scholarly journals Systematic Investigation of a Large Two-Stroke Engine Crankshaft Dynamics Model

Energies ◽  
2020 ◽  
Vol 13 (10) ◽  
pp. 2486 ◽  
Author(s):  
Konstantinos-Marios Tsitsilonis ◽  
Gerasimos Theotokatos ◽  
Nikolaos Xiros ◽  
Malcolm Habens
Keyword(s):  
Numerical Scheme ◽  
Linear Time ◽  
Piecewise Linear ◽  
Model Performance ◽  
Fast Response ◽  
Systematic Investigation ◽  
Computational Time ◽  
Dynamics Model ◽  
Engine Friction ◽  
Engine Crankshaft

The crankshaft dynamics model is of vital importance to a multitude of aspects on engine diagnostics; however, systematic investigations of such models performance (especially for large two-stroke diesel engines that are widely used in the power generation and shipping industries) have not been reported in the literature. This study aims to cover this gap by systematically investigating the parameters that affect the performance of a two-stroke diesel engine crankshaft dynamics model, such as the numerical scheme as well as the engine components inertia and friction. Specifically, the following alternatives are analysed: (a) two optimal performing numerical schemes, in particular, a stiff ordinary differential equations (ODE) solver and a fast solver based on a piecewise Linear Time-Invariant (LTI) scheme method, (b) the linear and the non-linear inertia-speed approaches, and (c) three engine friction submodels of varying complexity. All the potential combinations of the alternatives are investigated, and the crankshaft dynamics model performance is evaluated by employing Key Performance Indicators (KPIs), which consider the results accuracy compared to the measured data, the computational time, and the energy balance error. The results demonstrate that the best performing combination includes the stiff ODE solver, the constant inertia-speed approach and the most simplistic engine friction submodel. However, the LTI numerical scheme is recommended for applications that require fast response due to the significant savings in computational time with an acceptable compromise in the model results accuracy.

scholarly journals Performance Evaluation of Deep CNN-Based Crack Detection and Localization Techniques for Concrete Structures

Sensors ◽  
2021 ◽  
Vol 21 (5) ◽  
pp. 1688
Author(s):  
Luqman Ali ◽  
Fady Alnajjar ◽  
Hamad Al Jassmi ◽  
Munkhjargal Gochoo ◽  
Wasif Khan ◽  
...  
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Crack Detection ◽  
Concrete Structures ◽  
Model Performance ◽  
Training Data ◽  
Computational Time ◽  
Data Heterogeneity ◽  
Public Datasets ◽  
Detection And Localization

This paper proposes a customized convolutional neural network for crack detection in concrete structures. The proposed method is compared to four existing deep learning methods based on training data size, data heterogeneity, network complexity, and the number of epochs. The performance of the proposed convolutional neural network (CNN) model is evaluated and compared to pretrained networks, i.e., the VGG-16, VGG-19, ResNet-50, and Inception V3 models, on eight datasets of different sizes, created from two public datasets. For each model, the evaluation considered computational time, crack localization results, and classification measures, e.g., accuracy, precision, recall, and F1-score. Experimental results demonstrated that training data size and heterogeneity among data samples significantly affect model performance. All models demonstrated promising performance on a limited number of diverse training data; however, increasing the training data size and reducing diversity reduced generalization performance, and led to overfitting. The proposed customized CNN and VGG-16 models outperformed the other methods in terms of classification, localization, and computational time on a small amount of data, and the results indicate that these two models demonstrate superior crack detection and localization for concrete structures.

Three-Dimensional Paddle Shift Modeling for IC Packaging

2004 ◽  
Vol 127 (3) ◽  
pp. 324-334 ◽  
Author(s):  
Chien-Chang Pei ◽  
Sheng-Jye Hwang
Keyword(s):  
Structural Analysis ◽  
Integrated Circuits ◽  
Three Dimensional ◽  
Void Formation ◽  
Model Performance ◽  
Computational Time ◽  
Pressure Loading ◽  
Filling Process ◽  
Molding Process ◽  
Incomplete Filling

The plastic packaging process for integrated circuits is subject to several fabrication defects. For packages containing leadframes, three major defects may occur in the molding process alone, namely, incomplete filling and void formation, wire sweep, and paddle shift. Paddle shift is the deflection of the leadframe pad and die. Excessive paddle shift reduces the encapsulation protection for the components and may result in failures due to excessive wire sweep. Computer-aided analysis is one of the tools that could be used to simulate and predict the occurrence of such molding-process-induced defects, even prior to the commencement of mass production of a component. This paper presents a methodology for computational modeling and prediction of paddle shift during the molding process. The methodology is based on modeling the flow of the polymer melt around the leadframe and paddle during the filling process, and extracting the pressure loading induced by the flow on the paddle. The pressure loading at different times during the filling process is then supplied to a three-dimensional, static, structural analysis module to determine the corresponding paddle deflections at those times. The paper outlines the procedures used to define the relevant geometries and to generate the meshes in the “fluid” and “structural” subdomains, and to ensure the compatibility of these meshes for the transfer of pressure loadings. Results are shown for a full paddle shift simulation. The effect on the overall model performance of different element types for the mold-filling analysis and the structural analysis is also investigated and discussed. In order to obtain more accurate results and in a shorter computational time for the combined (fluid and structural) paddle shift analysis, it was found that higher-order elements, such as hexahedra or prisms, are more suitable than tetrahedra.

On Improving Algorithm Efficiency of Gas-Kick Simulations toward Automated Influx Management: A Robertson Differential-Algebraic-Equation Problem Approach

2021 ◽  
pp. 1-24
Author(s):  
Chen Wei ◽  
Yuanhang Chen
Keyword(s):  
Decision Making ◽  
Real Time ◽  
Model Prediction ◽  
Computational Efficiency ◽  
Prediction Accuracy ◽  
Numerical Scheme ◽  
Computational Time ◽  
Numerical Efficiency ◽  
Flow Models ◽  
Gas Kick

Summary Improved numerical efficiency in simulating wellbore gas-influx behaviors is essential for realizing real-time model-prediction-based gas-influx management in wells equipped with managed-pressure-drilling (MPD) systems. Currently, most solution algorithms for high-fidelitymultiphase-flow models are highly time consuming and are not suitable for real-time decision making and control. In the application of model-predictive controllers (MPCs), long calculation time can lead to large overshoots and low control efficiency. This paper presents a drift-flux-model (DFM)-based gas-influx simulator with a novel numerical scheme for improved computational efficiency. The solution algorithm to a Robertson problem as differential algebraic equations (DAEs) was used as the numerical scheme to solve the control equations of the DFM in this study. The numerical stability and computational efficiency of this numerical scheme and the widely used flux-splitting methods are compared and analyzed. Results show that the Robertson DAE problem approach significantly reduces the total number of arithmetic operations and the computational time compared with the hybrid advection-upstream-splitting method (AUSMV) while maintaining the same prediction accuracy. According to the “Big-O notation” analysis, the Robertson DAE approach shows a lower-order growth of computational complexity, proving its good potential in enhancing numerical efficiency, especially when handling simulations with larger scales. The validation of both the numerical schemes for the solution of the DFM was performed using measured data from a test well drilled with water-based mud (WBM). This study offers a novel numerical solution to the DFM that can significantly reduce the computational time required for gas-kick simulation while maintaining high prediction accuracy. This approach enables the application of high-fidelity two-phase-flow models in model-prediction-based decision making and automated influx management with MPD systems.

A Piecewise Linear Time-Varying Model for Modeling the Charging and Discharging Processes of a Lithium-Ion Battery

2014 ◽  
Vol 5 (2) ◽  
pp. 87-103
Author(s):  
Arab AlSharif ◽  
Manohar Das
Keyword(s):  
Test Data ◽  
Linear Time ◽  
Piecewise Linear ◽  
Lithium Ion ◽  
Battery Management System ◽  
Time Varying ◽  
Battery Management ◽  
Simulated Test ◽  
Li Ion ◽  
Charging And Discharging Processes

A piecewise linear, time-varying model for modeling the charging and discharging processes of Li-ion batteries is introduced in this paper. Such a model consists of a group of piecewise linear model segments, whose parameters are adapted online over time. Thus, the combined overall model is capable of modeling nonlinear time-varying processes, such as a Li-ion battery charging and discharging processes, quite well. Modeling results of both simulated test data and actual test data gathered from a high-power automotive-grade Li-ion cell are presented. The close matches between actual and model-predicted behaviors demonstrate the effectiveness of the proposed modeling approach and indicate the potential usefulness of such models for a battery management system.

scholarly journals Thermal modeling of hydrogen storage by absorption in a magnesium hydrides tank

2014 ◽  
Vol 5 ◽  
pp. A21 ◽  
Author(s):  
Karim Lahmer ◽  
Rachid Bessaïh ◽  
Angel Scipioni ◽  
Mohammed El Ganaoui
Keyword(s):  
Hydrogen Absorption ◽  
Numerical Scheme ◽  
Computational Time ◽  
Governing Equations ◽  
Time Step ◽  
Kinetic Reaction ◽  
Fully Implicit ◽  
Temporal Profiles ◽  
Tank Geometry ◽  
Reaction Equations

This paper summarizes numerical results of hydrogen absorption simulated in an axisymmetric tank geometry containing magnesium hydride heated to 300 °C and at moderate storage pressure 1 MPa. The governing equations are solved with a fully implicit finite volume numerical scheme used by a commercial software FLUENT. The effect of the different kinetic reaction equations modeling hydrogen absorption was studied by the introduction of a specific subroutine at each time step in order to consider which one will provide results close to available experimental results. Spatial and temporal profiles of temperature and concentration in hydride bed are plotted. Results show that suitable method for our two-dimensional study is a CV-2D technique because it generates the smallest error especially during the beginning of the reaction. Also, its computational time is the shortest one compared to the other methods.

scholarly journals Hybrid Harmony Search–Artificial Intelligence Models in Credit Scoring

Entropy ◽  
2020 ◽  
Vol 22 (9) ◽  
pp. 989
Author(s):  
Rui Ying Goh ◽  
Lai Soon Lee ◽  
Hsin-Vonn Seow ◽  
Kathiresan Gopal
Keyword(s):  
Artificial Intelligence ◽  
Feature Selection ◽  
Credit Scoring ◽  
Harmony Search ◽  
Model Performance ◽  
Hybrid Models ◽  
Performance Model ◽  
Computational Time ◽  
Support Vector ◽  
Artificial Intelligence Models

Credit scoring is an important tool used by financial institutions to correctly identify defaulters and non-defaulters. Support Vector Machines (SVM) and Random Forest (RF) are the Artificial Intelligence techniques that have been attracting interest due to their flexibility to account for various data patterns. Both are black-box models which are sensitive to hyperparameter settings. Feature selection can be performed on SVM to enable explanation with the reduced features, whereas feature importance computed by RF can be used for model explanation. The benefits of accuracy and interpretation allow for significant improvement in the area of credit risk and credit scoring. This paper proposes the use of Harmony Search (HS), to form a hybrid HS-SVM to perform feature selection and hyperparameter tuning simultaneously, and a hybrid HS-RF to tune the hyperparameters. A Modified HS (MHS) is also proposed with the main objective to achieve comparable results as the standard HS with a shorter computational time. MHS consists of four main modifications in the standard HS: (i) Elitism selection during memory consideration instead of random selection, (ii) dynamic exploration and exploitation operators in place of the original static operators, (iii) a self-adjusted bandwidth operator, and (iv) inclusion of additional termination criteria to reach faster convergence. Along with parallel computing, MHS effectively reduces the computational time of the proposed hybrid models. The proposed hybrid models are compared with standard statistical models across three different datasets commonly used in credit scoring studies. The computational results show that MHS-RF is most robust in terms of model performance, model explainability and computational time.

An Investigation Into Nonlinear Growth Rate of Two-Dimensional and Three-Dimensional Single-Mode Richtmyer–Meshkov Instability Using an Arbitrary-Lagrangian–Eulerian Algorithm

2014 ◽  
Vol 136 (9) ◽  
Author(s):  
Mike Probyn ◽  
Ben Thornber ◽  
Dimitris Drikakis ◽  
David Youngs ◽  
Robin Williams
Keyword(s):  
Growth Rate ◽  
Numerical Scheme ◽  
Initial Conditions ◽  
Numerical Study ◽  
Three Dimensional ◽  
Single Mode ◽  
Computational Time ◽  
Initial Growth ◽  
Arbitrary Lagrangian Eulerian ◽  
Layer Width

This paper presents an investigation into the use of a moving mesh algorithm for solving unsteady turbulent mixing problems. The growth of a shock induced mixing zone following reshock, using an initial setup comparable to that of existing experimental work, is used to evaluate the behavior of the numerical scheme for single-mode Richtmyer–Meshkov instability (SM-RMI). Subsequently the code is used to evaluate the growth rate for a range of different initial conditions. The initial growth rate for three-dimensional (3D) SM Richtmyer–Meshkov is also presented for a number of different initial conditions. This numerical study details the development of the mixing layer width both prior to and after reshock. The numerical scheme used includes an arbitrary Lagrangian–Eulerian grid motion which is successfully used to reduce the mesh size and computational time while retaining the accuracy of the simulation results. Varying initial conditions shows that the growth rate after reshock is independent of the initial conditions for a SM provided that the initial growth remains in the linear regime.

Sign in / Sign up

Export Citation Format

Share Document