution, and aberration from the experimentally obtained results.
On figure 6. the comparison of the line temperature distributions obtained experimentally and by simulation for heat conduction coefficients
l mix1, l mix2 and l mix3 are shown.
Figure 6. Comparison of the line temperature distributions obtained experimentally and numerically
|
8. Discussion of the results |
From the examples presented it is evident that the developed software and experimentaly obtained results give a good base for further analysis of the investigated problems.
It can be concluded that the simulation carried out with known thermal properties together with known process parameters gives a good results. Contrary when the properties are calculated by the law of mixtures and the uniform distribution is supposed the difference between simulation and experiment occurs.
On the base of both examples it can be seen that the developed software combined with IR thermography gives the new possibilities in measurement of the material thermal properties.
[1] D.P. Almond, P.M. Patel, Photothermal Science and Technique, London, Chapman & Hall, 1996.
[2] V. Vavilov, Transient thermal NDT: conception in formulae, Quantitative Infrared Thermography, QUIRT 92, Chatenay-Malabry, France, pp. 229-234, 1992.
[3] L.D. Favro, X. Han, P.K. Kuo, R.L. Thomas, Imaging the early time behaviour of reflected thermal-wave pulses, Thermosense XVII, An International Conference on Thermal Sensing and Imaging Diagnostic Applications, Orlando, Florida, USA, pp. 162-166, 1995.
[4] I. Boras, S. Svaic, Determination of the Defect Parameters in Specimen by Means of Thermography and Numerical Methods, Proceeding of The International Society for Optical Engineering, San Antonio, Texas, USA, Vol. 3396, pp. 271-281, 1998.
[5] I. Boras, S.Svaic, A.Galovic, Mathematical model for simulation of defects under material surface applied to thermographic measurements, Quantitative Infrared Thermography, QIRT 98, Lodz, Poland, pp. 53-58, 1998.
| 
