Heat Flux Model Validation Utilizing Machine Learning and Sub-surface Thermocouples for NSTX-U Plasma Facing Components

     A proof of concept convolutional neural network (CNN) has been developed to assist in operating tokamaks outside of existing empirical scalings for the heat flux width, l_q.  NSTX-U has designed new plasma facing components (PFCs) to withstand increased halo current forces as well as elevated heat fluxes driven by increased B_p and P_NBI compared to NSTX.  Larger graphite tiles are castellated to ~2.5 cm x 2.5 cm to reduce bending stresses. Maintaining PFCs below engineering limits will be an important consideration for operation of NSTX-U.  Sub-surface temperature transducers (thermocouples) will be utilized to demonstrate validation of the heat load model, using the castellated designs to quantify the shot-integrated energy deposited in the NSTX-U divertor.  A CNN has been trained using ANSYS simulations of PFC response to a variety of time-varying heat flux profiles.  In practice the CNN will accept time evolving thermocouple data and various 0-D engineering parameters and output the heat flux model parameters, such as the B_p scaling of l_q.  The CNN enables satisfactory validation of the heat flux model, despite a limited number of simulated NSTX-U shots, expected noise, and systematic errors in the thermocouple data.