Automation of hull plates classification in ship design system using neural network method

Manufacturing the complex surface plates in stern and stem is a major factor in cost of a preliminary ship design by computing process. If these hull plate parts are effectively classified, it helps to compute the processing cost and find the way of cut-down the processing cost. This paper presents a new method to classify surface plates effectively in preliminary ship design using neural network. A neural network based ship hull plate classification programme developed and tested for the automatic classification of ship design. The input variables are regarded as Gaussian curvature distributions on the plate. Various applicable rules of network topology are applied in ship design. By observing the results of the proposed method, the effectiveness of the proposed method are discussed. As a result, high prediction rate was achieved in ship design. And the proposed method will contribute to reduce the production cost of ship.     

Optimum tire contour design using systematic stom and neural network

An efficient multi-objective optimization method is presented making use of neural network and a systematic satisficing trade-off method (STOM), in order to simultaneously improve both maneuverability and durability of tire. Objective functions are defined as follows: the sidewall-carcass tension distribution for the former performance while the belt-edge strain energy density for the latter. A back-propagation neural network model approximates the objective functions to reduce the total CPU time required for the sensitivity analysis using finite difference scheme. The satisficing trade-off process between the objective functions showing the remarkably conflicting trends each other is systematically carried out according to our aspiration-level adjustment procedure. The optimization procedure presented is illustrated through the optimum design simulation of a representative automobile tire. The assessment of its numerical merit as well as the optimization results is also presented.     

基于神经网络的交流伺服控制系统设计

针对PID控制交流伺服系统无法同时获得稳定性和快速性,提出了一种基于神经网络的交流伺服控制系统设计方法.在交流伺服控制系统中,采用改进的BP神经网络作为系统控制器,利用其极强的非线性动态跟踪能力和自适应学习能力,及对控制对象的数学模型无依赖性,实现控制对象快速、准确定位.系统仿真表明,BP网络控制器具有快速跟踪性和较好的控制精度等优点.     

Prediction of error due to eccentricity of hole in hole-drilling method using neural network

The measurement of residual stresses by the hole-drilling method has been used to evaluate residual stresses in structural members. In this method, eccentricity can usually occur between the hole center and rosette gage center. In this study, we obtained the magnitude of the error due to eccentricity of a hole through the finite element analysis. To predict the magnitude of the error due to eccentricity of a hole in the biaxial residual stress field, it could be learned through the backpropagation neural network. The prediction results of the error using the trained neural network showed good agreement with FE analyzed results.     

Determination of process parameters in stereolithography using neural network

For stereolithography process, accuracy of prototypes is related to laser power, scan speed, scan width, scan pattern, layer thickness, resin characteristics and etc. An accurate prototype is obtained by using appropriate process parameters. In order to determine these parameters, the stereolithography (SLA) machine using neural network was developed and efficiency of the developed SLA machine was compared with that of the traditional SLA. Optimum values for scan speed, hatching spacing and layer thickness improved the surface roughness and build time for the developed SLA.     

碱锰高能电池自动装环机系统最佳方案设计

本文利用现代化设计理论和方法对电池自动装环机系统方案进行了最佳设计。在设计中对总功能进行“发散”求解，获得多个可供选择的方案，经技术—经济综合评价后，确定了最佳方案。经过技术实施和生产使用表明，该原理方案完全可以满足生产使用要求，技术上达到了国际先进水平，取得了显著的经济效益和社会效益。     

Prediction of deformations of steel plate by artificial neural network in forming process with induction heating

To control a heat source easily in the forming process of steel plate with heating, the electro-magnetic induction process has been used as a substitute of the flame heating process. However, only few studies have analyzed the deformation of a workpiece in the induction heating process by using a mathematical model. This is mainly due to the difficulty of modeling the heat flux from the inductor traveling on the conductive plate during the induction process. In this study, the heat flux distribution over a steel plate during the induction process is first analyzed by a numerical method with the assumption that the process is in a quasi-stationary state around the inductor and also that the heat flux itself greatly depends on the temperature of the workpiece. With the heat flux, heat flow and thermo-mechanical analyses on the plate to obtain deformations during the heating process are then performed with a commercial FEM program for 34 combinations of heating parameters. An artificial neural network is proposed to build a simplified relationship between deformations and heating parameters that can be easily utilized to predict deformations of steel plate with a wide range of heating parameters in the heating process. After its architecture is optimized, the artificial neural network is trained with the deformations obtained from the FEM analyses as outputs and the related heating parameters as inputs. The predicted outputs from the neural network are compared with those of the experiments and the numerical results. They are in good agreement. This paper was recommended for publication in revised form by Associate Editor Youngseog Lee Truong-Thinh Nguyen received the B.S and M.S degrees from Ho Chi Minh city National University, Viet Nam, in 1997 and 2000, respectively. Now, he is a doctoral candidate in the Department of Mechanical Engineering at Chonnam National University, Korea. His research interests are Induction Heating, Thermal deformations, applications of Neural Network and Fuzzy logic in Industry, intelligent control. Young-Soo Yang received a B.S. degree in Mechanical Engineering from Sungkyunkwan University in 1985. He then went on to receive his M.S. and Ph.D. degrees from KAIST in 1987 and 1991, respectively. Dr. Yang is currently a Professor at the School of Mechanical Engineering at Chonnam National University in Gwangju, Korea. He research interests are in the area of weld structure. Kang-Yul Bae is a Professor of Mechatronics Engineering Department at Jinju National University in Jinju, Korea. He received a B.S. degree in Mechanical Engineering from Busan National University in 1984. He also holds the following degrees of M.S. in Production Engineering and Ph.D. in Mechanical Engineering from KAIST in 1986 and 1994, respectively. He has industrial experience from 1986 to 1998 at Hyundai Heavy Industries, Co. Ltd. as a senior researcher. His teaching and research areas include manufacturing processes, automation, and mechatronics. Sung-Nam Choi received a B.S. and a M.S. degree in Department of Mechanical Engineering from Chonnam National University in 1989 and 1991. He then complete doctor course from the same University in 2006. He is currently a Senior Researcher at Non-Destructive Evaluation Center at Korea Electric Power Research Institute in Daejon, Korea. His research interests are in the area of weld integrity, fracture mechanics, and automated ultrasonic examination.     

Acoustic diagnosis of a pump by using neural network

A fundamental study for developing a fault diagnosis system of a pump is performed by using neural network. Acoustic signals were obtained and converted to frequency domain for normal products and artificially deformed products. The neural network model used in this study was 3-layer type composed of input, hidden, and output layer. The normalized amplitudes at the multiples of real driving frequency were chosen as units of input layer. And the codes of pump malfunctions were selected as units of output layer. Various sets of teach signals made from original data by eliminating some random cases were used in the training. The average errors were approximately proportional to the number of untaught data. Neural network trained by acoustic signals can detect malfunction or diagnose fault of a given machine from the results.     

Design of axial fan using inverse design method

The axial fans for cooling condensers were designed by inverse design code TURBOdesign-1. The parameters of the inverse design were set by DOE (design of experiments). By changing the design parameters, such as the distribution of the blade loading, spanwise circulation distribution and stacking, 32 different fan designs were created for the screening of parameters. The overall performance and the local flow field of these fans were computed using a commercial CFD code. The results of the CFD computations were analyzed by DOE. The pressure rise and efficiency were selected as the main responses, and the main effects of the design parameters on the responses were discussed. The main design parameters for the optimum design of the fan were decided from the results of the screening procedure. We designed the optimum axial fan by RSM (response surface method). The design center fan was made by RP (rapid prototype) and the performance was tested using a fan tester based on AMCA standards. These procedures ensured proper screening of parameters and optimum design of the axial fan. This paper was presented at the 9^th Asian International Conference on Fluid Machinery (AICFM9), Jeju, Korea, October 16–19, 2007.     

Optimum design of structures with stress and displacement constraints using the force method

A new structural analysis and optimization algorithm is developed to determine the minimum weight of structures with the truss and beam-type members under displacement and stress constraints. The algorithm combines the mathematical programming based on the sequential quadratic programming (SQP) technique and the finite element technique based on the integrated force method. The equilibrium matrix is generated automatically through the finite element analysis while the compatibility matrix is obtained directly using the displacement–deformation relations and the single value decomposition (SVD) technique. By combining the equilibrium and compatibility matrices with the force–displacement relations, the equations of equilibrium with the element forces as variables are obtained. The proposed method is extremely efficient to analyze and optimize the truss and beam structures under stress and displacement constraints. The computational effort required by the force method is found to be significantly lower than that of the displacement method. The effect of the geometric nonlinearity in the structural optimization problems under the stress and displacement constraints were also investigated and it is illustrated that the geometric nonlinearity is not an important issue in these types of problems and hence, it does not affect the final optimum solution significantly. Four examples illustrate the procedure and allow the results to be compared with those reported in the literatures.     

基于概率神经网络的CAPP系统中零件表面加工方法的研究

零件表面加工方法的选择是CAPP系统的重要组成部分,利用概率神经网络以逐步分解的方式来实现零件表面加工方法的选择,并给出了网络的构造、训练和组织训练数据的方法。结果表明,利用人工神经网络可以很好地选择零件表面的加工方法,克服了CAPP专家系统存在的工艺知识获取“瓶颈”问题和“无穷递归”、“推理薄弱”等推理过程中出现的问题。     

Improvement of the control performance of pneumatic artificial muscle manipulators using an intelligent switching control method

Problems with the control, oscillatory motion and compliance of pneumatic systems have prevented their widespread use in advanced robotics. However, their compactness, power/weight ratio, ease of maintenance and inherent safety are factors that could be potentially exploited in sophisticated dexterous manipulator designs. These advantages have led to the development of novel actuators such as the McKibben Muscle, Rubber Actuator and Pneumatic Artificial Muscle Manipulators. However, some limitations still exist, such as a deterioration of the performance of transient response due to the changes in the external inertia load in the pneumatic artificial muscle manipulator. To overcome this problem, a switching algorithm of the control parameter using a learning vector quantization neural network (LVQNN) is newly proposed. This estimates the external inertia load of the pneumatic artificial muscle manipulator. The effectiveness of the proposed control algorithm is demonstrated through experiments with different external inertia loads.     

A study on the fault diagnosis of roller-shape using frequency analysis of tension signals and artificial neural networks based approach in a web transport system

Rollers in the continuous process systems are ones of key components that determine the quality of web products. The condition of rollers (e.g. eccentricity, runout) should be consistently monitored in order to maintain the process conditions (e.g. tension, edge position) within a required specification. In this paper, a new diagnosis algorithm is suggested to detect the defective rollers based on the frequency analysis of web tension signals. The kernel of this technique is to use the characteristic features (RMS, Peak value, Power spectral density) of tension signals which allow the identification of the faulty rollers and the diagnosis of the degree of fault in the rollers. The characteristic features could be used to train an artificial neural network which could classify roller conditions into three groups (normal, warning, and faulty conditions). The simulation and experimental results showed that the suggested diagnosis algorithm can be successfully used to identify the defective rollers as well as to diagnose the degree of the defect of those rollers.     

人工神经网络在机械设计中的应用

本文通过对机械设计专家系统和人工神经网络的讨论，研究了人工神经网络和专家系统技术在机械设计智能系统中的综合应用问题，并提出了人工神经网络在机械设计中的总体应用方案，为进一步研究打下了基础。     

小波神经网络及其在机械工程领域中的应用

介绍小波神经网络的结构形式及其在机械工程领域的主要应用,并指出今后的发展趋势和需要深入探索的课题.     

工程系统设计指标的最优分配方法

概述了工程系统设计指标最优分配的三类方法,直接法,分解协调法和协同分配法。其中,协同分配法已经初步证明是一种有效的新方法。并简要的分析了它们的优缺点及适用范围。     

基于神经网络的机床-工件系统热误差补偿技术研究

以机床-工件系统的热变形为研究对象,应用神经网络理论建立机床-工件系统的热误差模型,对热误差神经网络模型的关键输入参数进行了分析讨论,提出了该模型的误差补偿策略。以某型号大尺寸回转支承滚道数控车削加工为例,建立了热误差模型,对回转支承滚道加工实施热误差补偿,结果表明,机床-工件系统的热误差模型有较强的预测能力,提出的补偿方法有较好的补偿效果。     

Defect diagnostics of SUAV gas turbine engine using hybrid SVM-artificial neural network method

A hybrid method of an artificial neural network (ANN) combined with a support vector machine (SVM) has been developed for the defect diagnostic system applied to the SUAV gas turbine engine. This method has been suggested to overcome the demerits of the general ANN with the local minima problem and low classification accuracy in case of many nonlinear data. This hybrid approach takes advantage of the reduction of learning data and converging time without any loss of estimation accuracy because the SVM classifies the defect location and reduces the learning data range. The results of test data have shown that the hybrid method is more reliable and suitable algorithm than the general ANN for the defect diagnosis of the gas turbine engine. This paper was recommended for publication in revised form by Associate Editor Tong Seop Kim Tae-Seong Roh received his B.S. and M.S. degrees in Aeronautical Engineering from Seoul National University in 1984 and 1986. He then went on to receive his Ph.D. degree from Pennsylvania State University in 1995. Dr. Roh is currently a Professor at the department of Aerospace Engineering at Inha University in Incheon, Korea. His research interests are in the area of combustion instabilities, rocket and jet propulsion, interior ballistics, and gas turbine engine defect diagnostics. Dong-Whan Choi received his B.S. degree in Aeronautical Engineering from Seoul National University in 1974. He then went on to receive his M.S. and Ph.D. degrees from University of Washington in 1978 and 1983. Dr. Choi served three years as a President of the Korea Aerospace Research Institute from 1999. He is currently a professor at the department of Aerospace Engineering at Inha University in Incheon, Korea. His research interests are in the area of turbulence, jet propulsion, and gas turbine defect diagnostics.     

Pareto genetic design of group method of data handling type neural network for prediction discharge coefficient in rectangular side orifices

The powerful method of Group Method of Data Handling (GMDH) was used for estimating the discharge coefficient of a rectangular side orifice. First, the existing equations for calculating the discharge coefficient were studied making use of experimental results. On the first hand, the factors affecting the discharge coefficient were determined, then five models were constructed in order to analyze the sensitivity in achieving accuracy by using different parameters. The results, obtained using statistical indexes (MARE=0.021 and RMSE=0.017), showed that one model out of the five models, on estimation using the dimensionless parameters of the ratio of depth of flow in main channel to width of rectangular orifice (Y_m/L), Froude number (Fr), the ratio of sill height to width of rectangular orifice (W/L) and width of main channel to width of rectangular orifice (B/L), presented the best results.     

基于遗传神经网络的压铸机座板结构优化设计

在对压铸机合模机构进行结构设计时,利用神经网络的非线性映射能力,通过少量样本的有限元分析结果,训练出表述结构参数间函数关系的神经网络模型,然后利用遗传算法的全局寻优性找到神经网络模型表述的目标函数的最优结构参数,从而解决结构优化设计的瓶颈和智能问题,利用这种优化设计策略,设计了压铸机合模机构座板,结果表明了该方法的高效性。