A materials compatibility and thermal stability investigation was conducted using five common liquid hydrocarbon fuels and two structural materials. The tests were performed at the NASA Glenn Research Center Heated Tube Facility under environmental conditions similar to those encountered in regeneratively cooled rocket engines. Scanning-electron microscopic analysis in conjunction with energy dispersive spectroscopy (EDS) was utilized to characterize the condition of the tube inner wall surface and any carbon deposition or corrosion that was formed during selected runs. Results show that the carbon deposition process in stainless steel tubes was relatively insensitive to fuel type or test condition. The deposition rates were comparable for all fuels and none of the stainless steel test pieces showed any signs of corrosion. For tests conducted with copper tubing, the sulfur content of the fuel had a significant impact on both the condition of the tube wall and carbon deposition rates. Carbon deposition rates for the lowest sulfur fuels (2 ppm) were slightly higher than those recorded in the stainless steel tubes with no corrosion observed on the inner wall surface. For slightly higher sulfur content (25 ppm) fuels, nodules that intruded into the flow area were observed to form on the inner wall surface. These nodules induced moderate tube pressure drop increases. The highest sulfur content fuels (400 ppm) produced extensive wall pitting and dendritic copper sulfide growth that

Load magnitude and location estimation on additively manufactured circular structures using deep learning Additively Manufactured (AM) metal parts have been used at critical engineering applications. Estimation of load and its location is important to avoid accidents. Since the excitation signals move through multiple paths, estimation of load and its location is difficult. In this study, surface response to excitation method (SuRE) was implemented with multiple steps by considering the challenges of the geometry. Two 3D printed stainless steel rocket nozzle type structures with two different sizes and a C shaped tube section were used. Two different PZT placements were used for the big nozzle while the PZTs were located at the opposite sides of the small nozzle. The large nozzle had 280 test cases for 9 categories, while the smaller nozzle had 60 test cases for 5 categories. Finally, the C shaped specimen had 48 test cases for three categories. Continuous wavelet transform (CWT) was used to obtain more representative presentation of the data to the deep learning algorithm. Convolutional Neural Networks (CNNs) were used for classification.