• Nuclear Engineering and Technology
• /
• 제55권6호
• /
• pp.2276-2282
• /
• 2023

We deploy artificial neural networks to unfold neutron spectra from measured energy-integrated quantities. These neutron spectra represent an important parameter allowing to compute the absorbed dose and the kerma to serve radiation protection in addition to nuclear safety. The built architectures are inspired from convolutional neural networks. The first architecture is made up of residual transposed convolution's blocks while the second is a modified version of the U-net architecture. A large and balanced dataset is simulated following "realistic" physical constraints to train the architectures in an efficient way. Results show a high accuracy prediction of neutron spectra ranging from thermal up to fast spectrum. The dataset processing, the attention paid to performances' metrics and the hyper-optimization are behind the architectures' robustness.

• Journal of Construction Engineering and Project Management
• /
• 제5권1호
• /
• pp.11-19
• /
• 2015

This paper presents optimized artificial neural networks (ANNs) claims prediction and decision awareness framework that guides owner organizations in their pre-bid construction project decisions to minimize claims. The framework is composed of two genetic optimization ANNs models: a Claims Impact Prediction Model (CIPM), and a Decision Awareness Model (DAM). The CIPM is composed of three separate ANNs that predict the cost and time impacts of the possible claims that may arise in a project. The models also predict the expected types of relationship between the owner and the contractor based on their behavioral and technical decisions during the bidding phase of the project. The framework is implemented using actual data from international projects in the Middle East and Egypt (projects owned by either public or private local organizations who hired international prime contractors to deliver the projects). Literature review, interviews with pertinent experts in the Middle East, and lessons learned from several international construction projects in Egypt determined the input decision variables of the CIPM. The ANNs training, which has been implemented in a spreadsheet environment, was optimized using genetic algorithm (GA). Different weights were assigned as variables to the different layers of each ANN and the total square error was used as the objective function to be minimized. Data was collected from thirty-two international construction projects in order to train and test the ANNs of the CIPM, which predicted cost overruns, schedule delays, and relationships between contracting parties. A genetic optimization backward analysis technique was then applied to develop the Decision Awareness Model (DAM). The DAM combined the three artificial neural networks of the CIPM to assist project owners in setting optimum values for their behavioral and technical decision variables. It implements an intelligent user-friendly input interface which helps project owners in visualizing the impact of their decisions on the project's total cost, original duration, and expected owner-contractor relationship. The framework presents a unique and transparent hybrid genetic algorithm-ANNs training and testing method. It has been implemented in a spreadsheet environment using MS Excel$^{(R)}$ and EVOLVERTM V.5.5. It provides projects' owners of a decision-support tool that raises their awareness regarding their pre-bid decisions for a construction project.

• Journal of Mechanical Science and Technology
• /
• 제20권2호
• /
• pp.251-261
• /
• 2006

The objective of this work is to apply artificial neural networks for solving inverse problems in the structural optimization of a fiber optic pressure sensor. For the sensor under investigation to achieve a desired accuracy, the change in the distance between the tips of the two fibers due to the applied pressure should not interfere with the phase change due to the change in the density of the air between the two fibers. Therefore, accurate dynamic analysis and structural optimization of the sensor is essential to ensure the accuracy of the measurements provided by the sensor. To this end, a normal mode analysis and a transient response analysis of the sensor were performed by combining commercial finite element analysis package, MSC/NASTRAN, and MATLAB. Furthermore, a parametric study on the design of the sensor was performed to minimize the size of the sensor while fulfilling a number of constraints. In performing the parametric study, the need for a relationship between the design parameters and the response of the sensor was fulfilled by using a neural network. The whole process of the dynamic analysis using commercial finite element analysis package and the parameter optimization of the sensor were automated within the MATLAB environment.

• 조명전기설비학회논문지
• /
• 제21권5호
• /
• pp.1-9
• /
• 2007

본 논문은 SynRM의 동손 및 철손을 최소화 하는 효율 최적화 제어를 제시한다. 퍼지와 신경회로망으로 구성된 적응 퍼지-신경회로망 제어기를 바탕으로 한 속도 제어기를 설계한다. 특정 전동기 토크를 발생하는 d-q축 전류 조합은 무수히 많이 존재한다. 효율 최적화 제어기의 목적은 정상상태에 확실한 동작점에서 d-q축 전류 조합을 찾는 것이다. 제시된 알고리즘은 양호한 동적 토크 제어를 유지하는 동안 속도 및 토크 변화를 감소시키기 위하여 전자기적 손실은 허용한다. HAI 제어기의 제어 성능은 다양한 동작 상태에서 평가된다. 결과 분석은 제시된 알고리즘의 타당성을 보여준다.

• 한국멀티미디어학회논문지
• /
• 제20권8호
• /
• pp.1406-1420
• /
• 2017

Artificial neural networks (ANNs) play an important role in the fields of function approximation, prediction, and classification. ANN performance is critically dependent on the input parameters, including the number of neurons in each layer, and the optimal values of weights and biases assigned to each neuron. In this study, we apply the particle swarm optimization method, a popular optimization algorithm for determining the optimal values of weights and biases for every neuron in different layers of the ANN. Several regression models, including general linear regression, Fourier regression, smoothing spline, and polynomial regression, are conducted to evaluate the proposed method's prediction power compared to multiple linear regression (MLR) methods. In addition, residual analysis is conducted to evaluate the optimized ANN accuracy for both training and test datasets. The experimental results demonstrate that the proposed method can effectively determine optimal values for neuron weights and biases, and high accuracy results are obtained for prediction applications. Evaluations of the proposed method reveal that it can be used for prediction and estimation purposes, with a high accuracy ratio, and the designed model provides a reliable technique for optimization. The simulation results show that the optimized ANN exhibits superior performance to MLR for prediction purposes.

• 대한기계학회논문집A
• /
• 제34권7호
• /
• pp.813-820
• /
• 2010

포틀랜드 시멘트용 분쇄기는 독일과 일본 등 선진국에서 도입된 고가의 대형 기계이다. 따라서 이에 대한 체계적 정비 및 보수가 원활히 진행되어야 포틀랜드 시멘트의 생산설비에 대한 안정성을 확보할 수 있다. 한편 국내에 도입된 수직형 롤러 분쇄기는 포틀랜드 시멘트의 원료인 석회석의 시간당 생산량이 5.5MN이나 되는 세계 최대 규모의 분쇄기로서 설계 수명이 $4{\times}10^{7%}$사이클 정도이나 대략 $4{\times}10^6\;{\sim}\;8{\times}10^6$ 사이클 정도에서 파괴되고 있어 계획 예방 정비에 대한 어려움이 있으며, 수직형 롤러 분쇄기의 보수비용을 절감하기 위하여 롤러 분쇄기에 대한 효과적인 재설계가 필요한 실정이다. 따라서 본 연구에서는 확률론적인 절차가 내재되어 있어 불확실성을 다룰 수 있고, 대량의 복잡한 비선형적인 관계도 단순화의 과정 없이 연관 관계를 자체 조직화할 수 있는 인간의 뇌와 가장 유사한 병렬연산모델인 신경회로망을 수직형 롤러 분쇄기에 적용하여 최적설계를 수행하였다.

• Structural Engineering and Mechanics
• /
• 제75권5호
• /
• pp.633-642
• /
• 2020

Artificial neural networks (ANNs) are known as intelligent methods for modeling the behavior of physical phenomena because of it is a soft computing technique and takes data samples rather than entire data sets to arrive at solutions, which saves both time and money. ANN is successfully used in the civil engineering applications which are suitable examining the complicated relations between variables. Functionally graded materials (FGMs) are advanced composites that successfully used in various engineering design. The FGMs are nonhomogeneous materials and made of two different type of materials. In the present study, the bending analysis of functionally graded material (FGM) beams presents on theoretical based on combination of mixed-finite element method, Gâteaux differential and Timoshenko beam theory. The main idea in this study is to build a model using ANN with four parameters that are: Young's modulus ratio (E_t/E_b), a shear correction factor (k_s), power-law exponent (n) and length to thickness ratio (L/h). The output data is the maximum displacement (w). In the experiments: 252 different data are used. The proposed ANN model is evaluated by the correlation of the coefficient (R), MAE and MSE statistical methods. The ANN model is very good and the maximum displacement can be predicted in ANN without attempting any experiments.

• 전자공학회논문지S
• /
• 제35S권4호
• /
• pp.67-75
• /
• 1998

In general, Evoluationary Algorithm(EAs) are refered to as methods of population-based optimization. And EAs are considered as very efficient methods of optimal sytem design because they can provice much opportunity for obtaining the global optimal solution. This paper presents a co-evolution scheme of artifical neural networks, which has two different, still cooperatively working, populations, called as a host popuation and a parasite population, respectively. Using the conventional generatic algorithm the host population is evolved in the given environment, and the parastie population composed of schemata is evolved to find useful schema for the host population. the structure of artificial neural network is a diagonal recurrent neural netork which has self-feedback loops only in its hidden nodes. To find optimal neural networks we should take into account the structure of the neural network as well as the adaptive parameters, weight of neurons. So we use the genetic algorithm that searches the structure of the neural network by the co-evolution mechanism, and for the weights learning we adopted the evolutionary stategies. As a results of co-evolution we will find the optimal structure of the neural network in a short time with a small population. The validity and effectiveness of the proposed method are inspected by applying it to the stabilization and position control of the invered-pendulum system. And we will show that the result of co-evolution is better than that of the conventioal genetic algorithm.

• Bulletin of the Korean Chemical Society
• /
• 제28권9호
• /
• pp.1472-1476
• /
• 2007

Artificial neural networks (ANNs) are successfully developed for the modeling and prediction of normalized polarity parameter (ETN) of 216 various solvents with diverse chemical structures using a quantitative-structure property relationship. ANN with architecture 5-9-1 is generated using five molecular descriptors appearing in the multi-parameter linear regression (MLR) model. The most positive charge of a hydrogen atom (q+), total charge in molecule (qt), molecular volume of solvent (Vm), dipole moment (μ) and polarizability term (πI) are input descriptors and its output is ETN. It is found that properly selected and trained neural network with 192 solvents could fairly represent the dependence of normalized polarity parameter on molecular descriptors. For evaluation of the predictive power of the generated ANN, an optimized network is applied for prediction of the ETN values of 24 solvents in the prediction set, which are not used in the optimization procedure. Correlation coefficient (R) and root mean square error (RMSE) of 0.903 and 0.0887 for prediction set by MLR model should be compared with the values of 0.985 and 0.0375 by ANN model. These improvements are due to the fact that the ETN of solvents shows non-linear correlations with the molecular descriptors.

• Steel and Composite Structures
• /
• 제47권5호
• /
• pp.601-613
• /
• 2023

Numerous methods have been proposed in predicting formability of sheet metals based on microstructural and macro-scale properties of sheets. However, there are limited number of papers on the optimization problem to increase formability of sheet metals. In the present study, we aim to use novel optimization algorithms in neural networks to maximize the formability of sheet metals based on tensile curve and texture of aluminum sheet metals. In this regard, experimental and numerical evaluations of effects of texture and tensile properties are conducted. The texture effects evaluation is performed using Taylor homogenization method. The data obtained from these evaluations are gathered and utilized to train and validate an artificial neural network (ANN) with different optimization methods. Several optimization method including grey wolf algorithm (GWA), chimp optimization algorithm (ChOA) and whale optimization algorithm (WOA) are engaged in the optimization problems. The results demonstrated that in aluminum alloys the most preferable texture is cube texture for the most formable sheets. On the other hand, slight differences in the tensile behavior of the aluminum sheets in other similar conditions impose no significant decreases in the forming limit diagram under stretch loading conditions.