PSI - Issue 72

E.B. Galkina et al. / Procedia Structural Integrity 72 (2025) 222–228

223

One of the challenges in additive manufacturing is the formation of residual stresses and strains that develop in products during the final stage of fabrication. Their presence can degrade mechanical properties and cause distortions in the shape and dimensions of the finished part. In this regard, a significant part of scientific research focuses on identifying the factors influencing the distribution and magnitude of residual stresses and strains, as well as developing methods for their prediction and mitigation. For instance, in the study by Xie et al. (2022), the authors analysed over 160 scientific papers on residual stress formation, the warping effect, and possible mechanisms governing their evolution in additive manufacturing. It has been observed that the primary cause of residual stresses is the shrinkage of newly deposited material layers, which are constrained by the underlying layers or surrounding material. Additionally, warping is most pronounced at the outermost edges of manufactured components. Various numerical and analytical models are being developed to predict residual stresses and strains. A review article by Al Rashid and Koç (2021) examines studies focused on modelling the printing process. A series of studies by Samy et al. (2021, 2022a, 2022b) devoted to modelling the FDM/FFF printing process using the finite element method, considering the interplay between material crystallization kinetics, viscoelastic and thermomechanical properties, and temperature variations during production should be noted. The developed model not only demonstrated strong quantitative agreement with experimental data but also enabled a significant reduction in residual stresses and strains by optimizing printing parameters Machine learning methods have recently been widely applied to various engineering problems, offering a potential alternative to costly field experiments and computationally intensive numerical modelling. The use of the K-Nearest Neighbors (KNN) method for predicting the warping effect was studied in the work by Song et al. (2020). A model trained on data from four temperature sensors achieved an accuracy of over 80% in detecting sample warping using only 20% of the experimental data. In addition to developing various mathematical models, efforts are also being made to enhance FDM technology. For example, in an experimental study by Han et al. (2021), a mechanism for additional laser heating of the deposited layer was integrated into the printer design to mitigate residual stresses. This additional heating promoted greater stress relaxation in the previous layer, resulting in a more than 10% increase in tensile strength and nearly a threefold improvement in elongation before fracture. For studying the FDM/FFF printing process, particularly for monitoring the stress-strain state of the material both during printing and throughout the operation of the final product, fiber-optic sensor (FOS) technology presents a promising solution. These sensors can be embedded in the material during production, enabling real-time monitoring. Examples of studies utilizing FOS for investigating residual stresses and strains include the work by Chen et al. (2021), where the authors used fiber Bragg gratings (FBG) to measure residual stresses in a circular plate, and Wang and Lasn (2022), which focused on evaluating the effect of layer thickness on the distribution of residual strains in PA6/CF samples using Optical Backscatter Reflectometry (OBR). Notably, in (Wang et al. (2020)), the authors used OBR to assess the influence of PLA sample dimensions on residual strain magnitude. The results indicated that sample length and width had minimal impact on the optical system's readings. However, noise was observed at the fiber entry and exit points, imposing limitations on the minimum sample length. This paper presents the results of residual strain measurements during the fabrication of samples from various materials commonly used in FDM 3D printing. Residual strains were recorded using distributed FOS based on Rayleigh scattering. Additionally, for the studied materials, the distribution of residual strains across the thickness of the fabricated samples, shaped as rectangular parallelepipeds, was analysed. 2. Materials and methods The samples were fabricated using FDM/FFF 3D printing with Designer XL Pro S2 3D printer by Picaso. This device features a closed chamber that maintains a stable temperature regime during printing, ensuring gradual cooling of the finished product and preventing sudden formation of residual strains that could distort the final geometry. The 3D printer and the FDM/FFF printing scheme are shown in Fig. 1. The manufactured samples have a rectangular parallelepiped shape with overall dimensions of 200×15×10 mm. For embedding FOS during sample fabrication, the model includes technological through-holes designed to accommodate the dimensions of the optical fibers. The sample geometry enables measurement of both strain and temperature, allowing for temperature compensation of fiber-optic strain sensors and assessment of residual strain distribution along the sample’s length and height. The sample geometry and through-hole dimensions are shown in Fig. 2.