Improving Fringe Projection Profilometry for in-situ Monitoring of Laser Powder Bed Fusion Processes

3D shape measurements are critical in a range of industries, ranging from manufacturing to art conservation. Many existing techniques are contact methods, which are time-consuming and material-contaminating. While non-contact technologies have their benefits, many current systems are limited in their functionality. The goal of this project is to improve a fringe projection-based system for 3D shape measurements, especially in laser powder bed fusion process (L-PBF) based metal AM. Fringe projection profilometry (FPP) is a low-cost, high-precision, and high-resolution optics-based 3D measurement technology. Several researchers have demonstrated that FPP can swiftly evaluate surface topography without contacting or interrupting the process during L-PBF. However, due to the complex physics and heterogeneous materials in L-PBF processes, the FPP method, which is based on dynamic light intensity, may result in inaccurate 3D reconstruction of printed layers and objects. Associated with the metal surface's high specular reflectivity, a glossy saturated portion is likely to present in a FPP camera image and conceal the fringe pattern, causing the height information linked with it to be lost. In this work, we propose two methods to improve FPP for in-situ monitoring of surface topography during L-PBF processes. The first method is to use a 12-step phase-shifted FPP approach instead of the conventional 3-step phase-shifted FPP method. The second method is to place an orthogonal linear polarizer in front of the camera and the projector lens, respectively, in an attempt to lessen the effect of specular reflection from the metal surface. The surface heights of printed sample blocks are measured using the proposed in-situ FPP methods as well as an ex-situ Keyence microscope. By comparing the FPP and Keyence measurement results, we can determine the influence of increasing projected fringe phase steps and that of employing linear polarizers. The experimental results show that the proposed methods can help reduce the overexposed area ratio from 24.52% to 0.28% in specific experiment setup. And the polarized FPP method could reduce the root-mean-square error from 15.75 to 10.97 in 3-step phase shifting, and also reduced the RMSE from 15.75 to 12.51 between 3-step non-polarized and 12-step with polarizers