College of Engineering (CoE)
http://hdl.handle.net/1853/5985
For more than 110 years, Georgia Tech has been producing engineers. Today, we are recognized as one of the nation's top ranked engineering colleges.2017-02-24T03:57:52ZMaterial Recognition from Heat Transfer given Varying Initial Conditions and Short-Duration Contact
http://hdl.handle.net/1853/56458
Material Recognition from Heat Transfer given Varying Initial Conditions and Short-Duration Contact
Bhattacharjee, Tapomayukh; Wade, Joshua; Kemp, Charles C.
When making contact with an object, a robot can
use a tactile sensor consisting of a heating element and a
temperature sensor to recognize the object’s material based on
conductive heat transfer from the tactile sensor to the object.
When this type of tactile sensor has time to fully reheat prior to
contact and the duration of contact is long enough to achieve
a thermal steady state, numerous methods have been shown
to perform well. In order to enable robots to more efficiently
sense their environments and take advantage of brief contact
events over which they lack control, we focus on the problem
of material recognition from heat transfer given varying initial
conditions and short-duration contact. We present both modelbased
and data-driven methods. For the model-based method,
we modeled the thermodynamics of the sensor in contact with
a material as contact between two semi-infinite solids. For the
data-driven methods, we used three machine learning algorithms
(SVM+PCA, k-NN+PCA, HMMs) with time series of raw temperature
measurements and temperature change estimates. When
recognizing 11 materials with varying initial conditions and 3-
fold cross-validation, SVM+PCA outperformed all other methods,
achieving 84% accuracy
2015-01-01T00:00:00ZBhattacharjee, TapomayukhWade, JoshuaKemp, Charles C.When making contact with an object, a robot can
use a tactile sensor consisting of a heating element and a
temperature sensor to recognize the object’s material based on
conductive heat transfer from the tactile sensor to the object.
When this type of tactile sensor has time to fully reheat prior to
contact and the duration of contact is long enough to achieve
a thermal steady state, numerous methods have been shown
to perform well. In order to enable robots to more efficiently
sense their environments and take advantage of brief contact
events over which they lack control, we focus on the problem
of material recognition from heat transfer given varying initial
conditions and short-duration contact. We present both modelbased
and data-driven methods. For the model-based method,
we modeled the thermodynamics of the sensor in contact with
a material as contact between two semi-infinite solids. For the
data-driven methods, we used three machine learning algorithms
(SVM+PCA, k-NN+PCA, HMMs) with time series of raw temperature
measurements and temperature change estimates. When
recognizing 11 materials with varying initial conditions and 3-
fold cross-validation, SVM+PCA outperformed all other methods,
achieving 84% accuracyComputational modeling of mechanical heart valves
http://hdl.handle.net/1853/56393
Computational modeling of mechanical heart valves
Yoganathan, Ajit P.
Issued as final report
2013-06-30T00:00:00ZYoganathan, Ajit P.Discrete Equivalent Wing Crack Based Damage Model for Brittle Solids
http://hdl.handle.net/1853/56382
Discrete Equivalent Wing Crack Based Damage Model for Brittle Solids
Jin, Wencheng; Arson, Chloé
The Discrete Equivalent Wing Crack Damage (DEWCD) model formulated
in this paper couples micro-mechanics and Continuum Damage Mechanics
(CDM) principles. At the scale of the Representative Elementary Volume
(REV), damage is obtained by integrating crack densities over the unit
sphere, which represents all possible crack plane orientations. The unit sphere
is discretized into 42 integration points. The damage yield criterion is expressed
at the microscopic scale: if a crack is in tension, crack growth is
controlled by a mode I fracture mechanics criterion; if a crack is in compression,
the shear stress that applies at its faces is projected on the directions
considered in the numerical integration scheme, and cracks perpendicular to
these projected force components grow according to a mode I fracture mechanics
criterion. The projection of shear stresses into a set of tensile forces
allows predicting the occurrence of wing cracks at the tips of pre-existing
defects. We assume that all of the resulting mode I cracks do not interact,
and we adopt a dilute homogenization scheme. A hardening law is introduced to account for subcritical crack propagation, and non-associated flow rules are adopted for damage and irreversible strains induced by residual crack displacements after unloading. The DEWCD model depends on only 6
constitutive parameters which all have a sound physical meaning and can be
determined by direct measurements in the laboratory. The DEWCD model
is calibrated and validated against triaxial compression tests performed on
Bakken Shale. In order to highlight the advantages of the DEWCD model
over previous anisotropic damage models proposed for rocks, we simulated:
(a) A uniaxial tension followed by unloading and reloading in compression;
and (b) Uniaxial compression loading cycles of increasing amplitude. We
compared the results obtained with the DEWCD model with those obtained
with a micro-mechanical model and with a CDM model, both calibrated
against the same experimental dataset as the DEWCD model. The three
models predict a non linear-stress/strain relationship and damage-induced
anisotropy. The micro-mechanical model can capture unilateral effects. The
CDM model can capture the occurrence of irreversible strains. The DEWCD
model can capture both unilateral effects and irreversible strains. In addition,
the DEWCD model can predict the apparent increase of strength and
ductility in compression when the confinement increases and the increasing
hysteresis on unloading-reloading paths as damage increases. The DEWCD
model is the only of the three models tested that provides realistic values of
yield stress and strength in tension and compression. This is a significant
advancement in the theoretical modeling of brittle solids. Future work will
be devoted to the prediction of crack coalescence and to the modeling of the
material response with interacting micro-cracks.
2016-01-01T00:00:00ZJin, WenchengArson, ChloéThe Discrete Equivalent Wing Crack Damage (DEWCD) model formulated
in this paper couples micro-mechanics and Continuum Damage Mechanics
(CDM) principles. At the scale of the Representative Elementary Volume
(REV), damage is obtained by integrating crack densities over the unit
sphere, which represents all possible crack plane orientations. The unit sphere
is discretized into 42 integration points. The damage yield criterion is expressed
at the microscopic scale: if a crack is in tension, crack growth is
controlled by a mode I fracture mechanics criterion; if a crack is in compression,
the shear stress that applies at its faces is projected on the directions
considered in the numerical integration scheme, and cracks perpendicular to
these projected force components grow according to a mode I fracture mechanics
criterion. The projection of shear stresses into a set of tensile forces
allows predicting the occurrence of wing cracks at the tips of pre-existing
defects. We assume that all of the resulting mode I cracks do not interact,
and we adopt a dilute homogenization scheme. A hardening law is introduced to account for subcritical crack propagation, and non-associated flow rules are adopted for damage and irreversible strains induced by residual crack displacements after unloading. The DEWCD model depends on only 6
constitutive parameters which all have a sound physical meaning and can be
determined by direct measurements in the laboratory. The DEWCD model
is calibrated and validated against triaxial compression tests performed on
Bakken Shale. In order to highlight the advantages of the DEWCD model
over previous anisotropic damage models proposed for rocks, we simulated:
(a) A uniaxial tension followed by unloading and reloading in compression;
and (b) Uniaxial compression loading cycles of increasing amplitude. We
compared the results obtained with the DEWCD model with those obtained
with a micro-mechanical model and with a CDM model, both calibrated
against the same experimental dataset as the DEWCD model. The three
models predict a non linear-stress/strain relationship and damage-induced
anisotropy. The micro-mechanical model can capture unilateral effects. The
CDM model can capture the occurrence of irreversible strains. The DEWCD
model can capture both unilateral effects and irreversible strains. In addition,
the DEWCD model can predict the apparent increase of strength and
ductility in compression when the confinement increases and the increasing
hysteresis on unloading-reloading paths as damage increases. The DEWCD
model is the only of the three models tested that provides realistic values of
yield stress and strength in tension and compression. This is a significant
advancement in the theoretical modeling of brittle solids. Future work will
be devoted to the prediction of crack coalescence and to the modeling of the
material response with interacting micro-cracks.Erosion analysis of coarse and fine grained sediments native to the state of Georgia
http://hdl.handle.net/1853/56381
Erosion analysis of coarse and fine grained sediments native to the state of Georgia
Krehbiel, Paul Richard
Sediment erosion in aquatic environments plays an important role in the design of bridges and other hydraulic structures with regard to scour, contaminant transport, and preservation of ecological systems. Erosion is the action of hydrodynamic forces overcoming the resistance by a sediment particle to being entrained and transported such that significant local erosion occurs. Sediments can be characterized as either non-cohesive or cohesive, as a classification determined by certain geotechnical properties. Non-cohesive sediments, consisting of sand and silt, primarily resist erosion due to the submerged weight of the particle and packing density of the sediment. Cohesive sediments, consisting of silt and clay, resist erosion via interparticle interactions, as determined by clay size fraction, water content or bulk density, and fines content, as well as other properties such as pH, organic matter, and mineralogy. Erosion of non-cohesive sediments that are primarily coarse-grained has been studied and documented by many researchers. While cohesive sediments have been investigated extensively, they are inherently more difficult to study because of the physicochemical properties that determine interparticle binding forces. This study focuses on a few geotechnical parameters to predict the erodibility of sediment mixtures on the coarse-fine transition boundary and mimic sediments native to Georgia. Previous researchers have investigated the erosion properties of coarse-sediment field samples in Georgia (Navarro 2004 and Hobson 2008) and predominantly fine, laboratory-prepared samples (Wang 2013 and Harris 2015). In order to span this collection of data, a series of samples was prepared and tested in an erosion flume in the Georgia Tech Hydraulics Laboratory using the same methodology as previous investigators to measure critical shear stress. The silt to sand ratio was held constant at 0.75, which is consistent with prior investigations of native Georgia sediments. Sand, silt, and Georgia Kaolinite were added to the samples, increasing the quantity of Kaolinite by weight in each subsequent sample from 10-30%. Sediment properties measured included water content, grain size distribution, clay size fraction, pH, temperature, and conductivity. Erosion rates for the mixtures were measured using a hydraulic flume. From these experiments, a critical shear stress for each mixture was determined based on water content. The critical shear stress data were analyzed as a function of measured geotechnical parameters using multiple regression analysis which provided a series of estimation equations. The relationships for critical shear stress derived in this research include a three-variable equation depending on water content, clay size fraction, and an interaction term; a fourth term that adds a fines content variable to the previous relationship; and a pair of equations that are implemented on separate coarse vs. fine data sets based on water content and particle size. While a weighted equation, which uses a combination of cohesive and non-cohesive equations, or two separate equations for coarse vs. fine sediment have merit, the optimal solution found in this research is the three-variable equation based on water content, clay size fraction, and an interaction term applied to all available data. However, more research should be conducted investigating the idea of two equations that are implemented on two separate data sets and on the criteria that best separate the data sets relative to cohesive vs. noncohesive erosion behavior. The results of this research can be used to find better predictions of sediment critical shear stress for Georgia sediments as a function of easily measured geotechnical parameters thereby providing better estimates of bridge scour and sediment stability for other structures in aquatic settings.
2016-12-15T00:00:00ZKrehbiel, Paul RichardSediment erosion in aquatic environments plays an important role in the design of bridges and other hydraulic structures with regard to scour, contaminant transport, and preservation of ecological systems. Erosion is the action of hydrodynamic forces overcoming the resistance by a sediment particle to being entrained and transported such that significant local erosion occurs. Sediments can be characterized as either non-cohesive or cohesive, as a classification determined by certain geotechnical properties. Non-cohesive sediments, consisting of sand and silt, primarily resist erosion due to the submerged weight of the particle and packing density of the sediment. Cohesive sediments, consisting of silt and clay, resist erosion via interparticle interactions, as determined by clay size fraction, water content or bulk density, and fines content, as well as other properties such as pH, organic matter, and mineralogy. Erosion of non-cohesive sediments that are primarily coarse-grained has been studied and documented by many researchers. While cohesive sediments have been investigated extensively, they are inherently more difficult to study because of the physicochemical properties that determine interparticle binding forces. This study focuses on a few geotechnical parameters to predict the erodibility of sediment mixtures on the coarse-fine transition boundary and mimic sediments native to Georgia. Previous researchers have investigated the erosion properties of coarse-sediment field samples in Georgia (Navarro 2004 and Hobson 2008) and predominantly fine, laboratory-prepared samples (Wang 2013 and Harris 2015). In order to span this collection of data, a series of samples was prepared and tested in an erosion flume in the Georgia Tech Hydraulics Laboratory using the same methodology as previous investigators to measure critical shear stress. The silt to sand ratio was held constant at 0.75, which is consistent with prior investigations of native Georgia sediments. Sand, silt, and Georgia Kaolinite were added to the samples, increasing the quantity of Kaolinite by weight in each subsequent sample from 10-30%. Sediment properties measured included water content, grain size distribution, clay size fraction, pH, temperature, and conductivity. Erosion rates for the mixtures were measured using a hydraulic flume. From these experiments, a critical shear stress for each mixture was determined based on water content. The critical shear stress data were analyzed as a function of measured geotechnical parameters using multiple regression analysis which provided a series of estimation equations. The relationships for critical shear stress derived in this research include a three-variable equation depending on water content, clay size fraction, and an interaction term; a fourth term that adds a fines content variable to the previous relationship; and a pair of equations that are implemented on separate coarse vs. fine data sets based on water content and particle size. While a weighted equation, which uses a combination of cohesive and non-cohesive equations, or two separate equations for coarse vs. fine sediment have merit, the optimal solution found in this research is the three-variable equation based on water content, clay size fraction, and an interaction term applied to all available data. However, more research should be conducted investigating the idea of two equations that are implemented on two separate data sets and on the criteria that best separate the data sets relative to cohesive vs. noncohesive erosion behavior. The results of this research can be used to find better predictions of sediment critical shear stress for Georgia sediments as a function of easily measured geotechnical parameters thereby providing better estimates of bridge scour and sediment stability for other structures in aquatic settings.