Neural network method for solving nonlocal two-temperature nanoscale heat conduction in gold films exposed to ultrashort-pulsed lasers

•Present an artificial neural network method and its algorithm for solving the NTTM in thin gold films exposed to ultrashort-pulsed lasers.•Analyze the well-posedness of the NTTM and convergence of the neural network solution to the analytical solution.•Predict the electron and lattice temperatures in gold films exposed to ultrashort-pulsed lasers. Recently, we have presented an artificial neural network (ANN) method for solving the parabolic two-temperature heat conduction equations (PTTM) in double-layered thin films exposed to ultrashort-pulsed lasers. In this article, we extend this study and develop an ANN method for the nonlocal (or weakly nonlocal) two-temperature model (NTTM), which takes into account both time and space nonlocal effects when the relaxation time becomes more effective such as in low temperature and when the characteristic length scale becomes comparable with the mean free path of heat carriers such as in nano systems under femtosecond laser irradiation where the PTTM may not be suitable. We analyze the well-posedness of the NTTM and the convergence of the ANN solution to the analytical solution. The ANN method is finally used to predict the electron and lattice temperatures in nanoscale gold films when exposed to ultrashort-pulsed lasers. As compared with the experimental data, the NTTM-based ANN method gives more accurate solutions than the PTTM-based ANN method.