Catarina Castro

Inverse methods in design of industrial forging processes

Abstract An approach to optimal shape design in forging is presented. The design problem is formulated as an inverse problem incorporating a finite element 3D analysis model and an optimisation technique conducted on the basis of design sensitivities. The mechanical analysis provides information to predict deformation, stresses and strains necessary to the shape optimisation problem. The objective is to minimise a function describing die underfill and excessive material waste. Analytical sensitivities of the objective function are required and the calculation of discrete derivatives based on the differentiation of the discrete problem equations is considered. Due to the time-dependent deformation process the direct differentiation method has been implemented. The capability of the proposed inverse approach to deal with optimal forging of industrial parts is demonstrated.

Eliminating Forging Defects Using Genetic Algorithms

Abstract In this article, an optimization method for metal forging process designs using finite element-based simulation is presented. Using as entry parameters the specifications of the final product the so-called inverse techniques developed for optimization problems allows the calculation of the optimal solution, the design parameters that produce the required product. An evolutionary genetic algorithm is proposed to calculate optimal shape geometry and temperature. An example demonstrating the efficiency of the developed method is presented considering a two-stage hot forging process. It considers optimization of the process parameters to reduce the difference between the realized and the prescribed final forged shape under minimal energy consumption, restricting the maximum temperature.

An efficient algorithm to estimate optimal preform die shape parameters in forging

An optimisation method for design of intermediate die shapes needed in some forging operations is presented. The basic problem consists of finding an optimal two‐step forging sequence by automatically designing the shape of the preforming tools. The optimisation problem is defined based on an inverse formulation. The objective function of the optimisation problem is a function describing the quality of the obtained part by measuring the die underfill. The finite element method is used to simulate the forging problem. The optimisation method is based on a modified sequential unconstrained minimisation technique and a gradient method. The sensitivity‐dependent algorithm requires computing the derivatives of the objective function with respect to the design variables defining the preform shapes. A direct differentiation method has been developed for this purpose. The optimisation scheme is demonstrated with two axisymmetric forging examples in which optimal preform dies are obtained.

Toward hemodynamic diagnosis of carotid artery stenosis based on ultrasound image data and computational modeling

AbstractThe ability of using non-expensive ultrasound (US) image data together with computer fluid simulation to access various severities of carotid stenosis was inquired in this study. Subject-specific hemodynamic conditions were simulated using a developed finite element solver. Individual structured meshing of the common carotid artery (CCA) bifurcation was built from segmented longitudinal and cross-sectional US images; imposed boundary velocities were based on Doppler US measurements. Simulated hemodynamic parameters such as velocities, wall shear stress (WSS) and derived descriptors were able to predict disturbed flow conditions which play an important role in the development of local atherosclerotic plaques. Hemodynamic features from six individual CCA bifurcations were analyzed. High values of time-averaged WSS (TAWSS) were found at stenosis site. Low values of TAWSS were found at the bulb and at the carotid internal and external branches depending on the particular features of each patient. High oscillating shear index and relative residence time values assigned highly disturbed flows at the same artery surface regions that correlate only moderately with low TAWSS results. Based on clinic US examinations, results provide estimates of flow changes and forces at the carotid artery wall toward the link between hemodynamic behavior and stenosis pathophysiology.

Blood flow simulation and vascular reconstruction

In medical practice, bypass grafts are commonly used as an alternative route around strongly stenosed or occluded arteries. In contrast to arterial bifurcations, surgically created anastomosis can be modified with the objective of enabling optimal graft geometry to yield a flow environment that improves its longevity. This paper presents a three dimensional numerical study of blood flow through bypass systems with different geometries. Coupled with the finite element solver a shape optimization framework considering a genetic algorithm is presented. Numerical results show the benefits of understanding blood flow hemodynamic at anastomosis junctions achieving design improvements. Minimizing recirculation zones and flow stagnation can be useful in surgical planning.

Blood Flow Simulation and Applications

In the vascular system altered flow conditions, such as separation and flow-reversal zones play an important role in the development of arterial diseases. Nowadays computational biomechanics modeling is still in the research and development stage. This chapter presents a numerical computational methodology for blood flow simulation using the Finite Element method outlining field equations and approaches for numerical solutions. Due to the complexity of the vascular system simplifying assumptions for the mathematical modeling process are made. Two applications of the developed tool to describe arterial hemodynamics are presented, a flow simulation in the human carotid artery bifurcation and a search for an optimized geometry of an artificial bypass graft.

Haemodynamic conditions of patient-specific carotid bifurcation based on ultrasound imaging

The purpose of this paper is to complement the characterisation of patient-specific carotid artery bifurcation haemodynamics based on image data obtained by Doppler ultrasound imaging. A methodology for patient-specific 3D luminal surface reconstruction followed by structured hexahedral meshing of the volume and blood flow simulation is presented. Quantitative descriptors of the flow based on wall shear stress (WSS) are used to compare healthy and stenosed carotid bifurcation haemodynamic disturbances. Independently on the presence of stenosis, the internal carotid artery has been identified as a region of abnormal high values of oscillating shear index and relative residence time and low values of time averaged WSS. For the healthy carotid bifurcation, WSS descriptors manage to capture flow disturbances at the external carotid artery. This work addresses the lack of quantitative analysis on anatomically realistic stenosed carotid bifurcations.

Computational simulation of carotid stenosis and flow dynamics based on patient ultrasound data – A new tool for risk assessment and surgical planning

PURPOSE There is nowadays extensive experimental and computational investigation on the pathophysiology of atherosclerosis, searching correlations between its focal nature and local hemodynamic environment. The goal of this work is to present a methodology for patient-specific hemodynamics study of the carotid artery bifurcation based on the use of ultrasound (US) morphological and blood flow velocity patient data. MATERIALS/METHODS Subject-specific studies were performed for two patients, using a developed finite element code. Geometrical models were obtained from the acquisition of longitudinal and sequential cross-sectional ultrasound images and boundary conditions from Doppler velocity measurements at the common carotid artery. RESULTS There was a good agreement between ultrasound imaging data and computational simulated results. For a normal and a stenosed carotid bifurcation the velocity, wall shear stress (WSS) and WSS descriptors analysis illustrated the extremely complex hemodynamic behavior along the cardiac cycle. Different patterns were found, associated with morphology and hemodynamic patient-specific conditions. High values of time-averaged WSS (TAWSS) were found at stenosis site and for both patients TAWSS fields presented low values within areas of high oscillating shear index and relative residence time values, corresponding to recirculation zones. CONCLUSION Simulated hemodynamic parameters were able to capture the disturbed flow conditions in a normal and a stenosed carotid artery bifurcation, which play an important role in the development of local atherosclerotic plaques. Computational simulations based on clinic US might help improving diagnostic and treatment management of carotid atherosclerosis.

Optimization of Forming Processes with Different Sheet Metal Alloys

Over the past decades relatively heavy components made of steel alloys comprise the majority of many manufactured parts due to steel’s low cost, high formability and good strength. The desire to produce lightweight parts has led to studies searching for lighter and stronger materials such as aluminum alloys. However, they exhibit lower elastic stiffness than steel resulting in higher elastic strains causing known distortions such as spring‐back and so decreasing accuracy of manufactured net‐shape components. This paper presents a developed computational method to optimize the design of sheet metal processes using genetic algorithms. An inverse approach is considered so that the final geometry of the bended blank closely follows a prescribed one. The developed computational method couples a finite element forming simulation and an evolutionary algorithm searching the optimal design parameters of the process. The developed method searches the optimal parameters that ensure a perfect net‐shape part. Differen...

Correlation between geometric parameters of the left coronary artery and hemodynamic descriptors of atherosclerosis: FSI and statistical study

AbstractThe hemodynamics conditioned by coronary geometry may play an important role in the creation of a pro-atherogenic environment in specific locations of the coronary tree. The aim of this study is to identify how several geometric parameters of the left coronary artery – cross-section areas, proximal left anterior descending artery length, angles between the branches and the septum, curvature and tortuosity – can be related with hemodynamic descriptors, using a computational fluid–structure interaction method. It is widely accepted that the hemodynamic indicators play an important role in identifying possible pro-atherogenic locations. A statistical study, using Pearson correlation coefficient and P value, was performed for a population study of 8 normal human left coronary arteries presenting right-dominant circulation. Within the study cases, arteries with high caliber (r = 0.88), high angles LMS-LAD (r = 0.49), LAD-LCx (r = 0.57) and LAD-Septum (r = 0.52), and high tortuosity LMS-LCx (r = 0.63) were correlated with a hemodynamic behavior propitious to plaque formation in the left anterior descending artery. In contrast, high proximal left anterior descending artery length (r = −0.41), high angle LMS-LCx (r = −0.59), high tortuosity LMS-LAD (r = −0.56) and LAD-LCx (r = −0.55) and high curvature of LMS (r = −0.60) and LCx (r = −0.56) can lead to non-favorable hemodynamic conditions for atheroma formation. Graphical abstract