Comparison between a neural model and a thermographic model in estimating the thickness of scale in water pipes

Abstract Cracks in metallic structures used in nuclear power plants may occur because to their continuous operation. To avoid accidents that will have serious consequences for the environment it is necessary to carry out non-destructive tests in the context of preventive maintenance. In this work, we present a simulation based on the 3D finite element method to detect the scale presence in a steel water pipe. We studied the effects of thickness and diameter of the pipe on the tartar detection. Then we studied the pipe thermal response as a function of scale thickness. Using the absolute thermal contrast, we made a comparison between a neural model and a thermo graphic model to estimate the scale thickness in steel water pipes. We found that the neural model results are better than the thermo graphic model.