A Deep Learning Framework for Full-Field Thermal Field Distribution Prediction from Digital Image Correlation Strain Measurements

Some pipelines face global plastic straining due to the nature of their installation process or harsh environmental conditions during operation. The ability of the girth welds to withstand these plastic strains is often evaluated on the basis of wide plate tests. Key for the validity of these tests is a representative measurement of remote strain, mostly obtained by linear variable differential transformers and/or strain gauges. The outcome of the remote strain measurement depends on the specimen geometry and the position of these sensors. In an attempt to investigate a specific geometric design of wide plate specimens and to find appropriate remote strain sensor positions, the authors have performed a series of tension tests on medium-sized wide plate specimens, supported by digital image correlation strain measurements. In addition, finite element simulations have been performed to evaluate whether the experimental observations can be extrapolated to a wider range of conditions. The results indicate that the strain distribution is mostly influenced by the weld strength mismatch, which governs the lateral restraint. For all experiments and simulations, nevertheless, the strain field was highly uniform in an identified zone, resulting in simple guidelines regarding specimen geometry and sensor positioning.

An approach has been developed to use the strain sensed in a membrane wing to estimate aerodynamic loads. Through experimental and theoretical correlation of membrane strain state due to a normally distributed load, the pressure over the surface of the wing was estimated. Elastic deformations and strains of the membrane wing were measured in a low-speed wind tunnel using digital image correlation. Simultaneously, aerodynamic loads were measured using a six-component load cell, inclinometer, and pitot tube. Digital image correlation displacement measurements were used to form a reduced-order model of the membrane displacements via a snapshot proper orthogonal decomposition method. Digital image correlation strain measurements were used to form virtual strain sensors, which served as the sensory input to the estimation. The Poisson equation for a two-dimensional linear-elastic membrane with out-of-plane deformation was used to calculate the normal pressure distribution from virtual strain sensors using proper orthogonal decomposition basis functions and a recursive least-squares minimization. Estimated pressure distributions were compared with a high-fidelity three-dimensional computational fluid dynamic model and pressures were calculated from digital image correlation deformations. Coefficients of lift and pitching moment for steady-state flow conditions were estimated and compared with measured wind-tunnel loads. Results show promise toward the application of a low-fidelity estimation approach for real-time load-estimation applications.

Strain concentrations in BCC micro lattices obtained by AM

Low-cycle fatigue properties and fracture location transition mechanism of dissimilar steel welded joints in towers of wind turbines

A two-dimensional (2D) phosphorescence-based method is presented for simultaneous measurement of strain and temperature. 6 μm Mg4FGeO6:Mn4+ phosphor particles were sprayed on the surface of the test target for use as a sensor. A measurement system was established, which included a high-speed camera, a 385 nm LED, and a signal generator. The signal generator controls the LED to output a 100 Hz light pulse to excite the phosphor. The phosphorescence is then detected by the high-speed camera with a frequency of 10 kHz. The temperature information is obtained from the phosphorescence decay rate using the lifetime method. A digital image correlation method was then used with the random markings formed by spraying. The 100 Hz temperature and strain measurement technique was implemented based on the current experimental conditions with an accuracy of 2% in temperature measurement and 16% and 9.3% in u- and v-direction respectively in strain measurement. The measurement procedure was used for measurements of temperature and deformation of an aluminum plate with a pulsed current in a temperature range of 0 °C–600 °C and a displacement range of 0–300 μm. The results from experiments show the possibility of 2D measurements of surface temperature and strain in a harsh environment with a non-invasive method.

The VDA238-100 standard for tight radius bending (v-bending) of sheet materials has received widespread acceptance with automotive suppliers and material producers to characterize local formability. However, the test fixture and tooling in the v-bend test standard is not amenable to direct strain measurement and the operator cannot terminate the test at the onset of crack initiation as the outer bend surface is not visible. Consequently, fracture is identified using a load threshold and the bend angle estimated from an analytical formula based upon the punch displacement and tooling geometry. Bend angles are not directly transferable and must be interpreted relative to the sheet thickness and bend radius unlike a strain measurement. By obtaining an in-situ strain measurement on the surface using digital image correlation (DIC), the plane strain fracture limit can be accurately identified at the onset of cracking and remove ambiguity in translating the bend angles to practical forming operations and simulations. A novel inverted VDA test frame was developed to incorporate DIC strain measurement during the bend test and a variety of advanced high strength sheet materials were evaluated. It was observed that the VDA bend test creates a homogeneous strain state of plane strain across the width of the sample along with a proportional strain path to fracture without necking that is ideal for fracture characterization. A correlation is developed to relate the bend angle with the major strain for the materials considered and accounts for the sheet thickness and bend radius. A comparison of the bend angle obtained using the formula in the VDA standard based on the punch displacement was in very good agreement with manual measurements and an algorithm to measure the bend angle using DIC analysis was developed.

Welding is one of the most widely used joining techniques for manufacturing complex engineering components. However, due to transient temperature cycles in the adjacent materials being joined, the local microstructure is altered and plastic strain in the surrounding materials is introduced resulting in a variation of mechanical properties across the weldment. When an extracted cross-weld test specimen is subjected to an external uniaxial load, there is interaction between the various constituents of the weld (such as the weld fused zone, heat affected zone and parent metal) resulting in development of constraint in the weaker material due to the adjacent stronger material. This affects the distribution of stress along and within the test specimen. The level of constraint developed is dependent on the strength mismatch between adjacent materials which may be graded across the heat affected zone and the specimen cross-sectional geometry. With development of digital image correlation (DIC) and its application to measuring deformation of solids, it has become feasible to measure local stress-strain properties of cross-weld samples and also specimens of variable geometry at both room and high temperature. Local deformation measured by DIC is correlated with nominal stress based on an iso-stress assumption, where it is assumed that the various regions of the weld are arranged in series and that the stress is uniform over the cross-section area at any given location in the specimen. Thus, local mechanical properties are inferred by mapping the measured local strain against the global stress. In this context the effect of constraint is important because it introduces a local stress distribution that will differ from the global net section stress and can therefore introduce errors in determination of local stress-strain properties. The influence of these factors raises questions concerning the interpretation of apparent local mechanical properties of cross-welds measured using DIC. Specifically, there is uncertainty as to whether the measured variation in local strain versus average cross-sectional stress (average global stress) represents the local mechanical properties of the material with negligible error. The severity of potential constraint errors will be influenced by the specimen geometry (particularly the tensile specimen thickness to width ratio for 2D DIC strain measurements) and the gradient of material property inhomogeneity along the specimen gauge length. In the current study, measurement of deformation in varying stress fields using DIC was carried out. Two distinct methods of introducing varying stress fields were employed: in the first case, the stress field was varied by designing an hour-glass shaped specimen with the specimen width varying as a function of the gauge length. In the second case, stress fields were varied due to a change in the specimen thickness in mismatched dissimilar metal joints. To reduce the complexity in the developed stress fields in the region near the interface, two materials were joined using solid state diffusion bonding giving a step change in the material properties across a planar interface. Test specimens spanning the interface were studied by finite element modelling and experimental tests with DIC deformation monitoring. Using DIC and a single hour-glass shaped specimen, multiple creep curves at different stress levels along the specimen gauge length have been successfully extracted for a test temperature of 525oC. The creep data obtained at 525oC from the hour-glass specimen test have been fitted to RCC-MR, Garofalo and Graham-Walles creep deformation models. A better correlation is achieved between the experimental data and the RCC-MR model predicted values when the local deformations from the single specimen are plotted as a function of time and true stress at the end of the specimen’s initial loading. From both elastic-plastic FE models and experimental local stress-strain curves extracted from DIC measurements, it can be concluded that a thin (1mm thick or less) 6mm wide test specimen should be used to extract the local mechanical properties from tensile specimens that have varying material properties along the gauge length, such as found in cross-weld test specimens. This is because the thinner specimen reduces the size of the strain concentration region adjacent to the interface where errors in measured local mechanical properties can occur. As a ‘rule of thumb’, the intrinsic stress-strain properties of the material should not be inferred from regions within 1mm of the interface for these 1mm thick samples because of concentration errors. The results obtained demonstrate the capability of DIC as a measurement technique for characterising the local deformation of specimens with varying stress fields. The findings also highlight some aspects of the technique which need to be considered when interpreting data from DIC measurements of local deformation from specimens with varying stress fields.

Simultaneous measurement of microscale fluid viscosity, temperature, and velocity fields by tracking Janus particle on microparticle image velocimetry

Critical comparative assessment of digital image correlation vs contact extensometry: ensuring the fidelity and depth of mechanical properties determination

Influence of image resolution on the performance of remote breathing rate measurement using thermal imaging technique

Simultaneous strain and temperature distribution sensing using two fiber Bragg grating pairs and a genetic algorithm

Assessment of digital image correlation for tensile characterization of fique woven textiles used as matrix-reinforcement

High-temperature mechanical tests coupled with Digital Image Correlation (DIC) on ceramics which exhibit rather low level of strain require to overcome extreme experimental conditions that usually can reduce significantly measurement accuracy. Thermal resistance of speckle pattern, black body radiation and heat haze are three main concerns, which should thus be taken into account while designing a high-temperature image acquisition setup. In this aim, an experimental procedure has been specifically designed in order to minimize the three above-mentioned disturbances. The main objective of this study is to select a suitable high-temperature resistant speckle pattern for mechanical characterization at 1200 °C (or above) on refractory ceramics. Most of tested speckle patterns were performed with white alumina adhesive and dark ceramic grains (silicon carbide or brown fused alumina). Different grain sizes of silicon carbide were tested. At first, different speckle patterns are compared in terms of DIC strain measurement uncertainty by discussing speckle features, some main DIC parameters and two image pre-treatments (low pass filter, image size reduction). Then, these speckle patterns are tested to analyse fracture behaviour of refractories through a Brazilian test. An enhanced digital image correlation technique (2P-DIC), dedicated to monitor the fracture behaviour, is applied to study the evolution of crack length. The best representation of crack progression has been achieved for sample surface covered with a fine SiC powder ranging from 50 to 100 µm. It is then possible to compare the fracture behaviour between 1200 °C and 20 °C and to show that the refractory exhibits more crack branching at 1200 °C in comparison with behaviour at room temperature.

The measurement of thermo-mechanical deformation in C4 interconnect and chip scale packages is a critical metrology problem for IC packaging. The ability to directly measure the deformation of the actual package, without applying special structures for metrology, is a major advantage of optical digital image correlation technology. Cutting edge automated digital image correlation technology allows in-plane deformation and strain to be measured directly at multiple locations on the bare cross-section or exterior surface of an IC package. In this paper the authors have shown the advantages of using digital image correlation based techniques for the measurement of advanced electronic packaging materials and components. The technique has moved on with the increasing power of computers which enable the required high intensity numerical calculations to be computed. The technique has been applied to the measurement of C4 based components. Highly accurate deformation measurements are possible with accuracy of 20 nm deformation resolution being possible. Minimal sample preparation was required with samples only requiring a minimal amount of roughening with silicon carbide paper. This compares favourably with comparable techniques such as moire interferometry which require a large amount of time-consuming sample conditioning and grating replication before any measurement can be made

Deep DIC: Deep learning-based digital image correlation for end-to-end displacement and strain measurement