Online heat flux estimation using artificial neural network as a digital filter approach

Surface heat flux estimation using temperature measurement data from the interior points is known as inverse heat conduction problem (IHCP). Several methods have been developed as solution techniques for IHCP's including analytical and numerical approaches. Digital filter representation for IHCP solution (Woodbury and Beck, 2013; Beck et al., 1985) is one of the methods which can be used for near real-time heat flux estimation. In this study, artificial neural network (ANN) is utilized as a digital filter, for near real-time heat flux estimation using temperature measurement data. Considering temperatures as the inputs and heat flux as the output, the weights can be interpreted as filter coefficients. The proposed approach is used for both constant and temperature dependent material properties. The method developed is tested through several test cases using exact solutions and numerical models. The results show that ANN can be used as a digital filter method for near real-time surface heat flux estimation. The advantages and disadvantages of the method are also discussed.