Validation of dynamic models of additive manufacturing parts

Metal Additive Manufacturing (AM) technologies, namely, Selective Laser Melting (SLM) and Electron Beam Melting (EBM) have become very popular in recent years. Many companies in manufacturing industry are investing in AM technologies and they are in the process of adapting AM into their manufacturing cycles. Companies are aiming to replace the conventional processes like casting and forging by AM technologies because of the advantages that AM offers. In order to achieve this and certify AM parts, substantial testing requirements have to be met, which are costly in terms of raw material and labor and time consuming. During the adaptation of AM technologies, vibration testing can be utilized to obtain representative dynamic models of AM parts. Mathematical model of the structure based on experimental route can be used to validate and update numerical models. Within the scope of this thesis, first of all, general information regarding additive manufacturing technologies is presented. Then, SLM process, which is a powder bed additive manufacturing method used in this thesis, is investigated in detail. The current problems in using this process are outlined and a literature survey is conducted regarding material properties of structures manufactured by conventional and AM technologies. At the end of the first chapter, the objectives and the scope of the thesis are described. In the second chapter, theoretical basis on vibration testing is given, including topics related to vibration measurement, identification of modal parameters and comparison and correlation. Some topics related to reliable vibration measurements such as averaging, aliasing, leakage and coherence are briefly described. Then, peak picking, circle fit and line fit methods, which are some of the basic methods of modal parameter estimation techniques, are explained. In this thesis, circle fit method in ICATS software is used in order to get natural frequencies and damping levels of structures. The third chapter presents material properties for FE models. Material properties for conventional and additive manufacturing samples are estimated using impulse excitation test method in order to provide material properties for FE models. In fourth chapter, finite element models and the corresponding results are obtained for conventional and additive manufacturing parts. FE model is developed so as to predict the first 6 elastic vibration modes. The FE models are developed using the ABAQUS software. Natural frequencies and mode shapes are obtained using FE models for both conventional and AM parts. The fifth chapter describes the vibration tests performed on structures manufactured using AM and conventional methods for comparison purposes. Experimental setup is prepared and some preliminary tests are conducted first. Natural frequencies and damping levels are obtained using vibration tests for both conventional and AM parts. In the sixth chapter, experimental and theoretical results are compared and FE model is validated for conventional part. FE models of AM parts which are manufactured using different process parameters are updated according to the natural frequencies identified experimentally. This process allowed estimation of more accurate material properties, yielding more realistic and representative FE model for AM parts. In the last chapter, some concluding remarks are made and suggestions for future work are provided. It is concluded that maximum 2.763% difference in natural frequency results between parts manufactured by conventional methods and parts manufactured by AM technology suggests that parts manufactured by AM technology can be used as counterparts for parts manufactured by conventional methods. Loss factor results show that similar or higher damping levels compared to conventional manufacturing methods can be obtained by AM technology. It is also concluded that a specific part with specific Young's modulus, damping level and density/porosity value can be manufactured by adjusting AM process parameters. It is shown in this thesis that natural frequencies decrease and damping levels increase with increasing porosity levels. For future work, it is suggested that the damping originated from welding or assembling via bolts and nuts in an assembled structure can be investigated by manufacturing an alternative "single piece assembly" using AM technology. The damping levels of this "single piece assembly" can be compared with those of the equivalent structure assembled by multiple subparts manufactured using conventional methods.