Browsing by Author "He, Zhaoming"
Now showing 1 - 20 of 29
- Results Per Page
- Sort Options
Item A High-Resolution Network with Strip Attention for Retinal Vessel Segmentation(2023) Ye, Zhipin; Liu, Yingqian; Jing, Teng; He, Zhaoming; Zhou, LingAccurate segmentation of retinal vessels is an essential prerequisite for the subsequent analysis of fundus images. Recently, a number of methods based on deep learning have been proposed and shown to demonstrate promising segmentation performance, especially U-Net and its variants. However, tiny vessels and low-contrast vessels are hard to detect due to the issues of a loss of spatial details caused by consecutive down-sample operations and inadequate fusion of multi-level features caused by vanilla skip connections. To address these issues and enhance the segmentation precision of retinal vessels, we propose a novel high-resolution network with strip attention. Instead of the U-Net-shaped architecture, the proposed network follows an HRNet-shaped architecture as the basic network, learning high-resolution representations throughout the training process. In addition, a strip attention module including a horizontal attention mechanism and a vertical attention mechanism is designed to obtain long-range dependencies in the horizontal and vertical directions by calculating the similarity between each pixel and all pixels in the same row and the same column, respectively. For effective multi-layer feature fusion, we incorporate the strip attention module into the basic network to dynamically guide adjacent hierarchical features. Experimental results on the DRIVE and STARE datasets show that the proposed method can extract more tiny vessels and low-contrast vessels compared with existing mainstream methods, achieving accuracies of 96.16% and 97.08% and sensitivities of 82.68% and 89.36%, respectively. The proposed method has the potential to aid in the analysis of fundus images.Item A model for noninvasive diagnosis of eye tumor and estimation of core body temperature by ocular surface temperature(2015-12) Tewolde, Senay; Chyu, Ming-Chien; Long, Kevin; James, Darryl; Sari-Sarraf, Hamed; He, ZhaomingThermal modeling of the human eye has been a subject of interest for years; however, the contributions of the existing models to clinical applications were limited. The purpose of this research is to develop the first algorithm that utilizes ocular surface temperature of the human eye (a) to detect the existence of an early stage intraocular tumor, and (b) to provide estimation of the core body temperature (CBT) on the back of the eye. Three versions of thermal models are developed, where two of them are developed for tumor detection and the other for CBT estimation. The first model assumes that the capillary vessels embedded in the vascular region are thermally insignificant, and the CBT is considered to be different from the blood temperature, whereas the second model and the model of the CBT estimation assume the capillary vessels to be thermally significant, causing a temperature difference between the solid structure of the vascular tissue and the blood, and the CBT is considered to be the same as the blood temperature. The thermal analyses of the eye model are performed by solving an inverse heat transfer problem using non-gradient-based optimization methods. The temperature distribution in the eye model is solved using Sundance, a high-performance parallel finite-element solution software, and HOPSPACK, a pattern search algorithm is employed to detect the heat source and to estimate CBT. Numerical results of the tumor detection algorithm show the capability of detecting a hypothetical small choroidal tumor, smaller than the smallest tumor stage classified by the Collaborative Ocular Melanoma Study (COMS). The algorithm is rigorously tested with realistic parameter uncertainties and a random error within the range of ±15 mK of noise-equivalent temperature difference available in commercial devices (e.g., IR cameras), and it has successfully demonstrated its clinical feasibility for patient-specific early stage tumor detection. In addition, the algorithm can quantify tumor heat generation rate for the purpose of diagnosis in terms of whether the tumor is benign or malignant. Parameters of the environmental conditions, viz. air-eye convective heat transfer coefficient and the surrounding temperature, are found to be the two most influential parameters in the tumor detection accuracy. The CBT is estimated using different levels of thermal resolutions, and in all cases, the algorithm has consistently demonstrated its capability of accurately estimating the tested CBT. The results highlight a wide range of possible applications of the current CBT algorithm in forensic, medical, sports, and other practices through a convenient, easily accessible, accurate, and noninvasive way.Item Accuracy and Stability of Smoothed Particle Hydrodynamics in Fluid-Structure Interaction Problems(2022-12) Isik, Doruk; He, Zhaoming; Idesman, Alexander; Christopher, Gordon; Bhattacharya, Sukalyan; Howle, Victoria E.An inevitable consequence of advances in computational power is humankind’s growing curiosity towards better understanding of its surroundings. Various methods and algorithms have been developed to address this curiosity by numerical modeling of complex problems. Smoothed particle hydrodynamics (SPH) is a numerical method that relies on approximating physical quantities through interactions of discrete points possessing individual material properties. Initially applied to study rotating polytropes, the usage areas have since then been extended to modeling of elasticity, fracture, heat transfer and incompressible flows. However, despite its relatively long history, the applicability of SPH for viscous incompressible internal flows is not fully understood and its limitations have not been fully assessed. This thesis is devoted to study the behavior of SPH when applied to such flows. Majority of the previous studies have studied the ability of SPH to capture flow phenomenon focusing primarily on mean integral quantities. This thesis underlines that ability of a method to model a phenomenon on a mean scale does not necessarily prove its viability to study problems where instantaneous flow structures play a role. Being an explicit numerical method, an instability in an instantaneous solution field might get amplified without being noticed, and adversely affects long term behavior. The accuracy and performance are tested through cases operating at a wide range of Reynolds numbers involving stationary as well as freely moving objects. The ability of SPH to model transitional flows past a stationary object, the effect of different viscous diffusion formulations on the solution field as well as its stability characteristics are analyzed. Relationship between the numerical speed of sound and the density field, along with the continuity equation formulations are also extensively investigated in a two-dimensional configuration with a freely migrating object using standard SPH and its commonly used variant. To improve its stability in such problems, a correction to the parameter controlling density diffusion is proposed. The understanding following these tests is then applied to model inertial migration of a neutrally buoyant sphere in a pipe with periodic corrugations both to test SPH in three-dimensional problems and to explain the physical mechanism responsible in such a migration.Item Accurate finite element simulation of stationary and moving dynamic cracks under impact loading(2018-08) Bhuiyan, A B M Abdul Ali; Idesman, Alexander V.; Parameswaran, Siva; He, Zhaoming; Qiu, Jenny; Nikishin, Sergey A.A new numerical technique that was developed for wave propagation problems has been implemented in current work for the simulation of elastodynamic fracture problems. Modeling of stresses near the crack tip and the calculation of dynamic stress intensity factors (DSIFs) have been done with the new technique. This technique includes the use of linear finite elements with reduced dispersion (RD) as well as the two stage time-integration method that quantifies and filters spurious high frequency oscillations. Several benchmark stationary and moving dynamic crack problems under impact loading have been solved. For the stationary cracks, it was found that the accuracy of stress calculations near the crack tip and the calculations of DSIFs can be significantly improved by using the linear elements with reduced dispersion. It was also interesting to note that the linear elements with reduced numerical dispersion yield much better results than the popular extended finite element method (XFEM) that uses special crack tip enrichment functions for the treatment of singularity in the vicinity of the crack tip. The implementation of RD is also much simpler than the implementation of XFEM. It is also interesting to mention that there is no necessity of the filtering stage for elastodynamic stationary crack problems under impact loading when using the finite elements with reduced dispersion (the spurious oscillations in DSIFs are very small and they decrease with mesh refinement). One benchmark impact problem with a propagating crack has been solved using XFEM. The numerical solution of this problem by XFEM includes spurious oscillations in the DSIFs. The simulation of moving cracks in XFEM framework depends on a variety of different factors such as the form of the enrichment functions used, the location of points with the enrichment functions for moving cracks, the quadrature rule in XFEM, etc. In the future research, we will study the combination of XFEM and the linear elements with reduced dispersion in order to improve the accuracy of the numerical results for propagating cracks.Item Air injection EOR and its application in shale oil reservoirs(2018-12) Zhang, Yao; Sheng, James J.; Bullard, Denny; Panacharoensawad, Ekarit; Gorell, Sheldon; He, ZhaomingA huge amount of shale oil in-place has been explored all over the world. The main problem in shale oil production is the extremely low oil recovery factor. Most of the shale oil remains in the formation and cannot be produced. The Enhanced Oil Recovery (EOR) techniques are required in shale oil plays. Many EOR methods have been applied in the shale oil reservoir, but the air injection method has not been used. Considering the availability of air and thermal effect caused by the oxidation reaction, the air injection could be a potential method for shale oil development. The main problem for air injection used in the light oil reservoir is oxygen may reach to the production well. The explosion could happen when the oxygen concentration in the production well reaching to the safety level (5%). In order to apply the air injection method in shale oil reservoir, the experimental works were performed to study the mechanism of air injection and the simulation studies were applied to estimate the performance of air injection in shale oil reservoirs. Three kinds of experiments were conducted in this study including Small Batch Reactor (SBR) experiments, core flooding tests, and sand pack tube tests. The SBR experiments were used to study the oxidation reaction between crude oil and oxygen. The air injection performance for oil displacement was studied by core flooding tests. The heat generated by the oxidation reaction was measured by the sand pack tube tests. According to the experimental results, the oxygen reaction rate is increased with the increment of temperature and pressure. The oil recovery factor by air flooding is increased with the increment of temperature. The heat generated by the oxidation reaction was observed and the temperature increased by 1 to 2℃. An analytical model was proposed to calculate the oxygen consumption under flooding conditions and the analytical model was verified by both of the experiment and simulation results. The oxygen consumption has a linear relationship with injection length squared divided by the pressure squared difference from the inlet to the outlet. The kinetic model used to describe the oxidation reaction was obtained by history matching the SBR and core flooding test results. The kinetic model was used in a field scale model to study the air injection. According to the simulation model, the oil can be produced by the air injection for 18 years without problem regarding oxygen consumption. The recovery factor of air injection is higher than that of nitrogen injection. The maximum temperature increment can reach to 11.1℃. The recovery factor is sensitive to the hydraulic fracture length and matrix permeability. A properly designed depletion production before air injection is recommended. The air injection can be applied in the shale oil reservoir for a long time (18 years) without a safety problem. The thermal effect caused by the oxidation reaction will benefit the oil production. The air injection could be a potential EOR method for the shale oil reservoir development.Item An in-vitro study of joint geometry and loading effects on anterior cruciate ligament strain and knee kinematics(2012-05-14) Breighner, Ryan E; Hashemi, Javad; Domire, Zachary; Mansouri, Hossein; Ekwaro-Osire, Stephen; Yang, Jingzhou; He, ZhaomingA frequent injury in both sport and recreational activities is a rupture of the anterior cruciate ligament or ACL. The ACL is a ligament in the knee that connects the femur to the tibia. It prevents excessive anterior tibial translation and contributes to overall knee joint stability. However, when excessively strained the ligament may rupture. This occurs over 250,000 times each year in the United States alone. Beyond the initial trauma of the injury, and costly reconstruction and rehabilitation, ACL injured individuals are at risk of early onset osteoarthritis and may require total knee replacement later in life. Approximately 70% of all ACL injuries are non-contact in origin, i.e. no direct contact to the knee led to the injury. Even more troubling is that non-contact anterior cruciate ligament injury is not especially well understood. This lack of understanding is likely the result of the short time period over which the injury occurs. In order to avoid risk to live subjects, and because of a lack of sufficient animal models, the most appropriate way of studying anterior cruciate ligament injury is through in-vitro simulation. In order to better understand the etiology of ACL injury and potential injury risk factors, a purpose built knee loading simulator was designed and built with the ability to simulate various athletic activities. This simulator is capable of controlling and monitoring muscle force levels, measuring strain in various knee ligaments, and measuring joint contact forces/pressure distributions in the knee during these activities. To better understand the influence of tibial geometry on ACL strain and injury, several studies of various knee-loading conditions were conducted on cadaver knees. The knees were first imaged using MRI, and measurements of their respective tibial geometries were taken. Subsequently, the knees were installed in the simulator and muscle forces were applied. In one of these studies, hip extensor-generated joint compressive forces were also applied, followed by an impulsive ground reaction force. An existing probabilistic model of injury risk, based on tibial plateau geometry was evaluated using these data. Additionally, another series of tests was conducted comparing strain in the ACL and MCL resulting from static valgus torques. The strain generated in the ACL and MCL were measured at various flexion angles under 200N of quadriceps activation and 80N of hamstrings force. The particular objective of these tests was to determine whether isolated ACL injury is possible under purely valgus loading. Another study examining the moderating effect of tibial geometry on muscle activity induced ACL strain was also conducted. The study looked at ACL strain generated from the application of quadriceps forces and hamstrings forces at different flexion angles and related these strains to the values of medial and/or lateral tibial slope in the tested knees. The results of these studies indicate that tibial slope and medial tibial depth are significant predictors of ACL strain and that pre-landing joint compression is protective of the ACL under dynamic loading. Additionally, it was shown that MCL strain increases more appreciably as a result of valgus loading as compared to the ACL. This information, coupled with the material properties of the two ligaments suggest that isolated ACL injury cannot result from purely valgus loadings. Additionally, tibial slope and medial tibial depth were shown to significantly affect ACL and MCL strain. Lastly, it was shown that medial and lateral tibial slopes moderate ACL strain due to muscle activity. The findings of the above studies are all novel. At the time of this writing, no one has related both medial tibial depth and lateral tibial slope to ACL strain under impulsive loading. Additionally, while joint compression as a means of protection for the ACL has previously been proposed, the method in which it was applied here eliminates potentially confounding posterior drawer effects from the hamstrings. The simultaneous measurement of ACL and MCL strains under valgus loading is also a significant contribution to the literature. The finding that tibial slope moderates muscle activity-induced ACL loading is also novel and serves not only as a further verification of tibial geometry as an ACL risk factor, but also shows that adequate muscle forces and joint flexion may be able to compensate for disadvantageous tibial geometries. It is the author’s hope that the information yielded by these studies can be incorporated into more effective prevention training programs so that the occurrence, severity, and overall costs of ACL injury can be reduced.Item Analysis of Brownian dynamics and unsteady particle-motion in viscoelastic fluids(2012-05) Azese, Martin; Bhattacharya, Sukalyan; Blawzdziewicz, Jerzy; Ibragimov, Akif; Christopher, Gordon; He, ZhaomingIn recent times, micro-rheological applications involve determination of viscoelastic properties for samples that are either too precious and fragile or in a state (like inside a cell) where macroscopic experiments are impossible. In such cases, direct measurements using rheometers are not possible, because then the system can be structurally destroyed. One way to circumvent this problem is to predict fluid-rheology from the random motion of a Brownian sphere in the medium. Thus, many past attempts tried to relate viscoelastic properties to features of stochastic motion like time-dependent velocity correlation or mean square displacement. All such theories, however, invariably involve heuristic assumptions inherited from classical studies on purely viscous fluid. This is why in this thesis the classical theories of statistical mechanics for Brownian dynamics are first reevaluated and then modified to suit the new technological demand. This research first focuses on the flow-analysis which describes hydrodynamic field inside a viscoelastic medium. Accordingly, a mathematically rigorous perturbation method is developed which isolates the leading order linear contributions from higher order non-linearities due to both convective acceleration and constitutive relation. As a result, the conditions for linearized analysis are identified, and the leading order fields as well as particle-motion are determined. Then the analysis concentrates on the leading order linearized hydrodynamic equation only, and scrutinizes the relevance of classical theories of statistical mechanics for micro-rheological applications. In this context, three key conclusions are drawn revealing the errors in the earlier concepts. Firstly, the validity of fluctuation-dissipation theorem are questioned, as it requires Markovian condition only true for memory-less systems without viscoelasticity and flow-inertia. Secondly, well-known Langevin equation for Brownian dynamics is rectified by including the effect of fluid-inertia in the equation of motion of the suspended body in a density-matched liquid. Thirdly, the equipartition principle is reinterpreted to find the correct normalization for correlation of Brownian forces where energy associated with the translation of a Brownian particle is considered to have an additional contribution from the induced flow in the liquid. Thus, we discard the fluctuation-dissipation postulate, and recommend an inertia-corrected modified Langevin formulation to be used in micro-rheological problems. We use our new theory to correctly describe the stochastic dynamics of a Brownian sphere in a viscoelastic liquid by relating its time-dependent velocity correlation function and mean square displacement to fluid-rheology. Resulting conclusions differ substantially from popular beliefs while maintaining agreements under the long-time or low-frequency limit under proper conditions. Thus, our alternative formulation can be used in microrheological measurements to predict large-frequency complex viscosity for which the failure of past theories are well-documented. Moreover, we analyze the classical problem involving a Brownian sphere in a purely viscous liquid with density similar to the suspended solid. The errors in the original Langevin formulation are highlighted where the inertia of the fluid is ignored in both equation of particle-motion and equipartition principle. Our new theory with proper corrections is used to find the unsteady velocity correlation and mean square displacement of the sphere. The computed temporal variations of these quantities differ substantially from the results obtained from the classical Langevin equation. Curiously, however, the long-time diffusion coefficients in both cases exactly coincide. It seems that the earlier analysis calculates the correct diffusivity, because the error in equation of motion and misinterpretation in equipartition principle nullify each other. As long-time diffusivity is a quantity which has been experimentally verified over a century, the aforementioned agreement can be viewed as a further verification of the new theory.Item Annulus tension in the tricuspid valve: The effects of annulus dilation and papilliary muscle movement(2012-08) Smith, Dylan; He, Zhaoming; Idesman, Alexander V.Tricuspid valve (TV) annular dilatation compromises TV coaptation and may lead to regurgitation. Annular mechanics involves interaction between the TV leaflets and right ventricle myocardium and plays a role in annular dilatation. Annulus tension (AT) is a leaflet force on the annulus and contributes to understanding of annular dilatation. The objective for this study was to determine the effect of PM position and annulus dilation on AT. Force transducers were attached to the porcine TV annulus using a TV closure test rig. The test rig allowed us to apply 40 mmHg of pressure on the TV and adjust the size of the annulus to the normal and dilatated size, which were 9.0 cm2 and 50% increase in area. Papillary muscles were secured and adjusted to simulate papillary muscle displacement, which are 5 mm apical displacement, 5mm lateral displacement, and 5 mm both apical and lateral displacement of the anterior papillary muscle. AT was measured in a normal and 3 pathological papillary muscle positions for a normal and a dilatated annuli. Eight TVs were tested. AT peaked at the commissures for both normal and dilatated annuli. The average AT in the normal TV was approximately 8.9 N/m with the highest AT along the posterior segment of the annulus. The average AT for the dilatated annulus was 16.3 N/m which was almost a 90% increase. The greatest average increase was found along the septal and anterior commissures and the smallest increase along the posterior commissure. As compared with average AT in the normal annulus, the average AT increased by 18.0%, 7.9%, and 30.3% for a 5mm apical, 5mm lateral, and both 5mm apical and lateral PM displacements, respectively. The similar AT distribution was found for both annuli. The apical anterior papillary muscle displacement increased AT evenly throughout the entire annulus. The lateral anterior papillary muscle displacement lowered AT in the anterior commissure and increased AT in other segments. Both apical and lateral anterior papillary muscle displacement caused no change of AT in the anterior commissure for the normal annulus, lowered AT in the anterior commissure for the dilated annulus, and increased AT in all other segments. In the normal annulus, the average AT is significantly increased(p<.05) for the apical and both apical and lateral anterior papillary muscle displacement, with no significant change for lateral anterior papillary muscle displacement. In the dilated annulus, the average AT is significantly increased only for the both apical and lateral anterior papillary muscle displacement, and not changed for the other two anterior PM conditions. TV AT is smaller in the TV than that in the mitral valve, which indicates the TV interacts with right ventricle wall in a weak manner. The increase in AT due to annulus dilatation and papillary muscle displacement helps to counteract TV annular dilatation.Item Anterior cruciate ligament response due to forces resulting from quadriceps muscle and ground reaction(2013-05) Bhuiyan, Ariful; Osire, Stephen E.; Hashemi, Javad; Hardy, Daniel; He, Zhaoming; Yang, JamesInstrumented cadaveric knees were used to quantify the association between quadriceps muscle force, ground reaction force, and strain in anterior cruciate ligament (ACL) through in vitro simulation of a vertical jump-landing process in a dynamic loading simulator. The research question for this project was set: “Can the ACL strain be influenced by the unopposed quadriceps force (QMF), low knee flexion angle, and constant ground reaction force (GRF)?” Fourteen cadaveric knees were mounted in the custom made dynamic loading simulator to measure strain on the anteromedial bundle of the ACL using a differential variable reluctance transducer (DVRT). An I-Scan pressure transducer (Model 4000) was used to measure contact pressure and area in the tibiofemoral joint. A 3D nonlinear dynamic finite element knee model was developed, and the results from this model were validated with the experimental results. The influence of the intrinsic risk factors, like the lateral tibial slope, medial tibial slop, and the curvature of the medial compartment do exist on the loading in ACL at the time of applying quadriceps muscle force. It was also observed that the tibia rotates internally prior to landing and during landing. During the landing phase, the peak pressure on the lateral compartment is very high, compared with the medial compartment. During the landing phase, both the contact area and pressure increases in the tibiofemoral joint. The influence of pressure induced joint conformity is also justified. It can be concluded from the obtained results from the experimental and numerical works that the unopposed quadriceps muscle forces coupled with ground reaction force at low knee flexion angle cannot cause ACL injury. Joint compressive loads induced by large muscle forces and GRF introduces the joint conformity, and this joint conformity produces the primary restraint against anterior tibial translation at low flexion angles along with menisci.Item Basis transform solution for Brinkman equation to describe unsteady hydrodynamic interactions(2021-05) Liu, Bo; Bhattacharya, Sukalyan; Blawzdziewicz, Jerzy; He, Zhaoming; Idesman, Alexander; Ibraguimov, AkifThis dissertation elucidates how inherently unsteady hydrodynamic interactions between two closely situated spheres in viscous liquid affect their time-dependent motion. The system represents two spherical Brownian particles for which temporal inertia is always comparable to the viscous forces even though convective inertia is negligible. Therefore, instead of Stokes equation, the linearized unsteady Navier-Stokes is considered. We apply Fourier transform in frequency space to convert this to Brinkman equation. Then a novel mathematical formulation is proposed to investigate vector field solution governed by Brinkman equation. The methodology relies on the expansions in multiple sets of vector basis functions corresponding to each sphere. The key result in the formulation is the mutual transformations between the basis functions of two such sets. This allows the derivation of the matrix relations coupling the unknown amplitudes with the given inhomogeneous boundary conditions. The presented mathematical theory is validated by complementing numerical calculations. Accordingly, the solution is constructed using the outlined method, and the error in the form of departure from the intended boundary condition is evaluated. This error vanishes very quickly with increasing number of basis solutions demonstrating high accuracy and exponential spectral convergence of the numerical scheme. Construction of hydrodynamic force or torque for each basis function in the derived vector field provide the frequency-dependent two-body frictions. On the contrary, inverse Fourier trans-form of these after adding appropriate inertial contributions yield time-dependent mobility response. The friction and mobility values are validated in limiting cases for short- and long-time limits. The scaling laws of these quantities are also explored as functions of the separation distance between two solid bodies revealing important physical insight into the complicated dynamics.Item Characterization of the mitral valve using accelerated wear testing(2010-12) Riggan, Courtney N.; He, Zhaoming; Hashemi, Javad; Patterson, Patrick E.There is a fundamental need for a better kind of long-term testing for artificial heart valves along with a biologically equivalent artificial mitral valve. Accelerated wear testing has been used for a long period of time in the engineering field to test the failure modes of materials and devices, but has not been widely accepted in the medical device area. The mitral valve is difficult to get to and therefore little is known about the leaflet motion. The lab has obtained a closed-system, compact, accelerated wear testing apparatus designed for aortic valves, and nothing is known about the device. The lab deals exclusively with mitral valves. The objective of this study is to characterize and fully understand this accelerated wear testing apparatus, to modify it to run with mitral valves and to define the mitral valve motion using pressure waveforms and leaflet motion. This was achieved by calibrating and verifying the device using an aortic valve with a known pressure curve. The device was then modified to fit mitral valves and the motor motion was measured to characterize the movement of the apparatus. A pressure waveform was then obtained for the mitral valve. Images of the mitral valve were obtained for these pressures and orifice area was calculated. This information gives a clear understanding about the design and function of the device and of the mitral valve.Item Design and Development of a Systemic Mock Circulation Loop with a Novel Beating Left Ventricular Simulator(2016-08-19) Baturalp, Turgut Batuhan; Ertas, Atila; Tate, Derrick; He, Zhaoming; Arvandi, Aliakbar; Barhorst, AlanMock Circulation Loops (MCLs) are used as a mechanical representation of the human cardiovascular system for in vitro testing of most cardiovascular devices. Thus, MCLs are essential to most of the cardiac device designs as a testbed. Cardiac device design procedures generally adopt mock circulatory systems before advancing to animal or clinical trials, which are much more troublesome and expensive. Therefore, various physiological cardiac failure or operating scenarios, need to be replicated in the MCL towards the testing of cardiovascular devices. Replicating these different scenarios requires not only a fully-automated MCL that can reproduce the necessary conditions precisely and switch between them without trouble, but also a beating LV simulator that mimics an actual human LV’s parameters, such as geometry, LV wall movement trajectories and elastance of the LV wall. A fully automated MCL design is proposed in this study and the beating LV simulator concept is analyzed, designed, prototyped and tested in the scope of this study. The novel and challenging part of this study was the development of the beating LV simulator. Due to the fact that designing a beating LV simulator which replicates geometry, wall motions, and muscle fiber orientations of an actual human LV chamber is a complex problem; a design analysis method known as Interpretive Structural Modeling (ISM) method, was applied to selected design approaches of the beating LV simulator’s actuation. Pneumatic muscles, flexible bands or strings, and artificial muscle materials were considered as candidates for different design approaches to replicate the wall motion and function of an actual human LV. ISM analysis showed that pneumatic muscle approach has proven to be the most promising. As a result, the pneumatic muscle approach was adopted to achieve realistic LV wall motions of various operating conditions. The beating LV simulator was prototyped by using Pneumatic Artificial Muscles (PAMs) and latex based elastic tissue material. Four PAMs were placed on the LV inner mold in the helical orientation and latex tissue material was applied on them. The fiber orientation of the cardiac muscles plays an essential role on the pumping performance of the actual LV chamber. Therefore, a thermally loaded FEA model was developed to investigate different fiber orientations in the future. The FEA model was validated by using experimental results of the beating LV simulator. Moreover, polymer based artificial heart valves (AHVs) were designed and prototyped for use in the MCL system, however, mechanical check valves were used in the experiments due to the short lifecycle of the prototyped AHVs. A state-space based numerical model with a time varying LV compliance was adopted and implemented. The numerical model draws an analogy to the systemic cardiovascular system, as an electrical circuit for MCL parameter estimation, and consistency investigation of the developed beating LV simulator’s experimental results. The numerical results were obtained for two cases. These were, healthy human resting operating cardiovascular conditions and the parameters of the numerical model were adjusted for the beating LV simulator’s experimental setup. The results of the prototyped beating LV simulator experiments showed that the goal of achieving realistic LV cardiac mechanics is possible. 50, 60, and 70 beats per minute cardiac cycles were generated in the experiment and an average flow rate of 2 liters per minute was obtained. The stroke volume of the beating LV simulator was 33.3 ml with an ejection fraction of 28.2% and the twist angle of the apex was measured as 21 degrees in cyclic loading. Additionally, the static loading experiment results were compared with the FEA model results and the comparison showed that the results are correlated.Item Determination of the mechanical properties of the Tricuspid Valve Annulus within the surrounding myocardium and leaflets(2017-07-10) Basu, Avik; He, Zhaoming; Lacerda, Carla; Idesman, Alexander; Parameswaran, Siva; Ren, BeibeiThe Tricuspid valve (TV) annulus is a heterogeneous structure located at the intersection of the leaflets of the TV and the myocardium. The TV annulus provides structural stability and assists in leaflet coaptation for the TV. This study helps to understand the mechanical properties of the TV annulus within the surrounding myocardium and leaflets. Since the annulus structure is intertwined within the myocardium and leaflets, incorporating these tissue components were important to get a comprehensive analysis of the mechanical properties for TV annulus mechanics. There are 3 leaflets (septal, posterior, and anterior) of the TV and there were 3 different annulus segments regions. Each segment had 30 mm length, 8 mm width, and 2 mm thickness. Eleven TVs were tested up to a 20% strain for all 3 TV annulus segments. Young’s Modulus (E) and extensibility (εT) values were calculated. Subsequent western blots for collagen I and III content along each annulus segment as well as histological analyses were conducted. GAG concentration was estimated along each annulus segment. All 3 TV annulus segments had many non-linear characteristics based on stress-strain curve outputs. The results show that the septal annulus E value (200 ± 70.5 kPa) was the greatest value out of all the 3 segments. The septal annulus segment is statistically greater (p < 0.02) than the anterior (109.6 ± 86.2 kPa), but not for the posterior (147.7 ± 56.2 kPa). Western blotting and histological analysis indicated that collagen I content was greatest along the septal annulus segment. Collagen III content was greatest along the posterior annulus segment, followed by the septal and anterior annulus segments. The εT values were 0.0607 ± 0.025, 0.0614 ± 0.016, and 0.0789 ± 0.033 for septal, posterior, and anterior annulus segments and showed no statistical difference. GAG concentration was 36.5 ± 30 µg/ml, 30 ± 31 µg/ml, and 33 ± 31 µg/ml for the septal, posterior, and anterior annulus segments, respectively. The collagen I and III content is directly related to the E and εT values along each annulus segment based on degree of stiffness. GAG concentrations for each segment suggest a counteracting mechanism to the high E values along each annulus segment to increase flexibility along the region. The septal annulus segment is able to withstand the effects of annulus dilation in comparison to the other annulus segments as shown from E and εT values. The E, εT, and GAG concentrations provide a new understanding of TV annulus mechanical properties which can be used to improve annuloplasty design in the future.Item Exploration of materials under compression of non-hydrostaticity and shear(2018-08) Gao, Yang; Ma, Yanzhang; Chyu, Ming; Yeo, Changdong; He, Zhaoming; Li, GuigenNon-hydrostatic compression is in general a circumstance that material is subject to in the nature. Understanding the behaviors of materials under such a condition makes it possible to gain insights into the physics and chemistry beyond their natural phenomena. In this dissertation, we attempt to explore the behaviors of materials under non-hydrostatic conditions to seek their potential in engineering applications. Three topics are covered in this work: the search of new scintillators via non-hydrostatic compression, the synthesis of novel phases from known materials under shear load, and the reaction of super-hard materials to shear stress. A diamond anvil cell and a rotational anvil cell accompanied with synchrotron X-ray diffraction, Raman spectroscopy, transmission electron microscope, and X-ray photoelectron spectroscopy were employed as diagnostic methods. MnWO4 was studied by synchrotron X-ray diffraction to 50.1 GPa in a diamond anvil cell. Comparison experiments under the hydrostatic and non-hydrostatic conditions were performed. A structural phase transformation is observed, of which the high-pressure phase is determined to be a triclinic structure. Under the non-hydrostatic condition, the transformation to a high-pressure phase of MnWO4 initiates at a far lower onset pressure. The low-pressure and the high-pressure phases are discovered to coexist in a wide range of pressure under both the hydrostatic and non-hydrostatic conditions, which indicates that the triclinic structure is energetically comparable to that of the wolframite one. Combined with previous reports, non-hydrostatic effect is believed to reveal the triclinic phase of wolframite tungstates at a far lower pressure. The discovery of this work suggests that the wolframite tungstates could be a new source of scintillating materials, and non-hydrostatic effects can be utilized to lower the condition required for their synthesis. Diamond synthesis from graphite is achieved at below 1 GPa and room temperature using a rotational anvil cell. By applying large plastic shear, graphite transformed into hexagonal and cubic diamonds at extremely low pressures of 0.4 and 0.7 GPa, respectively. The formation of a new orthorhombic diamond phase was also observed after pressure elevation to 3 GPa. It is discovered that shear, instead of pressure, plays the key role in this transformation. The discovery of this transformation suggests new mechanism of phase transformations with drastically reduced pressures by shear and is expected to new materials synthesis strategies. Furthermore, the formation of diamonds under unconventionally low pressures also opens up new thoughts in geophysics that the micro-diamonds at geological sites could have formed in the cold crust due to shear-related historical activities instead of the conventional subduction-exhumation process. Decomposition of B4C was observed at 1.0 GPa under large plastic shear using a rotational anvil cell. The products are determined to be a boron-very-rich compound, B50C2, and a pure carbon substance, nano-crystalline graphite. Amorphization of B4C is also observed in the quenched sample. The discovery of B4C’s decomposition and amorphization under large plastic shear suggests a new explanation, in addition to amorphization, to B4C’s mystery shear strength reduction over 20 GPa. It also reveals that shear combined with modest pressure is essential in initiating phase transformations and chemical reactions than hydrostatic compression. Furthermore, the discovery of such shear-induced decomposition of boron carbides may also open a new strategy of non-hydrostatic effects’ utilization in both engineering and chemistry.Item Intelligent diagnosis of heart murmurs in children with congenital heart disease(Hindawi, 2020) Wang, Jiaming; You, Tao; Yi, Kang; Gong, Yaqin; Xie, Qilian; Qu, Fei; Wang, Bangzhou; He, ZhaomingHeart auscultation is a convenient tool for early diagnosis of heart diseases and is being developed to be an intelligent tool used in online medicine. Currently, there are few studies on intelligent diagnosis of pediatric murmurs due to congenital heart disease (CHD). The purpose of the study was to develop a method of intelligent diagnosis of pediatric CHD murmurs. Phonocardiogram (PCG) signals of 86 children were recorded with 24 children having normal heart sounds and 62 children having CHD murmurs. A segmentation method based on the discrete wavelet transform combined with Hadamard product was implemented to locate the first and the second heart sounds from the PCG signal. Ten features specific to CHD murmurs were extracted as the input of classifier after segmentation. Eighty-six artificial neural network classifiers were composed into a classification system to identify CHD murmurs. The accuracy, sensitivity, and specificity of diagnosis for heart murmurs were 93%, 93.5%, and 91.7%, respectively. In conclusion, a method of intelligent diagnosis of pediatric CHD murmurs is developed successfully and can be used for online screening of CHD in children.Item Investigating the onset of slip in gait by employing probabilistic theory and optimization-based motion prediction(2014-05) Gragg, Jared; Yang, James; Ekwaro-Osire, Stephen; Smith, James L.; DeLucia, Patricia R.; He, ZhaomingSlips, trips, and falls have serious impact on humans and can cause serious injury or death. There is potential to reduce the likelihood of slips, and thus falls, which could reduce injuries and save money. The likelihood of a slip in gait is related to the available friction and required friction. Previous research has been dedicated to predicting the probability of slip; however, there are major drawbacks to the previous studies. In addition, the studies do not provide probabilities of slip for real world scenarios and ignore high potentials for slip by ignoring certain peaks in the required friction during level gait. Also, no one has extended the theory to include situations such as ramp gait, which in general has higher potential for slip compared to level gait. There are no studies that look at the sensitivity of the probability of slip to the input parameters to determine which parameters have the most influence on the probability of slip. Finally, there are no studies which incorporate simulations to predict gait adaptations that would reduce the probability of slip. This study addresses these drawbacks through the following objectives. First, a systematic method for predicting the probability of slip in gait, both level gait and ramp gait, was developed. It is critical to be able to predict the probability of slip in gait to determine whether a given gait-shoe-floor combination is hazardous or not. Second, a sensitivity analysis was performed to determine which of the input parameters has the highest influence on the probability of slip. Understanding the sensitivity of the input parameters on the probability of slip allows one to determine practical ways to reduce the probability of slip. Finally, a simulation method was developed that predicts gait adaptations that reduce the potential for slip.Item Investigation of left ventricular mechanics for a mitral valve implanted with a coaptation plate(2012-05) Krishnamoorthy, Srikumar; He, Zhaoming; Parameswaran, SivaIschemic left ventricular disease is one of the leading causes of death in North America, and around the world. There are a few surgical techniques in use, to correct Left ventricular ischemia but their effectiveness has not been particularly exceptional. A new method of correcting it by inducing coaptation by means of a coaptation plate was proposed, and the changes in left ventricular fluid mechanics effected by the implantation of the plate was assessed using techniques of computational fluid dynamics. The parameters for assessment included wall shear stress, vortex formation number, energy efficiency of the plate- implanted system. The velocity and energy distribution in the anterior left ventricular vortex was assessed in an attempt to answer the long-standing question of the importance of the left ventricular vortex to normal left ventricular function. Geometric models and associated meshes were rendered in the commercially available software package GAMBIT, and flow simulation was performed using the package software package FLUENT. The results of each case were tabulated, compared and an assessment was made.Item Left ventricle fluid mechanics under mitral valve edge-to-edge repair(2010-12) Shi, Liang; He, Zhaoming; Hoo, Karlene A.; Parameswaran, Siva; Ma, Yanzhang; Yang, JamesEdge-to-Edge repair (ETER) is one of the mitral valve (MV) repair techniques frequently used in MV disease treatment. The hemodynamic changes caused by ETER may cause clinical complications such as thrombus formation and affect left ventricle function. However, the detailed information about the hemodynamic changes after ETER is very limited and therefore needs further exploration. The objective of this study is to provide further understanding of the fluid mechanics under mitral valve ETER. We develop two specific aims to achieve the objective. The first one is to explore the peak diastolic hemodynamic change immediately downstream of MV after ETER by in vitro steady flow experiment. The findings of the experiment show us the changes of MV hemodynamics under different ETER suture configurations. The second specific aim is to study the vortex dynamics inside left ventricle by both pulsatile flow experiment and computational simulation. By the investigation of the vortex ring structure under ETER, we find vortex reconnection may appear under short suture configuration. ETER may have better washout effect compared with normal valve and short suture under ETER may have best performance in kinetic energy dissipation. The fluid dynamic changes under ETER in this study not only reveal the possible clinical complications under certain suture configurations but also show the effect of vortex development change on the left ventricular efficiency, which may have relation with left ventricular dysfunction. Therefore, we suggest surgeon’s attention on the selection of proper suture configurations in regard to the findings of this dissertation.Item Mechanical Analysis of the Connection Structure of a Double-Layered Valve Stent within an Annuloplasty Ring(2024) Dong, Ke; He, ZhaomingBackground: In this study, to address the failure of mitral valve repair surgery, a novel valve-in-ring model for an artificial mitral valve annuloplasty ring and a new doublelayer mitral valve were established. A suitable number and length of ventricular fixation struts within the annuloplasty ring, as well as the implantation depth, result in variations in stress and strain for the inner and outer stent layers. Methods: The compression and self-expansion model of the stent was established via finite element analysis. The changes in stress and strain were analyzed by setting the length and number of the ventricular fixed struts and implantation depth. Results: When only affected by factors such as blood pressure, the maximum stresses of stent structures with three and six ventricular fixed struts are 476 and 222 MPa, respectively, in the right posterior annular region. At implantation depths of 0, 0.5, 1, and 2 mm, the maximum stresses are located in the left posterior annular region of the outer stent and are 740, 697, 709, and 742 MPa, respectively, and the maximum displacements of the inner stent are all in the right posterior ventricular fixed strut region of the posterior annulus and are 3.71, 3.10, 2.48, and 1.87 mm, respectively. In the three and six ventricular fixed strut stents, when the ventricular fixed strut length is 3, 4, and 5 mm, the maximum stresses are 570, 557, and 621 MPa and 674, 666, 644 MPa, respectively. Conclusions: Appropriately increasing the number of ventricular fixed struts can effectively reduce damage to the stent inside the body, and the damage to the stent is relatively consistent across different implantation depths;however, the right side of the stent's posterior annulus is particularly susceptible to damage. However, if the implantation depth is lower, the impact on the inner stent will be more significant. As the number of ventricular fixed struts increases, the strut length variation has a relatively stable impact on stent damage.Item Mechanotransduction and control of valvular cell phenotype as tools to inform valvular pathophysiology(2018-08) Ali, Mir; Lacerda, Carla; Gill, Harvinder Singh; Vanapalli, Siva A.; He, ZhaomingValvular degenerative diseases cause significant morbidity and mortality in developed countries. Valvular interstitial cells (VICs) are responsible for pathogenesis of these diseases. In healthy adult valves, VICs remain as quiescent phenotype. In degenerative valves, they become phenotypically activated like myofibroblast. Controlling VIC activation has important therapeutic and tissue engineering implications. Mechanobiology also plays a crucial role in valvular pathophysiology. Various mechanical forces trigger VIC phenotypical transformations and valvular degenerative diseases. However, valvular mechanotransduction pathways of these forces are largely unknown. Valves have an outer endothelium layer consisting of endothelial cells. Endothelium maintains valvular physiology but function of endothelium in valvular mechanotransduction is unknown. In this dissertation, we studied VIC activation in terms VIC morphological distribution in vitro. We also demonstrated control over VIC activation in vitro via interaction with progenitor subpopulations and via variable substrate stiffness. Lastly, we studied valvular mechanotransduction and the endothelial function in it using high throughput analysis of valvular cell and tissue response to strain and substrate stiffness in vitro. In our study of VIC phenotypes, we detected six distinct VIC morphologies and their relative abundance in pathophysiological states. Morphologies were associated with VIC activation. This in vitro morphology based VIC phenotype detection is simpler and quicker compared to molecular marker-based techniques. We demonstrated control over VIC phenotype in vitro, inducing VIC deactivation by lowering substrate stiffness. Reversibility in VIC activation has important implications in maintaining quiescence for in vitro research with rare human VICs and for tissue engineering purposes. We also demonstrated control over VIC activation, in vitro, using interaction with two progenitor subpopulations of VICs resembling hematopoietic and mesenchymal stem cells. These subpopulation functions were amplified by increasing their concentrations to >50% in VIC cultures. Hematopoietic stem cell subpopulation induced deactivation in VICs. Mesenchymal subpopulations did not show any effect on VIC activation. This is the first study showing function of a progenitor VIC subpopulation in VIC activation and will augment progenitor VIC research which is at its infancy. Next, we focused on high throughput proteomic analyses of valvular cell and tissue response to variable substrate stiffness and physiologic cyclic strain. Cytoskeletal, signal transduction, oxidative stress and translation proteins were upregulated on stiff substrate suggesting their role in VIC mechanotransduction. Strain resulted in quiescent state with upregulated mechanoreceptors associated with pro-inflammatory and pro-angiogenic activities. Endothelium removal resulted in upregulation of translation suggesting a protective role of valvular endothelium. These high throughput studies will contribute to identifying cell signaling cascades and mechanotransduction pathways in VICs. Lastly, we also developed a computational model of diastolic heart left ventricle to determine blood flow and wall stress related changes introduced by medical device ‘Mitraclip’. Despite some limitations, the model can be used in future to simulate other left ventricular diastolic phenomena. This dissertation improves our understanding and control of VIC phenotype structure and function as well as links molecular, cellular and tissue-level valvular responses to static and dynamic stimuli. Taken together, these results expand the picture of valvular cell mechanotransduction, morphology, phenotypical transformation, signaling networks and tissue maintenance.