Doctoral Degrees (Mechanical and Mechatronic Engineering)
Permanent URI for this collection
Browse
Browsing Doctoral Degrees (Mechanical and Mechatronic Engineering) by browse.metadata.advisor "Becker, Thorsten Hermann"
Now showing 1 - 6 of 6
Results Per Page
Sort Options
- ItemAdaptive digital image correlation using neural networks(Stellenbosch : Stellenbosch University, 2023-03) Atkinson, Devan James; Becker, Thorsten Hermann; Neaves, Melody; Stellenbosch University. Faculty of Engineering. Dept. of Mechanical and Mechatronic Engineering.ENGLISH SUMMARY: Subset size selection is crucial to the accuracy and precision of digital image correlation (DIC) measured displacements. Increasing the subset size improves noise suppression (reducing random errors) at the cost of spatial resolution (ability to accurately measure complex displacement fields). The tradition of global correlation parameter assignment is suboptimal because the speckle pattern quality and displacement field complexity can vary spatially. Dynamic subset selection (DSS), which assigns location specific optimal subset sizes, is challenging because the metrological performance of correlation is dictated by complex interactions between correlation parameters (subset size and shape function) and image set properties (noise, speckle pattern and displacement field complexity). This dissertation uses an open-source DIC framework to investigate the potential of artificial neural networks (ANNs) for error prediction and DSS, prior to the DIC process, from purely image information. ANNs are capable of modelling complex relationships within noisy, incomplete data without imposing fixed relationships, inspiring their recent resurgence for DIC applications. Despite the plethora of open-source DIC algorithms available, none offer spatially and temporally independent assignment of correlation parameters. Subsequently, a modular, open-source DIC framework capable of such flexibility is developed. This framework is predominantly consistent with current state-of-the-art practices and performs on par with well-established open-source and commercial DIC algorithms. Drawing direct links between the well-documented theory of DIC and its nuanced practical implementation, bridges this gap in literature which has acted as a barrier to newcomers intending to develop the capabilities of DIC. This framework, implemented in 117 and 202 lines of MATLAB code for 2D and stereo DIC, respectively, is attractive as a starting point to further the capabilities of DIC. The feed-forward ANN developed using this DIC framework, predicts random errors based on the speckle pattern quality (contained within a subset) and standard deviation of image noise more accurately and precisely than established theoretical derivations. A DSS framework is developed which uses this ANN to appoint subset sizes, based on the local speckle pattern, that offer random errors consistent with a stipulated threshold value. Appropriate selection of the random error threshold offers a favourable compromise between noise suppression and spatial resolution for up to moderate displacement gradients. Consequently, in the presence of varying speckle pattern quality this framework outperforms the traditional approach of trialand- error global subset size selection for the same mean subset size. Speckle pattern characteristics outside the training scope reveal the generalisability limitations of the DSS method, and associated ANN, as it performs on par with the traditional global subset size approach, motivating the need to broaden its training scope. Investigation of convolutional neural networks for dynamic shape function selection is initiated, showing they are capable of quantifying displacement field complexity between image pairs to guide spatially and temporally independent shape function assignment. The dissertation reveals that ANNs are an attractive approach to model the correlation parameter assignment. Furthermore, such models facilitate dynamic correlation parameter assignment from purely image information such that they can operate as a pre-process to DIC.
- ItemFatigue crack growth rate threshold of laser powder bed Fusion Ti-6AI-4V(Stellenbosch University, 2021-12) Dhansay, N. M.; Becker, Thorsten Hermann ; Stellenbosch University. Faculty of Engineering. Dept. of Mechanical and Mechatronic Engineering.ENGLISH ABSTRACT: Typically, producing Ti-6Al-4V through the laser powder bed fusion (LPBF) technique, results in the material having large residual stresses and martensitic microstructure. These stresses and microstructure have been shown to result in Ti-6Al-4V having poor fatigue properties. However, the insight into the fatigue failure mechanisms caused by the residual stress and microstructure has been limited. LPBF is one of many additive manufacturing (AM) techniques in which parts are built in a layer- wise manner with the use of powdered material and consolidated through high power laser melting. This allows for complex geometries and previously impossible geometries to be manufactured with minimal material wastage. Many industries are aware of the potential in this manufacturing technique and have shown interest in it becoming a viable option for the manufacturing of some of their components. In particular, the use of LPBF produced Ti-6Al-4V is of interest to the aerospace and biomedical industries. This is because Ti-6Al-4V is already well established in many existing industries. For LPBF produced Ti-6Al-4V to be a viable option in industry, researchers need to have an adequate understanding of the material and the implications for its mechanical properties. Fatigue property investigations have largely focused on the fatigue life approach (crack initiation) and Paris regime (region II, crack propagation). However, in recent years, the near-threshold regime (region I, crack propagation) has become of interest, albeit limited in approach. The consensus within literature shows that the large tensile residual stresses, martensitic microstructure and porosity results in poor fatigue properties. Unfortunately, the insight into the fatigue fracture mechanisms brought about by the residual stress and microstructure is not yet well established. Furthermore, the near-threshold fatigue crack growth rate regime (FCGR) experiences crack closure mechanisms which result in premature near-threshold values. The majority of the near-threshold investigations on LPBF produced Ti-6Al-4V do not account for the crack closure mechanisms and therefore produce premature results. As a result of the low crack growth rates achieved in the near-threshold regime, a window into observing the fatigue crack initiation mechanisms is obtained. More specifically, how the fatigue crack initiation mechanisms are influenced by residual stress, martensitic microstructure and changing microstructural morphology. Literature has shown that porosity acts as crack initiation sites and reduce the fatigue life of a component. Furthermore, surface and near-surface porosity have been shown to have a more severe impact on fatigue life than internal porosity. It is through using the near-threshold FCGR approach in which one can calculate the allowable pore size under operational loads. Using the load-shedding technique to obtain near-threshold FCGRs, the results showed anisotropic behaviour dependent on residual stress levels and R-ratios. The fatigue fracture mechanisms were predominantly governed by transgranular quasi-cleavage mechanisms. Furthermore, the fracture is directed by PBG morphology which results in anisotropic crack- closure effects. In addition, primary ฮฑ lath orientation is governed by the crystallographic texture of the PBG which influences the mechanisms of crack propagation. With the increase in grain size, the presence of ฮฒ improves the near-threshold FCGRs in the duplex anneal (DA) condition due to the superior plastic flow abilities more than in the as-fabricated and stress relief conditions. This study investigates the near-threshold FCGRs of LPBG produced Ti-6Al-4V in the as- fabricated, stress relief and bi-modal conditions in three build orientations. In addition, crack closure mechanisms are accounted for by implementing variable R-ratio testing. This research presents the influencing mechanisms of residual stress and microstructure on fatigue behaviour.
- ItemFracture mechanics-based fatigue life assessment of additively manufactured Ti-6Al-4V(Stellenbosch : Stellenbosch University, 2024-02) Macallister, N; Becker, Thorsten Hermann ; Blaine, DC; Stellenbosch University. Faculty of Engineering. Dept. of Mechanical and Mechatronic Engineering.ENGLISH ABSTRACT: This dissertation presents a study on fracture mechanics-based fatigue life assessment for Additively Manufacturing (AM). The mature laser powder bed fusion (LPBF) process with the Ti-6Al-4V alloy in particular is selected for study, as it is well suited to the South African context with regard to economic climate, strong AM relationships and abundant mineral titanium reserves available. Furthermore, the Ti-6Al-4V alloy is a staple of aerospace, automotive and biomedical industries which are amongst the largest promoters for using AM technology, and for whom fatigue characterisation remains a prevalent topic as many end-use applications are intended for cyclic loading. Though significant research in fatigue behaviour exists, the conundrum of reliably certifying fatigue life in AM parts persists. This problem stems from the complex relation between AM print parameters, build orientation, surface roughness, inherent defects, residual stresses, meso- and microstructure; and establishing reportable fatigue strength baseline values required by industry. Moreover, as the AM environment promises saving in cost and time, full fatigue testing schemas are undesirable. As such, alternate damage-tolerant methods are becoming increasingly popular, where adopting fracture mechanics-based frameworks accompanied by limited or non-destructive testing could aid in certification. For this purpose, the dissertation first presents a novel version of the fatigue predictive NASGRO model where parameters are established that are unique to LPBF produced Ti-6Al-4V meso- and microstructures. In establishing these parameters for LPBF produced Ti-6Al-4V, the influence of process inherent microstructure, residual stress, and orientational dependant meso- structure is considered through examining near-threshold in combination with steady-state fatigue crack growth rates. The analysis shows that the descriptors of material constraint are sensitive to build orientation and microstructure. Furthermore, the effect of residual stresses is observed to not be severe. In this a clear effect of build orientation and meso- and microstructure is established for selecting NASGRO model parameters. Secondly, the proposed NASGRO formulation is translated into a comprehensive novel damage-tolerant fracture mechanics-based model to estimate fatigue life. Non-uniform defect populations, typical of AM material, in terms of size, shape and location are captured through X-ray tomography and surface profilometry and used as inputs modelled as equivalent crack lengths. The fatigue strength estimations are shown to be sensitive to fatigue crack growth rate threshold parameters and short crack growth mechanic descriptions. Furthermore, by introducing multiple crack initiations, the fatigue estimates are shown as distributions and are sensitive to defect number. Finally, sub-size specimen testing is investigated as a potentially elegant solution to accompany fatigue life assessments for threshold validation. Where results show inconsistent near-threshold fatigue behaviour linked to the microstructure. In this, considering unique meso- and microstructural features of LPBF produced Ti-6Al-4V, the domain and suitability in using sub-size specimens for fatigue crack growth rate threshold testing is discussed. Overall, this dissertation walks the path required in establishing reliable damage tolerant fatigue life estimation approaches for LPBF produced Ti-6Al-4V. Providing fundamental insights into interactions of fracture mechanic mechanisms and descriptions necessary for reliably modelling fatigue behaviour, therefore contributing to the developing frameworks and philosophies in AM to help in certification of fatigue performance of LPBF produced Ti-6Al-4V components.
- ItemLaser powder bed fusion produced Ti-6AI-4V: microstructural transformations and changes in deformation behaviour through thermal treatments.(Stellenbosch : Stellenbosch University, 2021-12) Ter Haar, G.; Becker, Thorsten Hermann ; Stellenbosch University. Faculty of Engineering. Dept. of Mechanical and Mechatronic Engineering.ENGLISH ABSTRACT: The current 4th Industrial Revolution has brought about enormous potential for social-economic development in South Africa. Additive manufacturing, a key technology of the 4th Industrial Revolution, and titanium, an abundantly mined but under-developed mineral commodity, collectively inspired industrial embracing of titanium-powder additive manufacturing, and specifically, laser powder bed fusion produced Ti-6Al-4V. Manufactured part quality concerns and the limited understanding of the innovative process-structure-property interactions impede its industry acceptance. Through developing fundamental insights into interactions between the microstructure and thermal treatments, mechanical performance can be improved, tailored, and optimised. Since conventional process routes cannot be applied to additively manufactured components, novel thermal treatments need to be developed. This dissertation documents experimental research into thermal treatments of laser powder bed fusion produced Ti-6Al-4V. By developing novel insights into the materialโs unique metallurgical response to thermal treatments, tensile behaviour of the material is improved. The research investigates three temperature regions for thermal treatments post-fabrication and one in-situ approach. Thermal treatments between 750 โ 960 ยฐC are used to develop insight into the microstructureโs large-scale morphological transformation. Thermal treatments at 960 ยฐC achieve fragmentation and grain globularisation. This is followed by quenching to attain a superior bi-modal microstructure. Thermal treatments at temperatures below 650 ยฐC are used to develop insight into the initial stages of martensite decomposition phase transformation and material stress relief. Microhardness and tensile properties revealed material embrittlement. Fine precipitates in the microstructure were identified using high-resolution transmission electron microscopy. Based on these findings, two theories for the cause of material embrittlement based on two possible transformation routes of the initial stages of martensite decomposition are proposed. Mechanical properties of laser powder bed fusion produced Ti-6Al-4V depend on build orientation. Although the unique columnar-shaped and textured prior-ฮฒ microstructure is identified as a probable cause of deformation anisotropy, limited insight into the cause of anisotropic deformation exists. Thermal treatments above 975 ยฐC are used to globularise the columnar prior-ฮฒ grain morphology. Microstructural anisotropy between two orientations is quantified using electron backscatter diffraction maps, and the influence of microstructure on deformation and crack initiation is studied. Insight into the deformation behaviour and the identified relation between prior-ฮฒ crystallographic texture and ฮฑ-lath morphological texture is used to formulate a theory of the probable cause of material deformation anisotropy. The study lastly investigates a novel approach to thermal treatments by using high-energy process parameters to achieve in-situ heating. An iterative approach for part optimisation using non-default process parameters is undertaken. While findings indicate that an improved microstructural and residual stress state can be achieved, detrimental effects of part oxidation and part-edge bulging are also observed.
- ItemMeasuring fracture properties using digital image and volume correlation: decomposing the J-integral for mixed-mode parameters(Stellenbosch : Stellenbosch University, 2017-12) Molteno, Matthew Robert; Becker, Thorsten Hermann ; Stellenbosch University. Faculty of Engineering. Dept. of Mechanical and Mechatronic Engineering.ENGLISH ABSTRACT: The degradation of material properties over time is one of the core constituents of structural integrity monitoring. For this purpose, the materialsโ resistance to brittle fracture is the most widely used property in the determination of safe operation and predicting component life. Measuring such properties by conventional methods presents challenges: large numbers of samples are needed, accounting for mixed-mode loading. Non-contact measurements by digital image correlation shows promise in extracting such parameters from in situ loaded structures, accounting for their complex geometry and loading. This thesis aims to develop a framework that allows for the extraction of mixed-mode brittle fracture properties, using three-dimensional (3D) image correlation techniques: multi-camera Digital Image Correlation (stereo-DIC), and Digital Volume Correlation (DVC). The ๐ฝ-integral method was chosen for obtaining the fracture resistance, with the decomposition method used for determining separate ๐ฝ๐ผโ๐ผ๐ผ๐ผ components (opening, sliding and shearing: modes I, II and III), and stress intensity factors (๐พ๐ผโ๐ผ๐ผ๐ผ), as both methods are resilient to crack tip and displacement errors. The first stage of development is a finite element based method for the removal of experimental displacement artifacts from DIC or DVC results, intended as a pre-processor for the ๐ฝ-integral. Subsequently, a theoretical derivation is provided to link the two volumetric forms: the volume and the path-area integrals. This results in a proposed โhybridโ integral, which benefits from the resilience to crack-front positional errors of the path-area approach, and resilience to random displacement noise of the volume integral. Initial testing of the ๐ฝ-integral based decomposition method was on the surface using stereo-DIC and the Arcan fixture to induce mixed-mode loading. The results are verified with mixed-mode fracture toughness values measured from PMMA and compared to literature and ASTM 1820 tests. The typical image correlation errors close to fracture are quantified by a methodology of applying the ๐ฝ-integral to analytical crack tip fields to which displacement errors are added artificially. It was found that the ๐ฝ-integral is most prone to DIC errors under anti-symmetric (mode II/III) loading. Testing in the volume used X-ray computed tomography to acquire images and DVC for displacement maps. The measurements were verified on two configurations: a SENT polyurethane composite specimen (mode I), and a shear loaded inclined notch in Magnesium alloy-WE43 (mixed-mode). The ๐ฝ-integral was verified against values from finite element fields resulting again in larger errors in mixed-loading. Decomposition of the volume integral requires an approach to separate the anti-symmetric ๐ฝ-integral. Two extensions are proposed: the first using a ratio derived from mode II/III Williams series formulas, and the second using the interaction integral. Both approaches are verified on DVC displacements. This thesis finds that ๐ฝ based decomposition offers a versatile method to extract ๐พ๐ผโ๐ผ๐ผ๐ผ values from non-standard 3D geometries and loadings. However, the pre-processing of fields to minimise errors is essential when mode II-III displacement fields are prominent. Although it is beyond the scope of this thesis, the hope is that this work will assist in the adoption of full-field measurement techniques as a standard testing practice in structural integrity assessments, and lead to better informed maintenance and inspection schedules.
- ItemThermal power plant steel creep deformation measurement using digital image correlation(Stellenbosch : Stellenbosch University, 2020-12) Van Rooyen, Melody; Becker, Thorsten Hermann ; Stellenbosch University. Faculty of Engineering. Dept. of Mechanical and Mechatronic Engineering.ENGLISH ABSTRACT: Following the rapid expansion of populations and industries with a concomitant rise in the demand for electricity in South Africa, increased focus has been directed to the structural health monitoring of critical components of aging thermal power plants. Long-term operation at high temperatures and loads encountered by plant components results in the degradation of material properties through creep exposure. Assessing creep damage using conventional creep testing poses several challenges in terms of large sampled material requirements for manufacturing standard specimen geometries. This is often not possible for service-retrieved material due to the limited availability for destructive sample removal. This dissertation documents a novel experimental technique that employs digital image correlation (DIC) to characterise the creep damage of service-exposed power plant steel through creep deformation measurement across nonuniform temperature and stress fields. Through the full-field property of DIC, multiple properties are measured from single specimens ofX20CrMoV12-1 (X20) piping steel supplied by Eskom in virgin and various stages of service exposure. Initial development of the technique involves a stereo-DIC setup adapted for high temperature deformation measurement.A temperature profile ranging from 550โ600ยฐC is applied using resistive heating from a Gleeble thermo mechanical simulator and measured using infrared imaging.Near-uniform stress profiles result from appropriate specimen geometry and region of interest selection. Several temperature-dependent elastic moduli and Poissonโs ratios are measured from single specimens of virgin X20. Creating an extension to accelerated creep tests(shorter-term tests conducted at higher stresses and temperatures than encountered in service)used on virgin and various ex-service X20 states is explored.Creep curves at several temperatures are successfully realised from single specimens and are useful for identifying differences in creep resistance between ex-service levels in line with traditional damage classification through surface cavity density replication. Micro structural quantification using electron microscopy shows subgrain and M23C6precipitate growth as well as Laves phase precipitation in higher damage states of X20. For middle damage classes, it is highlighted that several complementary damage assessment methods are necessary to identify differences in material deterioration. Zener-Hollomon parameters calculated from DIC-measured strains over the temperature profiles serve as suitable damage indicators. Single camera DIC measurement of creep curves across spatially varying stress profiles is achieved with a waisted specimen design in a traditional creep testing setup for medium-term tests. Virgin X20 creep curves at 140โ150MPaare used to calibrate baseline parameters for a continuum damage mechanics model. Subsequently, the model is optimised for subgrain and precipitate parameters using the corresponding ex-service X20 creep responses. These micro structural-based parameters serve as comparative damage sensors that agree with traditional cavity and hardness-based methods. This dissertation reveals that a wealth of creep data can be extracted from fewer X20 specimens, which is highly beneficial in characterising material integrity in a supplementary manner to existing methods.Expansion to small samples through small punch creep testing is also initiated. Although beyond the scope of this work, future aspirations are that these approaches will be integrated into life management philosophies to better guide inspection and maintenance strategies.