The data was interpreted in the following article:

Machine Learning-Assisted High-Throughput Exploration of Interface Energy Space in Multi-Phase-FieldModel with CALPHAD potential

Vahid Attari, Raymundo Arroyave

Texas A&M University

Link to article: https://doi.org/10.1186/s41313-021-00038-0

For more information please refer to Open Phase-field Microstructure Database (OPMD) curated at https://microstructures.net.

The hdf5 file contains image data of inclusion based microstructured material and the homogenized effective heat conductivity thereof. The microstructure is defined with a representative volume element with periodic boundary conditions. 30.000 images are contained and split into two inclusion subclasses of circular and rectangular inclusions. Features computed via the 2-point correlation function can be found. The features and effective properties have been used for a regression problem in the related paper. Example code to access the data and recreate the features is attached.

Prediction of microstructure evolution during material processing is essential to control the material properties. Simulation tools for microstructure evolution prediction based on physical concepts are computationally expensive and time-consuming. Therefore, they are not practical when either there is an urgent need for microstructure morphology during the process, or there is a need to generate big microstructure datasets. Essentially, microstructure evolution prediction is a spatiotemporal sequence prediction problem, where the prediction of material microstructure is difficult due to different process histories and chemistry. We propose a Predictive Recurrent Neural Network (PredRNN) model for the microstructure prediction, which extends the inner-layer transition function of memory states in LSTMs to spatiotemporal memory flow. As a case study, we used a dataset from spinodal decomposition simulation of FeCrCo alloy created by the phase-field method for training and predicting future microstructures by previous observations. The results show that the trained network predicts quantitatively accurate microstructure morphologies while it is several orders of magnitude faster than the phase field method. The trained model aims to generate future material microstructures by learning from the historical microstructures, where the visual dynamics are believed to have modular structures that can be learned with compositional subsystems. A PredRNN (https://github.com/thuml/predrnn-pytorch) model was trained to predict Fe-Cr-Co microstructures evolution during the spinodal decomposition.

This dataset belongs to the paper "Solving the puzzle of hierarchical martensitic microstructures in NiTi by (111)-oriented epitaxial films" and contains all raw data used for the paper. It includes SEM, TEM, Texture measurements and inverse polfigures. It also contains the MATLAB code for calculating variant orientations, twin boundary and habit plane orientations, and inverse pole figures. Information about sample, measurement techniques and further data description can be found in README.txt.

OOF is software for analyzing the properties of materials with complex microstructures. Starting with a micrograph of a real or simulated material, OOF allows users to assign material properties to features of the microstructure, to construct a finite element mesh from the microstructure geometry, and to perform virtual experiments on the microstructure.

See paper: Sachin L. Shrestha, Andrew J. Breen, Patrick Trimby, Gwénaëlle Proust, Simon P. Ringer, and Julie M. Cairney; An Automated Method of Quantifying Ferrite Microstructures using Electron Backscatter Diffraction (EBSD) Data

This data set is a collection of multi-energy X-ray CT data and DCM generated 3D compositional microstructural data for an anthracite coal sample which are comprised with void, coal matrix, minerals. Lineage: A DCM software is required to open the data files with a “.dcm” file extension. More details about the sample and computation can be found at http://dx.doi.org/10.1016/j.fuel.2012.11.079. The DCM software is available for download at https://data.csiro.au/dap/landingpage?pid=csiro:9448, and more information about DCM is available at http://research.csiro.au/dcm.

This dataset is about book subjects and is filtered where the books is Microstructure and function of cells, featuring 10 columns including authors, average publication date, book publishers, book subject, and books. The preview is ordered by number of books (descending).

TechnicalRemarks: Scanning electron microscopy images are provided as tif files named according to the figure number in the manuscript. Short notations highlighting the sample condition are added. Additional micrographs not presented in the manuscript are highlighted by "-add". All micrographs are taken with backscattered electron contrast. Orientation imaging microscopy data by electron backscatter diffraction is provided in the form of jpg files of inverse pole figure color coded maps (for both phases BCC and C14). The legends of the inverse pole figure maps are added as well. The images are 40 x 40 µm² with 0.1 µm step size. Additionally, image quality maps were added. Number data is provided as ASCII files with tabulator separation. Files including multiple sample conditions contain columns with “condition” designation.

The data includes EBSD orientation maps of the specimens preheated at 1200 degrees celsius, and deformed at 1100 degrees celsius with 30% reduction and control cooled at the rates of 1, 18, and 95 degrees per second. The resultant microstructures correspond to quasipolygonal ferrite plus granular bainite, and lath bainite, respectively.

The data examines the design of magnesium alloys for improved ductility by the edition of rare earth elements. These elements, such as cerium and gadolinium modify the texture of wrought products and also refine the grain size.

The MICA-MICs dataset provides raw and fully processed multimodal neuroimaging data acquired in 50 healthy control participants at a filed strength of 3T. Modalities include high-resolution anatomical (T1-weighted), microstructurally-sensitive (quantitative T1), diffusion-weighted and resting-state functional imaging. In addition, MICA-MICs provides ready-to-use connectomes built across multiple parcellation schemes based on brain anatomy, function, and histology (18 parcellations in total). Processed matrices are available for each imaging modality across a range of parcellation scales.

This is the data and methodology guide for the paper " Message Traffic and Short-Term Illiquidity in High-Speed Markets", which is published in Emerging Market Review. In this paper, we use detailed message-level data from a high-speed market that flags the orders of high-frequency traders (HFTs), agency algorithmic traders, and non-algorithmic traders to show that only the unexpected part of HFTs’ net buying pressure, computed from the inflow of aggressive and non-aggressive orders, precedes increases in both immediacy costs and price impacts in the short run. Consistent with market-making theories of active risk management, updates of outstanding limit orders relate to preceding efficient price returns and enhance the overall signaling capacity of the HFTs’ order flow. Market-wide HFTs’ net buying pressure adds extra power in anticipating single-stock short-term illiquidity.

The data consists of a series of Scanning Electron Microscope (SEM) images from NiTi wire samples heat treated under different conditions, to study the effect of processing on the microstructure (mainly focusing on nanoscale precipitates) of the samples. The aim of the project is to reveal the correlation among processing, microstructure, and functional properties of NiTi shape memory wires.

Data files from the experiments presented in this publication include: raw .tif images of collected multigrain X-ray diffraction, filtered peak files (.flt), crystallographic information files (.cif), and the parameter files (.prm) necessary for grain indexing.

This dataset is about book subjects and is filtered where the books is Steels : microstructure and properties, featuring 2 columns: book subject, and publication dates. The preview is ordered by number of books (descending).

The dataset contains information on the crystallographic orientation and on the grain size of minerals in the brittle-viscous shear zones. The methodology used to generate it is polarized light microscopy, scanning electron microscopy (SEM) and electron backscatter diffraction (EBSD)

MGH Connectome Diffusion Microstructure Dataset

This repository includes the dataset described in "A multi-site, year-round turbulence microstructure atlas for the deep perialpine Lake Garda", Scientific Reports, 2021, DOI: 10.1038/s41597-021-00965-0

A multi-site, year-round dataset comprising a total of 606 high-resolution turbulence microstructure profiles of shear and temperature gradient in the upper 100 m depth is made available for Lake Garda (Italy). Concurrent meteorological data were measured from the fieldwork boat at the location of the turbulence measurements. During the fieldwork campaign (March 2017-June 2018), four different sites were sampled on a monthly basis, following a standardized protocol in terms of time-of-day and locations of the measurements. Additional monitoring activity included a 24-h campaign and sampling at other sites. Turbulence quantities were estimated, quality-checked, and merged with water quality and meteorological data to produce a unique turbulence atlas for a lake. The dataset is open to a wide range of possible applications, including research on the variability of turbulent mixing across seasons and sites (demersal vs pelagic zones) and driven by different factors (lake-valley breezes vs buoyancy-driven convection), validation of hydrodynamic lake models, as well as technical studies on the use of shear and temperature microstructure sensors.

Methods Between March 2017 and June 2018, an extensive turbulence monitoring campaign was carried out in Lake Garda (Italy) using a free-falling, internally recording microstructure profiler MicroCTD, specifically developed by Rockland Scientific International Inc. (RSI) for application in lakes, reservoirs and estuaries. Along with turbulence-related quantities, CTD (conductivity and temperature) and water quality (chlorophyll-a and turbidity) profiles, and meteorological data were measured. All vertical profiles were measured down to 100 m depth, while meteorological data were collected directly at the fieldwork boat (using a Davis Vantage Pro2 6152 meteorological station), providing representative weather conditions at the time and location of the turbulence measurements. The fieldwork activity was conducted from an inflatable rubber motorboat. A Secchi disk (black and white, 20 cm diameter) completed the fieldwork boat's onboard equipment, and the anecdotal experience of the boat captain enriched the fieldwork activity.

Four reference sampling sites were established in the northern, deep and elongated part of Lake Garda: three along a transverse transect where the lake is about 2.5 km wide (West Station - WS, Central Station - CS, and East Station - ES), and one in a sheltered bay a few kilometers (~4 km) to the south (Limone Station - LS). The four sites were sampled on a monthly basis, following a standard procedure aimed at minimizing the differences among the monitoring days in terms of scheduling of the fieldwork activity (time-of-day and monitoring sites sequence). To achieve good statistical significance, a minimum of three and up to six vertical profiles were measured at each reference site. The monitoring campaign included also other activities, such as an intensive 24-hour session in May 2018 (May 7th - 8th), and the occasional collection of vertical profiles at additional locations. These sites included the deepest point of the lake (Deep Station - DS) and a shallower site in the southeast basin, besides occasional profiles in other points of the lake (Other Stations - OS). Overall, 606 profiles were measured.

The microstructure profiler has a length of 1 m, a maximum operational depth of 100 m and is equipped with turbulence and water quality sensors located at the front bulkhead, which is protected by a sensor guard. Turbulence properties were measured with two microstructure airfoil shear probes and two fast-response temperature sensors (type FP07), sampled at high frequency (512 Hz). Water quality profiles were measured with a stable and reliable conductivity/temperature (CT) sensor and a fluorescence/turbidity (FT) sensor (JFE-Advantech Sensors). The sampling rate is 64 Hz for the CT sensor and 512 Hz for the FT sensor. Finally, a two-axis vibration sensor (i.e., a pair of piezo-accelerometers) sampling at 512 Hz and a two-axis inclinometer (pitch and roll angles) sampling at 64 Hz monitored the dynamics of the profiler flight. The buoyancy of the micostructure profiler was regulated to achieve a downward profiling speed (W) of about 0.75 m/s (specifically, W=0.74 ± 0.04 m/s, mean ± standard deviation, based on the entire dataset). This profiling speed is within the range recommended for turbulence measurements using shear probes, but higher than that typically used when employing fast-response thermistors. Here we challenged the use of the latter type of sensors, trying to find a compromise to exploit the capabilities of the two types of sensors across the wide range of turbulence intensities observed in the lake.

The processing of the microstructure data was based on the scripts provided by RSI (ODAS libraries v4.4), properly modified for the specific purposes of the analysis. The data processing workflow is described in detail in the manuscript presenting the dataset, which has been structured aimed at providing a self-contained, step-by-step reference compendium synthesizing the key procedures and the best practices available in the literature. For the sake of synthesis, here only the main steps are summarized, while we refer the interested reader to the manuscript for the full and detailed description of the data processing:

shear probes and fast-response thermistors were converted into physical units knowing the sensitivity of the shear probes and the resistance of the FP07 (after calibrating each cast against the precise CT sensor);
the vertical profile was partitioned into 3 m long, 50% overlapping segments. For each segment, the frequency spectra of shear and temperature gradient were derived by ensemble averaging the fast Fourier transform power spectra computed for 1 m long, 50% overlapping subsegments, detrended and Hanning tapered;
the measured frequency spectrum was converted into the corresponding wavenumber spectrum knowing the profiling speed and according to the Taylor's frozen turbulence hypothesis;
the raw microstructure shear signals were cleaned by despiking and high-pass filtering, corrected for the probe’s spatial response, and denoised removing high-frequency noise due to instrument vibrations using the piezo-accelerometers data. Then, the resulting shear signals were used to derive the shear power spectra;
the vertical temperature gradient signals were calculated by applying a high-pass filter to the pre-emphasized temperature signals, and the corresponding temperature gradient spectra were derived. The spectra were then corrected for the frequency response of the probes;
the despiked and response-corrected shear and temperature gradient wavenumber spectra, processed as outlined above, were used to calculate vertical profiles of dissipation rates of turbulent kinetic energy (TKE) and temperature variance using the (empirical) Nasmyth and (theoretical) Kraichnan spectra as reference, respectively;
all estimates were checked based on a number of quality metrics, and a detailed quality screening and inter-sensor cross-validation was carried out.

The Matlab scripts for estimating TKE and temperature variance dissipation rates according to the procedures described in the manuscript are provided together with the dataset. Some steps rely on functions provided by RSI through the ODAS library. We provided some alternatives/suggestions for users not having access to this library.

EBSD data sets of ice samples presented in Fan et al., 2020 (the Cryosphere paper). These ice samples were deformed under uniaxial compression with constant displacement rates (~1e-5/s) at -10, -20 and -30 °C to progressively higher strains.