Computational Effort Research Articles

Graph Neural Networks to Simulate Flexible Pavement Responses Using Three-Dimensional Finite Element Analysis Data

The significant computational costs and efforts required for accurate three-dimensional (3D) finite element (FE) pavement response calculations necessitate an expedited approach. This study proposes a graph neural network (GNN)-based simulator for the modeling of 3D pavement structural responses under tire loading. The GNN model was trained using 240 simulations of 3D pavement FE data of flexible pavement structures. The simulator represented the state of pavement structure meshes in FE analysis at any given timestep as a graph, with FE nodes encoded as graph nodes and mesh edges as graph edges. The dynamic behaviors of pavement FEs were computed via learned message-passing between two graphs within two continuous timesteps. The one-step mean squared error (MSE) and rollout MSE were used as evaluation metrics for the GNN model. The results showed that, given an initial state of FE responses, the model could perform accurate one-step predictions, extending to trajectory predictions with one-step MSE as low as [Formula: see text] and rollout MSE around [Formula: see text]. The prediction framework is efficient; it requires a week of model training but only a mere 5 min of prediction for each single case. This a contrast to traditional 3D FE analyses that can span hours to weeks for a single case. The hyperparameters, including the number of message-passing steps M and the number of historical timesteps C, were established as 10 and 1, respectively, based on model performance and computation time. It was also observed that data normalization before training could significantly reduce model simulation noise.

Robust Automated Tumour Segmentation Network Using 3D Direction-Wise Convolution and Transformer

AbstractSemantic segmentation of tumours plays a crucial role in fundamental medical image analysis and has a significant impact on cancer diagnosis and treatment planning. UNet and its variants have achieved state-of-the-art results on various 2D and 3D medical image segmentation tasks involving different imaging modalities. Recently, researchers have tried to merge the multi-head self-attention mechanism, as introduced by the Transformer, into U-shaped network structures to enhance the segmentation performance. However, both suffer from limitations that make networks under-perform on voxel-level classification tasks, the Transformer is unable to encode positional information and translation equivariance, while the Convolutional Neural Network lacks global features and dynamic attention. In this work, a new architecture named TCTNet Tumour Segmentation with 3D Direction-Wise Convolution and Transformer) is introduced, which comprises an encoder utilising a hybrid Transformer-Convolutional Neural Network (CNN) structure and a decoder that incorporates 3D Direction-Wise Convolution. Experimental results show that the proposed hybrid Transformer-CNN network structure obtains better performance than other 3D segmentation networks on the Brain Tumour Segmentation 2021 (BraTS21) dataset. Two more tumour datasets from Medical Segmentation Decathlon are also utilised to test the generalisation ability of the proposed network architecture. In addition, an ablation study was conducted to verify the effectiveness of the designed decoder for the tumour segmentation tasks. The proposed method maintains a competitive segmentation performance while reducing computational effort by 10% in terms of floating-point operations.

This is SpArta: Rigorous Optimization of Regionally Resolved Energy Systems by Spatial Aggregation and Decomposition

Energy systems with high shares of renewable energy are characterized by local variability and grid limitations. The synthesis of such energy systems, therefore, requires models with high spatial resolution. However, high spatial resolution increases the computational effort. Here, we present the SpArta method for rigorous optimization of regionally resolved energy systems by Spatial Aggregation and decomposition. SpArta significantly reduces computational effort while providing a near-optimal solution at the full spatial resolution of sector-coupled energy systems.SpArta first reduces problem size by spatially aggregating the energy system using clustering. The aggregated problem is then relaxed and restricted to obtain a lower and an upper bound. The spatial resolution is iteratively increased until the difference between upper and lower bound satisfies a predefined optimality gap.Finally, each cluster of the aggregated problem is redesigned at full resolution. For this purpose, SpArta decomposes the original synthesis problem into subproblems for each cluster. Combining the redesigned cluster solutions yields an optimal feasible solution of the full-scale problem within a predefined optimality gap. SpArta thus optimizes large-scale energy systems rigorously with significant reductions in computational effort. We apply SpArta to a case study of the sector-coupled German energy system, reducing the computational time by a factor of 7.5, compared to the optimization of the same problem at full spatial resolution. As SpArta shows a linear increase in computational time with problem size, SpArta enables computing larger problems allowing to resolve energy system designs with improved accuracy.

Mapping flagellated swimmers to surface-slip driven swimmers

Flagellated microswimmers are ubiquitous in natural habitats. Understanding the hydrodynamic behavior of these cells is of paramount interest, owing to their applications in bio-medical engineering and disease spreading. Since the last two decades, computational efforts have been continuously improved to accurately capture the complex hydrodynamic behavior of these model systems. However, modeling the dynamics of such swimmers with fine details is computationally expensive due to the large number of unknowns and the small time-steps required to solve the equations. In this work we propose a method to map fully resolved flagellated microswimmers to coarse, active slip driven swimmers which can be simulated at a reduced computational cost. Using the new method, the slip driven swimmers move with the same velocity, to machine precision, as the flagellated swimmers and generate a similar flow field with a controlled accuracy. The method is validated for swimming patterns near a no-slip boundary, interactions between swimmers and scattering with large obstacles.

A novel lightweight dynamic focusing convolutional neural network LAND-FCNN for EEG emotion recognition

The inefficiency of model inference and the limited sample data are significant issues in electroencephalogram (EEG) emotion recognition models based on neural networks. Therefore, we propose a novel lightweight adaptive dynamic focusing convolutional neural network (LAND-FCNN). We use Partial Convolution (PConv) and Batch interactive attention in constructing the Backbone network. PConv reduces the computational cost by selectively extracting features from a subset of input channels, utilizing feature graph redundancy. Batch interactive attention establishes connections between various types of samples to address sample scarcity without relying on data augmentation. Then Backbone serves as the encoder for the Glance and Focus Dynamic Inference Networks (GFnet). The patch proposal network in GFnet adaptively captures task-relevant areas of each sample, allowing encoders to process only those areas. This increases inference speed and ensures reliable results with minimal computational effort. Extensive validation of diverse datasets with varying distributions confirms the efficacy and precision of LAND-FCNN.

Improving analytic approximations of TSP tour lengths with adjustment factors

Optimizing Traveling Salesman Problem (TSP) tours requires substantial computational effort, leading researchers to develop approximations relating tour length to the number of visited points, n. Existing models, such as the √nA predictor, effectively approximate tour lengths for large-capacity vehicles but sacrifice accuracy for small n values relevant for most practical applications. Consequently, this study addresses this gap by proposing models with uniform node distributions, which incorporate realistic factors, such as central vs. random starting points and various service zone shapes. These factors are then integrated into a single equation, enhancing applicability. Furthermore, the exponent of n is statistically estimated to be significantly different from 0.5, challenging previous studies. Our proposed model estimates TSP tour lengths more accurately, particularly for small n values, and maintains accuracy for large n values, with errors below 3.11% for up to 600 points. This model offers a more precise and versatile alternative to current models.

Modeling gas flow in low-permeability formations: An efficient combination of mixed finite elements and high order time integration schemes

Numerical simulation of gas flow in low permeability formations is a challenging task due to the high nonlinearity induced by (i) the compressibility of the gas, (ii) the Klinkenberg slippage effect and (iii) the Langmuir adsorption of the gas on the pore surface. Because of these nonlinearities, modeling gas flow in low permeability formations requires a great deal of computational effort. In this work, we develop an efficient numerical model using advanced spatial and temporal discretization methods for a simultaneous solution of the coupled equations of gas flow, cubic Peng-Robinson equation of state, slippage effect and Langmuir adsorption. The spatial discretization is performed with the lumped hybrid formulation of the mixed finite element method which is well adapted for fluid flow in heterogeneous porous media. The time integration is performed with high-order methods via the method of lines (MOL) which allows large time steps and efficient solution of the nonlinear system of equations. Numerical experiments, performed for gas extraction in a heterogeneous domain, point out the high efficiency of the new model since it can be until 10 times more efficient than the classical first-order time discretization method. Results of global sensitivity analysis show that the intrinsic gas-phase permeability and the Klinkenberg factor are the most influential parameters controlling the pumping rate in the case of gas extraction in a homogeneous domain.

Detection of tuberculosis using customized MobileNet and transfer learning from chest X-ray image

One of the most contagious diseases in the world, tuberculosis (TB) is brought on by the bacteria Mycobacterium tuberculosis. This hazardous scenario can cause life losses and requires expert doctors and several hours to detect the disease. Using the MobileNet transfer learning model, a computationally lightweight model has been proposed in this study. The optimal model for the diagnosis of tuberculosis has been determined after testing numerous variants on the base model with pre-trained weights. A computationally light transfer learning model is proposed to obtain the maximum overall accuracy of 98.66%. The improvement over the best existing model is quite significant. The transfer learning model (DenseNet) utilized in this existing model is based on a very complex convolutional neural network (CNN), and as a result, the model requires greater amounts of time. The performance of the other existing models is relatively less in comparison to our proposed model, and the methods have a number of other drawbacks. Our goal in this work is to create a more accurate model that requires less computational effort. When compared to previous models, our model has a very less number of trainable parameters, which causes the model to converge more quickly and predict more accurately. Our approach also has the benefit of being simply able to modify its weights when the system is further updated with new datasets. Additionally, because of its lightweight architecture, it can be installed on mobile devices as well as used in web-based applications with ease. To analyze and validate the proposed method, we have collected data from Kaggle and used MC, CHN and NIH datasets.

3DFRINet: A Framework for the Detection and Diagnosis of Fracture Related Infection in Low Extremities Based on 18F-FDG PET/CT 3D Images

Fracture related infection (FRI) is one of the most devastating complications after fracture surgery in the lower extremities, which can lead to extremely high morbidity and medical costs. Therefore, early comprehensive evaluation and accurate diagnosis of patients are critical for appropriate treatment, prevention of complications, and good prognosis. 18Fluoro-deoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) is one of the most commonly used medical imaging modalities for diagnosing FRI. With the development of deep learning, more neural networks have been proposed and become powerful computer-aided diagnosis tools in medical imaging. Therefore, a fully automated two-stage framework for FRI detection and diagnosis, 3DFRINet (Three Dimension FRI Network), is proposed for 18F-FDG PET/CT 3D imaging. The first stage can effectively extract and fuse the features of both modalities to accurately locate the lesion by the dual-branch design and attention module. The second stage reduces the dimensionality of the image by using the maximum intensity projection, which retains the effective features while reducing the computational effort and achieving excellent diagnostic performance. The diagnostic performance of lesions reached 91.55% accuracy, 0.9331 AUC, and 0.9250 F1 score. 3DFRINet has an advantage over six nuclear medicine experts in each classification metric. The statistical analysis shows that 3DFRINet is equivalent or superior to the primary nuclear medicine physicians and comparable to the senior nuclear medicine physicians. In conclusion, this study first proposed a method based on 18F-FDG PET/CT three-dimensional imaging for FRI location and diagnosis. This method shows superior lesion detection rate and diagnostic efficiency and therefore has good prospects for clinical application.

Multi-sensor data fusion-enabled lightweight convolutional double regularization contrast transformer for aerospace bearing small samples fault diagnosis

Aiming at the problems of low information utilization and lack of feature mining capability in multi-sensor fusion networks, this study presents a multi-sensor data fusion-enabled lightweight convolutional double regularization contrast transformer for aerospace bearing small samples fault diagnosis. Firstly, a metric termed integrated cliff entropy is devised to assign weights to vibration signals from diverse sensor channels. It aims to enhance the cyclic impulse characteristics within the fused signals, thereby facilitating more precise fault identification. Secondly, a lightweight Diwaveformer architecture is constructed as the backbone of contrast learning. It enables the global and local features of faulty signals to be comprehensively extracted with less computational effort. Finally, a double contrast loss is constructed to optimize the distribution of intra-class and inter-class features to improve the fault identification ability of the network with small samples. Additionally, a discard regularization method is designed to remove the projection head during the contrast learning process, further advancing the model lightweight. Our method achieved accuracies of 95.54% and 92.56% on two aerospace bearing datasets with extremely sparse training samples, which proved its superior performance.

EnTdecker - A Machine Learning-Based Platform for Guiding Substrate Discovery in Energy Transfer Catalysis.

Due to the magnitude of chemical space, the discovery of novel substrates in energy transfer (EnT) catalysis remains a daunting task. Experimental and computational strategies to identify compounds that successfully undergo EnT-mediated reactions are limited by their time and cost efficiency. To accelerate the discovery process in EnT catalysis, we herein present the EnTdecker platform, which facilitates the large-scale virtual screening of potential substrates using machine-learning (ML) based predictions of their excited state properties. To achieve this, a data set is created containing more than 34,000 molecules aiming to cover a vast fraction of synthetically relevant compound space for EnT catalysis. Using this data predictive models are trained, and their aptitude for an in-lab application is demonstrated by rediscovering successful substrates from literature as well as experimental validation through luminescence-based screening. By reducing the computational effort needed to obtain excited state properties, the EnTdecker platform represents a tool to efficiently guide substrate selection and increase the experimental success rate for EnT catalysis. Moreover, through an easy-to-use web application, EnTdecker is made publicly accessible under entdecker.uni-muenster.de.

A low-cost, penalty parameter-free, and pressure-robust enriched Galerkin method for the Stokes equations

This paper proposes a low-cost, penalty parameter-free, and pressure-robust Stokes solver based on the enriched Galerkin (EG) method with a discontinuous velocity enrichment function. The EG method employs the interior penalty discontinuous Galerkin (IPDG) formulation to weakly impose the continuity of the velocity function. However, despite its advantage of symmetry, the symmetric IPDG formulation requires a lot of computational effort to choose an optimal penalty parameter and compute different trace terms. To reduce such effort, we replace the derivatives of the velocity function with its weak derivatives computed by the geometric data of elements. Therefore, our modified EG (mEG) method is a penalty parameter-free numerical scheme that has reduced computational complexity and conserves the optimal convergence orders. Moreover, we achieve pressure robustness for the mEG method by employing a velocity reconstruction operator on the load vector on the right-hand side of the discrete system. The theoretical results are confirmed through numerical experiments with two- and three-dimensional examples.

Experimental investigation of the total flow resistance in emergent and submerged rigid canopy flows

In canopy flows, flow resistance mainly originates from vegetation drag and depends on vegetation characteristics and flow conditions. In the present study, a series of experiments were performed in various hydraulic scenarios with high stem Reynolds numbers (2641 ≤ Red≤ 17333) using relatively sparse rigid canopies, represented with four different dimensionless vegetation densities (0.0044, 0.0098, 0.0174 and 0.0392), on a smooth bed. A novel drag plate mechanism was developed to measure the total flow resistance due to the emergent and submerged vegetation arrays in a staggered pattern under subcritical flow conditions. Manning’s roughness coefficient and Darcy–Weisbach friction factor were adopted to represent the total flow resistance in the analyses. Simple empirical relationships based on roughness concentration and submergence ratio were derived to determine the total flow resistance parameters within a broad range of stem Reynolds numbers. Although relationships were proposed in a simple form to be used for direct practical applications, they show similar or better performance in the prediction of total flow resistance parameters than the existing equations in the literature, which require considerable computational effort. Additionally, analyses demonstrated that the results of the present study and those of similar studies regarding canopy flow resistance are in good agreement. Accordingly, the novel drag plate looks promising for measuring flow resistance due to vegetation and bed conditions similar to those in nature.

Sesame: A power spectrum emulator pipeline for beyond-ΛCDM models

The mysterious nature of the dark sector of the Λ-cold-dark-matter (ΛCDM) model is one of the main motivators behind the study of alternative cosmological models. A central quantity of interest for these models is the matter power spectrum, which quantifies structure formation on various scales and can be cross-validated through theory, simulations, and observations. Here, we present a tool that can be used to create emulators for the non-linear matter power spectrum, and similar global clustering statistics, for models beyond ΛCDM with very little computation effort and without the need for supercomputers. We use fast approximate N-body simulations to emulate the boost, B(k, z) = Pbeyond − ΛCDM(k, z)/PΛCDM(k, z), and then rely on existing high-quality emulators made for ΛCDM to isolate Pbeyond − ΛCDM(k, z). Since both the ΛCDM and beyond-ΛCDM models are simulated in our approach, some of the lack of power on small scales due to the low force-resolution in the simulations is factored out, allowing us to extend the emulator to k ∼ 3 − 5 h Mpc−1 and still maintain good accuracy. In addition, errors from the simulation and emulation process can easily be estimated and factored into the covariance when using the emulator on data. As an example of using the pipeline, we create an emulator for the well-studied f(R) model with massive neutrinos, using approximately 3000 CPU hours of computation time. Provided with the paper is a fully functioning pipeline that generates parameter samples, runs a Boltzmann solver to produce initial conditions, runs the simulations, and then gathers all the data and runs it through a machine learning module to develop the emulator. This tool, named Sesame, can be used by anyone to generate a power spectrum emulator for the cosmological model of their choice.

Actuator Surface Modeling of Rotors at the Ship–Helicopter Dynamic Interface

The aerodynamics of the ship–helicopter dynamic interface play an important role in determining ship–helicopter operating limits. A review of prior studies, in conjunction with experimental work, indicates that a time-accurate rotor model and a physics-based turbulence-resolving flow solver with two-way coupling between these components are required to accurately predict rotor–airwake and rotor–rotor interactions. Here, a recently developed actuator surface model is validated against two further multirotor cases and employed for two large-scale ship–helicopter dynamic interfaces with up to three concurrent rotorcraft and five rotors. Simulations of the Sikorsky X2 rotor with significant rotor–rotor interaction match resolved unsteady rotor loading well using 1.2% of the computational effort of resolved blade calculations. Computations of the outwash from a Boeing CH-47D in hover gave results within experimental error bars for most azimuths, where discrepancies may be attributed to the estimated trim and division of rotor loading. Simulations of the dynamic interface between the Simple Frigate Shape 1 and a hovering UH-60a were presented for three wind-over-deck angles and three positions for 40 knots of wind speed. The computed thrust and root mean square thrust coefficients in the closed-loop pilot control band are presented and demonstrated to have an to have an expected dependence on immersion in the ship airwake. Finally, a large-scale dynamic interface computation was performed with two CH-47Ds and one UH-60a operating concurrently over the deck of the landing helicopter assault ship at 40 knots of wind speed and 0° and 30° angle. Results highlighted the importance of understanding rotor-wake/airwake interactions dependent on rotorcraft type and position and wind-over-deck. The computation provided unsteady thrust and torque for each of the five rotors from 60 s of data using just 11,424 central processing unit hours on a grid of 19.5×106 points. These simulations show that hybrid Navier–Stokes simulations employing actuator line or surface methods are capable of providing high-fidelity, time-accurate predictions of rotor loads at the ship–helicopter dynamic interface at substantially lower cost.

Exploring the potential and challenges of phase change materials in future heat storage systems via computations and experiments

The current work describes computational and experimental efforts to enable new heat storage technologies. Approximately 50% of the energy consumed globally is used to generate and manage heat. This is still predominantly acquired from the combustion of fossil fuels, consequently generating just over 40% of the total global CO2 emissions. To achieve a sustainable future, there is a need for sustainable, decarbonised and dependable energy sources. Heat storage is a key technology that could help achieving this goal. Recently published work by the current authors has shown the importance of heat storage in decarbonising heating (and cooling) of buildings using phase change materials (PCMs). The black-box approach followed evaluated the energetic, cost, and environmental impacts of a novel concept PCM storage scheme with microgeneration. The results obtained showed that this system had advantages over current and future competing heating technologies. However, PCM technology remains at low technological readiness levels because of the lack of understanding of the complex heat transfer physics associated with their phase change transition. For example, understanding and characterizing the propagation of the solid-liquid front during charging and discharging in such heat storage modules will be important in predicting and optimising their performance. In the current work, computations have been performed to understand and optimise the effects of container geometry and size on the charging and discharging heat flux profiles of a simple heat storage module. Furthermore, fundamental experiments have also been implemented to identify the complex governing mechanisms of heat transfer and storage in such materials during transient operation.

Artificial intelligence-based surrogate modeling for computational cost-effective optimization of hydrogen liquefaction process

The chemical processes are inherently complex, and obtaining desired product quality and high energy efficiency by optimizing the design variables is challenging. The progress in computing technologies can be a blessing since machine learning models can process the information and data for learning purposes and eventually help ease the simulation and optimization efforts. Against this backdrop, an artificial neural network-based surrogate model is developed in the present study and applied to the challenging process of hydrogen liquefaction. Optimization of hydrogen liquefaction processes is still an open challenge because of convergence issues and many design variables; only a handful of algorithms are considered successful. One of those algorithms, particle swarm optimization (PSO), is used here to compare computational effort and results accuracy. The percentage error of prediction of the ANN-based surrogate model compared to PSO was 4% for the minimum internal approach temperature, which is usually considered as a constraint, and 0.04% for specific energy consumption considered as the objective function. The optimization of the proposed model significantly reduced the time needed for PSO optimization of the existing model by more than 99.99%. As an extension of the current work, the ‘shallow' neural network model, with one hidden layer used in the present study, can be used to analyze the whole liquefaction process.

Enhancing Multiple Sequence Alignment with Genetic Algorithms: A Bioinformatics Approach in Biomedical Engineering

This study aimed to create a genetic information processing technique for the problem of multiple alignment of genetic sequences in bioinformatics. The objective was to take advantage of the computer hardware's capabilities and analyze the results obtained regarding quality, processing time, and the number of evaluated functions. The methodology was based on developing a genetic algorithm in Java, which resulted in four different versions: Gp1, Gp2, Gp3 and Gp4 . A set of genetic sequences were processed, and the results were evaluated by analyzing numerical behavior profiles. The research found that algorithms that maintained diversity in the population produced better quality solutions, and parallel processing reduced processing time. It was observed that the time required to perform the process decreased, according to the generated performance profile. The study concluded that conventional computer equipment can produce excellent results when processing genetic information if algorithms are optimized to exploit hardware resources. The computational effort of the hardware used is directly related to the number of evaluated functions. Additionally, the comparison method based on the determination of the performance profile is highlighted as a strategy for comparing the algorithm results in different metrics of interest, which can guide the development of more efficient genetic information processing techniques.

Numerical-experimental study on the crashworthiness of a windshield A-pillar

This document reports a numerical and experimental crashworthiness study of a vehicle structural component. Particularly, a windshield A-pillar made of a novel high-strength steel was tested on an apparatus for impact tests that was in-house designed so as to reproduce the effects of a vehicle side impact crash. In parallel, numerical simulations were performed using the explicit finite element code LS-DYNA, considering different elastic-plastic material constitutive laws with damage. A comparison between numerical and experimental outcomes demonstrated a satisfactory correlation, with the GISSMO model that provided a good correlation and needed a relatively low computational effort.

A study on the applicability of simplified few-group GET (Generalized Equivalence Theory) to cylindrical molten salt fast reactor

One of the essential requirements for molten salt reactor (MSR) design is methodology for analyzing multi-physics phenomena, such as the behavior of liquid fuel. In the research of Molten Salt Fast Reactors (MSFRs), the Neutron Diffusion Equation (NDE) is widely employed. This study introduces a method to enhance the accuracy of neutronic analysis of MSFR using the NDE. Using the simple structure of MSFR and the characteristics of liquid nuclear fuel, it is intended to enable the application of the simple equivalence method, which is difficult to perform on the existing fast reactor. A straightforward yet effective approach named the simplified Generalized Equivalence Theory (simplified GET) is proposed for applying flux-volume-weighted homogenized cross-sections and representative Discontinuity Factor (DF) values obtained at material. This approach, while similar to the General Equivalence Theory (GET) method, significantly simplifies the enhancement of accuracy in reactor analysis, minimizing computational efforts. Our investigation spans from the initial core to the burned core, ensuring the applicability of this simple method throughout the reactor's operational life. The proposed method demonstrates promising results, offering a viable solution to improve the accuracy of NDE-based calculations in MSFRs.