Validation Of Computational Models Research Articles

Cerebral Aneurysm Occlusion at 12-Month Follow-Up After Flow-Diverter Treatment: Statistical Modeling for V&V With Real-World Data

Background: Flow-Diverter (FD) porosity has been pointed as a critical factor in the occlusion of cerebral aneurysms after treatment.Objective: Verification and Validation of computational models in terms of predictive capacity, relating FD porosity and occlusion after cerebral aneurysms treatment.Methods: Sixty-four aneurysms, with pre-treatment and follow-up images, were considered. Patient demographics and aneurysm morphological information were collected. The computational simulation provided by ANKYRAS provided FD porosity, expansion, and mesh angle. FD occlusion was assessed and recorded from follow-up images. Multiple regression Logit and analysis of covariance (ANCOVA) models were used to model the data with both categorical and continuous models.Results: Occlusion of the aneurysm after 12 months was affected by aneurysm morphology but not by FD mesh morphology. A Time-To-Occlusion (TTO) of 6.92 months on average was observed with an SE of 0.24 months in the aneurysm population surveyed. TTO was estimated with statistical significance from the resulting model for the data examined and was capable of explaining 92% of the data variation.Conclusions: Porosity was found to have the most correction power when assessing TTO, proving its importance in the process of aneurysm occlusion. Still, further Verification and Validation (V&V) of treatment simulation in more extensive, multi-center, and randomized databases is required.

Effect of artificial surface shapes and their malpositioning on the mechanics of the replaced ankle joint for possible better prosthesis designs

BackgroundThe clinical outcomes of total ankle replacement are limited by prosthesis component malpositioning during surgery. The goal of this study is to assess the mechanical impact of this malpositioning in a validated computer model. MethodsIn a previously developed multi-body dynamic model of the human ankle complex three different artificial implants were designed, each one presenting a different approximation of the natural articular surfaces of the corresponding specimen. The most common implant translational and rotational malpositionings were defined and mimicked. Dynamic simulations of joint motion were run for the various surfaces and malpositionings. The same input loading conditions derived from a previous in-vitro experiment on the corresponding natural specimen were applied. FindingsFrom load-displacement graphs it was observed that all three artificial surfaces reproduced well physiological motion between the calcaneus and the tibia/fibula, with a maximum difference of 2°. It was found that antero-posterior translation of either the tibial or the talar component and inclination of the tibial component in the sagittal plane led to considerable increases in the range of motion. Antero-posterior and dorsiflexion of the tibial component resulted in an increased internal-external rotation by up to 3.5° and 4.0°, respectively. The corresponding increase of inversion-eversion was 5.0° and 6.5°. InterpretationThis study showed that relatively small surgical errors have great consequences in replaced joint mechanics. The present model can be used in future studies to analyse the effect of malpositioning with any specific current total ankle prosthesis.

Effect of inflow conditioning for dry powder inhalers

The transport of pharmaceutical dry powder inside an optically accessible inhaler-like device is studied using both macro- and microscopic high-speed imaging. The investigation aims to systematically study the effect of inflow modifications on the dispersion characteristics of agglomerates inside a dry powder inhaler (DPI) geometry. An inhaler device was designed with geometrical features akin to commercial inhalers used in the current market and research oriented inhalers such as the Twincer®: two offset inlet channels (one with a powder pocket), a clockwise swirling chamber and a single outlet channel. At the device outlet, a vacuum pump was fitted with an actuator and calibrated to achieve a steady state inhalation with a peak flowrate of 85 and 125 L/min. Airflow conditions at the intake of the device were strategically perturbed in order to induce powder fluidisation and dispersion using turbulence grids and through physically obstructing channel streams in order to achieve changes in flow behaviour (e.g., flow separation). Complete fluidisation of the powder bed was observed with image processing enabling statistics on de-agglomerated fragment size and velocity. A range of behaviour was noted including local turbulence through introduction of a grid, bimodal fragment size behaviour for cohesive mannitol powder, as well as introduction of low velocity zones in the device through flow splitting. The geometry enables simple systematic study of inflow conditions into a DPI-like device with the data being useful for study of a given powder formulation (mannitol) and validation of computational models.

Measurement of brain simulant strains in head surrogate under impact loading.

Impact-induced traumatic brain injury (TBI) is a major source of disability and mortality. Knowledge of brain strains during impact (accelerative) loading is critical for the overall management of TBI, including the development of injury thresholds, personal protective equipment, and validation of computational models. Despite these needs, the current understanding of brain strains in humans or humanlike surrogates is limited, especially for injury causing loading magnitudes. Toward this end, we measured full-field, in-plane (2D) strains in a brain simulant using the hemispherical head surrogate. The hemispherical head was mounted on the Hybrid-III neck and subjected to impact loading using a linear impactor system. The resulting head kinematics was measured using a triaxial accelerometer and angular rate sensors. Dynamic, 2D strains in a brain simulant were obtained using high-speed imaging and digital image correlation. Concurrent finite element (FE) simulations of the experiment were also performed to gain additional insights. The role of stiff membranes of the head was also studied using experiments. Our results suggest that rotational modes dominate the response of the brain simulant. The wave propagation in the brain simulant as a result of impact has a timescale of ~100ms. We obtain peak strains of ~20%, ~40%, ~60% for peak rotational accelerations of ~838, ~5170, ~11,860rad/s2, respectively. Further, peak strains in cortical regions are higher than subcortical regions by up to ~70%. The agreement between the experiments and FE simulations is reasonable in terms of spatiotemporal evolution of strain pattern and peak strain magnitudes. Experiments with the addition of falx and tentorium indicate significant strain concentration (up to 115%) in the brain simulant near the interface of falx or tentorium and brain simulant. Overall, this work provides important insights into the biomechanics of strain in the brain simulant during impact loading.

How to Validate in silico Deployment of Coronary Stents: Strategies and Limitations in the Choice of Comparator.

This study aims at proposing and discussing useful indications to all those who need to validate a numerical model of coronary stent deployment. The proof of the reliability of a numerical model is becoming of paramount importance in the era of in silico trials. Recently, the ASME V&V Standard Committee for medical devices prepared the V&V 40 standard document that provides a framework that guides users in establishing and assessing the relevance and adequacy of verification and validation activities performed for proving the credibility of models. To the knowledge of the authors, only a few examples of the application of the V&V 40 framework to medical devices are available in the literature, but none about stents. Specifically, in this study, the authors wish to emphasize the choice of a relevant set of experimental activities to provide data for the validation of computational models aiming to predict coronary stent deployment. Attention is focused on the use of ad hoc 3D-printed mock vessels in the validation plan, which could allow evaluating aspects of clinical relevance in a representative but controlled environment.

Parametric Regression Applied for Determination of Electrical Parameters of Synchronous and Induction Generators Operating in Parallel on the Electrical Energy Repowering System

The purpose of this work is to determine the values of electrical parameters of synchronous and induction machines to validate electrical interactions between an induction generator and a synchronous generator. The generators are connected in two ways: (i) isolated from the common bus and (ii) parallel to the common bus in steady state, subject to nonlinear load. They are old and refurbished machines; thus, the parametric regression methodology is used to determine the electrical parameter values. After the computational model validation, analyses are performed by various system configurations to confirm the repowering and to analyze the system harmonic current flow. The results obtained comparing the computer simulation and experimental tests prove that the validated model represents the real system. With an experimentally validated computational model, it is possible to verify the occurrence of system repowering and the increased value of harmonic distortions in the induction generator terminals, acting as a preferential path for harmonic currents.

Computational Modeling of the Impact of Solid Particles on the Gas Hold-up in Slurry Bubble Columns

Three-phase bubble column reactors, despite their broad diffusion in chemical process, are often difficult to describe given the intricate mutual phase interactions. Moreover, validation of computational models is likewise complex, since it is considerably challenging to obtain exhaustive and precise experimental data from these systems. The aim of this work is to predict reliably the effect of disperse solid particles on the overall fluid dynamics of a bubble column by computational fluid dynamics.The small size of particles allowed us to approximate the solid-liquid mixture as a single pseudo-homogeneous phase in the Euler-Euler framework. The simulation settings were tuned on two-phase gas-liquid systems and then extended to the three-phase gas-solid-liquid columns. Experimental data from different setups and correlations for the global gas hold-up in slurry systems were used to validate the model. In particular, what emerges from experiments is that the presence of the solid reduces gas hold-up, as a consequence of the higher density of the slurry in comparison to the pure liquid and the increased coalescence of the bubbles.Results show that this model is able to predict the reduction of the hold-up with a solid volumetric loading up to 20%, achieving good agreement with experimental data. Moreover, what stands out from the simulations is that the addition of the solid also changes the shape of the flow map curve (global hold-up as a function of the gas superficial velocity), switching to a smoother transition between the behavior at low and high gas superficial velocity.

Severe Calcaneus Injury Probability Curves Due to Under-Body Blast.

The lower extremity is the most frequently injured body region to mounted soldiers during underbody blast (UBB) events. UBB events often produce large deformations of the floor and subsequent acceleration of the lower limb that are not sufficiently mitigated by the combat boot, leaving the calcaneus bone vulnerable to injury. Biomechanical experiments simulating UBB loading scenarios were conducted in a laboratory environment using isolated postmortem human subject (PMHS) leg components. Each leg component was tested twice: one sub-injurious test followed by a injury-targeted test. This enabled the use of interval censoring for each specimen in the survival statistical analysis to generate the human injury probability curves (HIPCs). Foot contact forces were measured in both the hindfoot and forefoot. Strains and acoustic emission signals at the calcaneus and distal tibia were utilized to determine injury timing. The footplate velocities of the injury tests ranged 8-13 m/s with time-to-peak velocity of 1.8-2.5 ms while the velocities of non-injury tests ranged from 4 to 6 m/s with the same time-to-peak. The majority of the injuries were severe calcaneus fractures (Sanders III-IV). Secondary injuries included fractures to the distal tibia, talus, cuboid and cuneiform. These injury outcomes were found to be consistent with those reported in UBB injury literature. The HIPCs for the severe calcaneus fracture were developed using the vertical heel contact force as the injury correlation measure through survival analysis statistical method in the form of lognormal function. This work represents the first set of HIPCs dedicated to the severe calcaneus fracture using the biomechanical force measurement closest to the injury location. This injury probability curve will enable biomechanical response validation of computational models, development of ATD injury assessment reference curve, and injury prediction capability for computational models or ATDs in the UBB environment.

Tests of circular concrete-filled steel tubular stub columns with artificial notches representing local corrosions

Concrete-filled steel tubular (CFST) columns have been used in civil structures due to their affordability, high strength, superior ductility, and demonstrated fire resistance. However, there are remaining issues regarding the performances of the CFST columns with the locally corroded steel tubes, hindering the widespread acceptance of this type of members in practice. This paper focused on the circular CFST stub columns in which the steel tubes are locally corroded. To develop the critical data needed for understanding the behavior of such stub columns, a group of thirty specimens were designed, constructed and tested. Based on the recommendations from prior research, the local corrosions in the steel tubes of the specimens were represented by the artificial notches. The test parameters varied among the specimens included the notch orientations, lengths, depths, and widths. The test results revealed the influences of these parameters on different aspects of the responses of the stub columns with the corroded steel tubes. Computer models were also developed for the specimens and validated by the test results. Based on the test results and the results from parametric analyses using the validated computer models, a practical model to compute the peak resistance of a CFST stub column with the locally corroded steel tube was proposed and verified.

Advanced full-scale laboratory dynamic load testing of a ballasted high-speed railway track

Understanding ballast layer dynamic response and long-term behavior under moving train loads is necessary for optimization of railway track performance. This paper discussed an advanced testing effort undertaken in a controlled laboratory environment for the measurement of dynamic load responses via multitude of installed sensors and the development of validated computer models and simulation tools suitable for analyzing full-scale ballasted track under dynamic loading. Constructed in a large rectangular metal frame, the laboratory model consisted of a full-scale track with eight ties, ballast, subballast, and embankment. Three different speeds and axle load configurations were applied sequentially onto the full-scale ballasted track using eight actuators, which realistically captured both slow moving freight and high-speed passenger train loads. Vibration velocities were captured at different locations on ballast surfaces and ties. Dynamic soil stresses were recorded at the bottom of the ballast and subballast layers. Both transient and permanent deformations due to repeated dynamic loads were measured in the track substructure layers and for all eight ties. “SmartRock” advanced sensors were used for the first time to track ballast particle movements under dynamic loading. Further, the test data from the experiments were compared with preliminary Discrete Element Method (DEM) simulations.

The importance of uncertainty quantification in model reproducibility.

Many computer models possess high-dimensional input spaces and substantial computational time to produce a single model evaluation. Although such models are often ‘deterministic’, these models suffer from a wide range of uncertainties. We argue that uncertainty quantification is crucial for computer model validation and reproducibility. We present a statistical framework, termed history matching, for performing global parameter search by comparing model output to the observed data. We employ Gaussian process (GP) emulators to produce fast predictions about model behaviour at the arbitrary input parameter settings allowing output uncertainty distributions to be calculated. History matching identifies sets of input parameters that give rise to acceptable matches between observed data and model output given our representation of uncertainties. Modellers could proceed by simulating computer models’ outputs of interest at these identified parameter settings and producing a range of predictions. The variability in model results is crucial for inter-model comparison as well as model development. We illustrate the performance of emulation and history matching on a simple one-dimensional toy model and in application to a climate model.This article is part of the theme issue ‘Reliability and reproducibility in computational science: implementing verification, validation and uncertainty quantification in silico’.

Computational FEM Model and Phantom Validation of Microwave Ablation for Segmental Microcalcifications in Breasts Using a Coaxial Double-Slot Antenna.

Introduction Cancer is the second leading cause of death worldwide. Breast cancer is the second most common cause of cancer-related mortality, accounting for 11.6% of the total number of deaths. The main treatments for this disease are surgical removal of the tumor, radiotherapy, and chemotherapy. Recently, different minimally invasive technologies have been applied (e.g., emission of electromagnetic waves, thermal and chemical means) to overcome the important side effects of these treatment modalities. The objective of this study was to develop and evaluate a predictive computational model of microwave ablation. Materials and Methods The predictive computational model of microwave ablation was constructed by means of a dual-slot coaxial antenna. The model was compared with an experiment performed using a breast phantom, which emulates the dielectric properties of breast tissue with segmental microcalcifications. The standing wave ratio (SWR) was obtained for both methods to make a comparison and determine the feasibility of applying electromagnetic ablation to premalignant lesions in breasts. Specifically, for the analysis of segmental microcalcifications, a breast phantom with segmental microcalcifications was developed and two computational models were performed under the same conditions (except for blood perfusion, which was excluded in one of the models). Results The SWR was obtained by triplicate experiments in the phantom, and the measurements had a difference of 0.191 between the minimum and maximum SWR values, implying a change of power reflection of 0.8%. The average of the three measurements was compared with the simulation that did not consider blood perfusion. The comparison yielded a change of 0.104, representing a 0.2% change in power reflection. Discussion. Both experimentation in phantom and simulations demonstrated that ablation therapy can be performed using this antenna. However, an additional optimization procedure is warranted to increase the efficiency of the antenna.

Multiscale modeling of twitch contractions in cardiac trabeculae.

Understanding the dynamics of a cardiac muscle twitch contraction is complex because it requires a detailed understanding of the kinetic processes of the Ca2+ transient, thin-filament activation, and the myosin-actin cross-bridge chemomechanical cycle. Each of these steps has been well defined individually, but understanding how all three of the processes operate in combination is a far more complex problem. Computational modeling has the potential to provide detailed insight into each of these processes, how the dynamics of each process affect the complexity of contractile behavior, and how perturbations such as mutations in sarcomere proteins affect the complex interactions of all of these processes. The mechanisms involved in relaxation of tension during a cardiac twitch have been particularly difficult to discern due to nonhomogeneous sarcomere lengthening during relaxation. Here we use the multiscale MUSICO platform to model rat trabecular twitches. Validation of computational models is dependent on being able to simulate different experimental datasets, but there has been a paucity of data that can provide all of the required parameters in a single experiment, such as simultaneous measurements of force, intracellular Ca2+ transients, and sarcomere length dynamics. In this study, we used data from different studies collected under similar experimental conditions to provide information for all the required parameters. Our simulations established that twitches either in an isometric sarcomere or in fixed-length, multiple-sarcomere trabeculae replicate the experimental observations if models incorporate a length-tension relationship for the nonlinear series elasticity of muscle preparations and a scheme for thick-filament regulation. The thick-filament regulation assumes an off state in which myosin heads are parked onto the thick-filament backbone and are unable to interact with actin, a state analogous to the super-relaxed state. Including these two mechanisms provided simulations that accurately predict twitch contractions over a range of different conditions.

Verification and Validation of Computational Models Used in Biopharmaceutical Manufacturing: Potential Application of the ASME Verification and Validation 40 Standard and FDA Proposed AI/ML Model Life Cycle Management Framework.

A wide variety of computational models covering statistical, mechanistic, and machine learning (locked and adaptive) methods are explored for use in biopharmaceutical manufacturing. Limited discussion exists on how to establish the credibility of a computational model for application in biopharmaceutical manufacturing. In this work, we tried to use the American Society of Mechanical Engineers (ASME) Verification and Validation 40 (V&V 40) standard and FDA proposed AI/ML model life cycle management framework for Software as a Medical Device (SaMD) in biopharmaceutical manufacturing use cases, by applying to a set of curated hypothetical examples. We discussed the need for standardized frameworks to facilitate consistent decision making to enable efficient adoption of computational models in biopharmaceutical manufacturing and alignment of existing good practices with existing frameworks. In the study of our examples, we anticipate existing frameworks like V&V 40 can be adopted.

Subglottal pressure oscillations in anechoic and resonant conditions and their influence on excised larynx phonations

Excised larynges serve as natural models for studying behavior of the voice source. Acoustic resonances inside the air-supplying tubes below the larynx (i.e., subglottal space), however, interact with the vibratory behavior of the larynges and obscure their inherent vibration properties. Here, we explore a newly designed anechoic subglottal space which allows removing its acoustic resonances. We performed excised larynx experiments using both anechoic and resonant subglottal spaces in order to analyze and compare, for the very first time, the corresponding subglottal pressures, electroglottographic and radiated acoustic waveforms. In contrast to the resonant conditions, the anechoic subglottal pressure waveforms showed negligible oscillations during the vocal fold contact phase, as expected. When inverted, these waveforms closely matched the inverse filtered radiated sound waveforms. Subglottal resonances modified also the radiated sound pressures (Level 1 interactions). Furthermore, they changed the fundamental frequency (fo) of the vocal fold oscillations and offset phonation threshold pressures (Level 2 interactions), even for subglottal resonance frequencies 4–10 times higher than fo. The obtained data offer the basis for better understanding the inherent vibratory properties of the vocal folds, for studying the impact of structure-acoustic interactions on voice, and for validation of computational models of voice production.

Anatomy of a Psychological Theory: Integrating Construct-Validation and Computational-Modeling Methods to Advance Theorizing

Discussions about the replicability of psychological studies have primarily focused on improving research methods and practices, with less attention paid to the role of well-specified theories in facilitating the production of reliable empirical results. The field is currently in need of clearly articulated steps to theory specification and development, particularly regarding frameworks that may generalize across different fields of psychology. Here we focus on two approaches to theory specification and development that are typically associated with distinct research traditions: computational modeling and construct validation. We outline the points of convergence and divergence between them to illuminate the anatomy of a scientific theory in psychology—what a well-specified theory should contain and how it should be interrogated and revised through iterative theory-development processes. We propose how these two approaches can be used in complementary ways to increase the quality of explanations and the precision of predictions offered by psychological theories.

GMSH and GetDP Software: Validation of Computer Models to Calculating Electromagnetic Force and Torque

The purpose of this article is to assess the reliability of the particular solutions of electromechanical problems obtained in practice by grid methods using the open source software GMSH & GetDP. The paper considers the issues of validation of computer models for calculating the electromagnetic force (EMF) and torque (EMT) obtained in the GMSH & GetDP software package. The results of calculations and compares for electromagnet-ic devices for the TEAM 20 and TEAM 24 test problems from the Compumag international community problems package are presented. A comparative assessment of the obtained results with the test ones was carried out. The relative average deviation between the measured and calculated values is: 7.6 % for induction; 7.5 % for the electromagnetic torque; the maximum deviation in electromagnetic force does not exceed 3.4%. The efficiency of using the GMSH & GetDP software for solving magnetostatic problems in a three-dimensional setting is shown. The presented results will be used to form the verification and validation matrices in assessing the reliability of numerical models, as well as in the development of digital twins of electromechanical devices

The moment arms and leverage of the human finger muscles

The moment arm of a muscle’s force represents the muscle’s leverage or mechanical advantage in producing a joint moment. It is indicative of the muscle’s potential to contribute to actuation of a joint in a particular joint motion direction and defines the role of the muscle, for example, as a joint flexor or abductor. The aims of this study were, firstly, to measure the moment arms of the flexor and extensor muscles of the metacarpophalangeal (MCP), proximal interphalangeal (PIP) and distal interphalangeal (DIP) joints, and the moment arms of the major abductor and adductor muscles of the metacarpophalangeal (MCP) joint of each finger in the hand; secondly, to assess the effect of change in joint angle on these moment arms; and thirdly, to determine if there are differences in a given flexor or extensor’s muscle moment arms between the joints it spans on a given finger, and across its tendon slips to multiple fingers. The tendon-excursion method was used to measure instantaneous muscle moment arms in nine fresh-frozen entire forearm cadaver specimens. Joint flexion angle was found to have significant effects on the moment arms of the extensor muscles at the MCP and PIP joints (p < 0.05). In contrast, the digital flexor muscles maintained relatively constant moment arms through the range of joint flexion. The moment arms of the digital flexors and extensors spanning multiple joints in a finger were largest at the MCP joints and smallest at the DIP joints. The findings demonstrate greater torque generating capacity for tasks such as grasping at the proximal interphalangeal joints, and smaller torque capacity for finer movement control at the distal interphalangeal joints. The dataset generated in this study may be useful in the development and validation of computational models used in surgical planning, and rehabilitation.

Predicting cold gas-solid flow in a pilot-scale dual-circulating fluidized bed: Validation of computational particle fluid dynamics model

This study uses the Eulerian-Lagrangian multiphase particle-in-cell (MP-PIC) approach, adopted in the computational particle fluid dynamics (CPFD) code, Barracuda VR®, to simulate a cold-flow dual-circulating fluidized bed. The hydrodynamics is simulated using a novel segregated approach, in which the experimental solids circulation rate is specified as an input parameter. CPFD predictions of pressures and particle concentration distributions are compared with experimental results for varying bed material circulation rate of 76–153 kg/m2s and superficial gas velocity in the reactor riser from 3.3 m/s to 6.4 m/s. Overall trends in these quantities are well captured by the model. Detailed sensitivity studies of the reactor riser are carried out using a riser-only approach with a fixed bed mass inventory. Apart from the drag model selection, the pressure constant Ps and the exponent β of the MP-PIC particle stress model are found to have the greatest influence on the fluidization behavior and on the predicted solids flux.

Computational modelling of survival of Aspergillus flavus in peanut kernels during hot air-assisted radio frequency pasteurization

In recent years, radio frequency (RF) heating is getting popular as an alternative pasteurization method for agricultural commodities and low moisture foods. Computer simulation is an effective way to help understand RF interactions with food components and predict temperature distributions among food samples after RF treatments. In this study, a computer model based on Joule heating and thermal inactivation kinetic of A. flavus was established to predict both temperature distribution and microbial reduction among peanut kernels after RF processing. For the process validation, three 2-g peanut samples inoculated with 40 μL A. flavus were placed at three representative locations among 2.17 kg peanut kernels and subjected to various processing conditions in a 27.12 MHz, 6 kW RF heating unit together with hot air system. Results showed that the average difference of the sample temperature and microbial reduction between simulation and experiment was small with RMSE values of 0.009 °C and 0.012 °C, and 0.31 log CFU/g and 0.42 log CFU/g for peanut moisture contents of 7.56% and 12.02% w. b., respectively. Nonuniform RF heating resulted in the least lethality of A. flavus at the cold spot. The validated computer model was further used to estimate microbial reduction distributions at other target temperatures based on predicted temperature profiles. This computer model may help design the RF pasteurization protocols for peanut kernels without extensive experiments in food industry.