Neural-network process modeling of a continuous manufacturing operation

Neural-network techniques for the development of models of critical parameters in continuous forest products manufacturing processes are described. Predictive models of strength parameters in particleboard manufacturing were developed utilizing both backpropagation and counterpropagation neural network techniques. The modeled strength parameters were modulus of rupture and internal bond. The backpropagation neural network model did not provide sufficient accuracy in predicting the values of the strength parameters. Counterpropagation was successful at predicting modulus of rupture within ± 10% and internal bond within ± 15%. The trained counterpropagation network can be used to improve process control and reduce the amount of substandard and scrap board produced. Efforts are underway to refine the counterpropagation network and further improve its predictive capability, as well as to evaluate alternative neural network paradigms.     

Nonlinear interval model control of quasi-keyhole arc welding process

This paper addresses the development of a nonlinear model based interval model control system for the quasi-keyhole arc welding process, a novel arc welding process which has advantages over the laser welding process and conventional arc welding processes. The structure of the nonlinear model chosen was proposed based on an analysis of the quasi-keyhole process to be controlled. Because of the variations in the manufacturing conditions, the parameters of the nonlinear model are uncertain but bounded by fixed intervals if the range of the manufacturing conditions is specified. To determine the intervals, extreme operating conditions/parameters (manufacturing conditions) were used to conduct experiments. Each experiment gives a set of model parameters and the interval for each parameter is given by the minimum and maximum among the values obtained from different experiments. Closed-loop control experiments have verified the effectiveness of the developed system as a robust control which requires no re-adjustment and can function properly when fluctuations/variations in manufacturing conditions, and thus the process dynamics, change, vary, or fluctuate.     

Neuro evolutionary model for weld residual stress prediction

Welding is an efficient reliable metal joining process in which the coalescence of metals is achieved by fusion. Localized heating during welding, followed by rapid cooling, induce residual stresses in the weld and in the base metal. Determination of magnitude and distribution of welding residual stresses is essential and important. Data sets from finite element method (FEM) model are used to train the developed neural network model trained with genetic algorithm and particle swarm optimization (NN–GA–PSO model). The performance of the developed NN–GA–PSO model is compared neural network model trained with genetic algorithm (NN–GA) and neural network model trained with particle swarm optimization (NN–PSO) model. Among the developed models, performance of NN–GA–PSO model is superior in terms of computational speed and accuracy. Confirmatory experiments are performed using X-ray diffraction method to confirm the accuracy of the developed models. These developed models can be used to set the initial weld process parameters in shop floor welding environment.     

COMBINING A RADIAL BASIS NEURAL NETWORK WITH TIME SERIES ANALYSIS TECHNIQUES TO PREDICT MANUFACTURING PROCESS PARAMETERS

The accurate prediction of the values of critical quality parameters of a product during the production stage is a key factor in the success of a manufacturing operation. Neural network algorithms have been used to successfully predict process parameter values. However, techniques to further improve the predictive capability of neural network models are sought. Thus, an analysis was conducted to determine if the predictive capability of the network would he improved if the prediction from a time series model of a manufacturing process parameter were included in the training data set of a radial basis function neural network model. A manufacturing process data set was evaluated, and the use of the time series model prediction significantly improved the neural network's prediction of critical process parameters. Often in a manufacturing environment, the collection of adequate amounts of data for network training is difficult. This integrated technique offers potential for improving network performance without collecting additional data.     

基于参数估计误差的辊道窑温度场优化控制方法

辊道窑烧结过程的温度是决定锂离子电池正极材料产品质量的关键. 然而, 根据炉内有限个测温点的温度 建立起描述整个温度场的模型往往非常困难, 导致无法优化控制烧结过程的温度分布; 而控制方法的设计一般需要 进行参数估计, 已有参数估计方法大多依赖于观测器/预测器的状态误差信息, 无法直接反映待估计参数的变化特 征且方法的准确性取决于观测器/预测器的性能. 为此, 本文提出一种基于参数估计误差的温度场自适应动态规划 (adaptive dynamic programming, ADP)优化控制方法. 首先, 基于传热机理建立二维多孔介质能量守恒方程, 构建包 含角系数的边界条件以反映热辐射作用; 考虑到竖直方向温度变化较大, 通过转换边界条件建立起辊道窑一维温 度场模型, 并根据正极材料的特性获得模型参数. 然后, 采用ADP中的策略迭代(policy iteration, PI) 优化设计温度场 控制器, 神经网络(neural network, NN)用于PI中的评价网络以逼近代价函数; 基于权值参数的估计值与真实值之差 构建参数估计误差, 通过将估计误差的信息融入到评价NN参数更新过程, 提出基于参数估计误差的NN权值更新算 法, 以提高参数估计误差的收敛性, 实现有限时间内NN权值的快速收敛. 最后, 通过仿真验证所提建模和控制方法 的有效性.     

Training a neural network to select dispatching rules in real time

Dispatching rules are often suggested to schedule manufacturing systems in real-time. Numerous dispatching rules exist. Unfortunately no dispatching rule (DR) is known to be globally better than any other. Their efficiency depends on the characteristics of the system, operating condition parameters and the production objectives. Several authors have demonstrated the benefits of changing dynamically these rules, so as to take into account the changes that can occur in the system state. A new approach based on neural networks (NN) is proposed here to select in real time, each time a resource becomes available, the most suited DR. The selection is made in accordance with the current system state and the workshop operating condition parameters. Contrarily to the few learning approaches presented in the literature to select scheduling heuristics, no training set is needed. The NN parameters are determined through simulation optimization. The benefits of the proposed approach are illustrated through the example of a simplified flow-shop already published. It is shown that the NN can automatically select efficient DRs dynamically: the knowledge is only generated from simulation experiments, which are driven by the optimization method. Once trained offline, the resulting NN can be used online, in connection with the monitoring system of a flexible manufacturing system.     

Application of a genetic algorithm in solving the capacity allocation problem with machine dedication in the photolithography area

Wafer fabrication is a complicated manufacturing process with high process capability. Hence, maximizing machine capacity to meet customer deadlines is a very important issue in this field. This study proposes an integer programming model and a heuristic algorithm approach to solve the loading balance problem for the photolithography area in the semiconductor manufacturing industry. Considering process capability, machine dedication, and reticle constraints, we aim to minimize the difference in loading between machines. Process capability means that each product must be processed in machines that meet the process specification. Machine dedication means that if the first critical layer of a wafer is assigned to a certain machine, then the following critical layers of such wafer must be processed in this certain machine to ensure wafer quality. This research compares the results of two methods and finds the best parameter settings of the genetic algorithm (GA). The computational performance results of the GA shows that we can find the near-optimal solution within a reasonable amount of time. Finally, this research analyzes machine capability and reticle flexibility to determine the best percentage that can be used as reference for application in the semiconductor industry.     

A neural network model to determine the plate width set-up value in a hot plate mill

Performance of the process reducing the slab width in hot plate mill called edging is critical to produce rolled products with a desired dimension, which otherwise increase the yield loss caused by trimming. This process, therefore, requires a stringent width control performance. In this paper, an edger set-up model generating the desired slab width required for the control is proposed based upon the neural network approach. This neural network model accounts for variation of the dimension of incoming slabs to predict the preset value of the width as accurately as possible. A series of simulations were conducted to evaluate the performance of the neural network estimator for a variety of operating conditions needed for producing rolled products of various dimensions. The results show that the proposed model can estimate the preset value of the slab width with good accuracy, thereby enhancing the dimensional accuracy of rolled products. The estimation performance is discussed in detail for various process operation conditions.     

基于鲁棒迭代学习的灌浆压力自适应控制

为了提高非线性、不确定和时变性灌浆过程中压力的控制精度,在分析灌浆过程数学模型的基础上,提出了灌浆压力的PID控制器参数的自适应调节方法.由神经网络预测模型对灌浆系统进行非线性建模,然后基于神经网络学习误差迭代优化PID控制参数.为了确保控制器参数矩阵在调节时灌浆压力能收敛于灌浆设计压力,采用了李亚普洛夫误差增量迭代函数,使得对每次采样时刻系统误差PID调节向量能渐近收敛于最优值,从而使模型跟踪误差最小.通过迭代反馈调节方法的压力输出同手工控制方法对比研究,仿真结果表明,此方法有更好的自适应能力,较好地跟踪了灌浆设计压力曲线.     

Process parameter optimization of lap joint fillet weld based on FEM–RSM–GA integration technique

This study introduces a welding process design tool to determine optimal arc welding process parameters based on Finite Element Method (FEM), Response Surface Method (RSM) and Genetic Algorithms (GA). Here, a sequentially integrated FEM–RSM–GA framework has been developed and implemented to reduce the weld induced distortion in the final welded structure. It efficiently incorporates finite element based numerical welding simulations to investigate the desired responses and the effect of design variables without expensive trial experiments. To demonstrate the effectiveness of the proposed methodology, a lap joint fillet weld specimen has been used in this paper. Four process parameters namely arc voltage, input current, welding speed and welding direction have been optimized to minimize the distortion of the structure. The optimization results revealed the effectiveness of the methodology for welding process design with reduced cost and time.     

Optimizing the auto-brazing process quality of aluminum pipe and flange via a Taguchi-Neural-Genetic approach

This work describes an application of an integrated approach using the Taguchi method (TM), neural network (NN) and genetic algorithm (GA) for optimizing the lap joint quality of aluminum pipe and flange in automotive industry. The proposed approach (Taguchi-Neural-Genetic approach) consists of two phases. In first phase, the TM was adopted to collect training data samples for the NN. In second phase, a NN with a Levenberg-Marquardt back-propagation (LMBP) algorithm was adopted to develop the relationship between factors and the response. Then, a GA based on a well-trained NN model was applied to determine the optimal factor settings. Experimental results illustrated the Taguchi-Neural-Genetic approach.     

Learning-based multi-pass adaptive scheduling for a dynamic manufacturing cell environment

Because the essential attributes are uncertain in a dynamic manufacturing cell environment, to select a near-optimal subset of manufacturing attributes to enhance the generalization ability of knowledge bases remains a critical, unresolved issue for classical artificial neural network-based (ANN-based) multi-pass adaptive scheduling (MPAS). To resolve this problem, this study develops a hybrid genetic /artificial neural network (GA/ANN) approach for ANN-based MPAS systems. The hybrid GA/ANN approach is used to evolve an optimal subset of system attributes from a large set of candidate manufacturing system attributes and, simultaneously, to determine configuration and learning parameters of the ANN according to various performance measures. In the GA/ANN-based MPAS approach, for a given feature subset and the corresponding topology and learning parameters of an ANN decoded by a GA, an ANN was applied to evaluate the fitness in the GA process and to generate the MPAS knowledge base used for adaptive scheduling control mechanisms. The results demonstrate that the proposed GA/ANN-based MPAS approach has, according to various performance criteria, a better system performance over a long period of time than those obtained with classical machine learning-based MPAS approaches and the heuristic individual dispatching rules.     

Gait Pattern Adaptation for an Active Lower-Limb Orthosis Based on Neural Networks

This work deals with neural network (NN)-based gait pattern adaptation algorithms for an active lower-limb orthosis. Stable trajectories with different walking speeds are generated during an optimization process considering the zero-moment point (ZMP) criterion and the inverse dynamic of the orthosis–patient model. Additionally, a set of NNs is used to decrease the time-consuming analytical computation of the model and ZMP. The first NN approximates the inverse dynamics including the ZMP computation, while the second NN works in the optimization procedure, giving an adapted desired trajectory according to orthosis–patient interaction. This trajectory adaptation is added directly to the trajectory generator, also reproduced by a set of NNs. With this strategy, it is possible to adapt the trajectory during the walking cycle in an on-line procedure, instead of changing the trajectory parameter after each step. The dynamic model of the actual exoskeleton, with interaction forces included, is used to generate simulation results. Also, an experimental test is performed with an active ankle–foot orthosis, where the dynamic variables of this joint are replaced in the simulator by actual values provided by the device. It is shown that the final adapted trajectory follows the patient intention of increasing the walking speed, so changing the gait pattern.     

A genetic algorithm calibration method based on convergence due to genetic drift

The selection of Genetic Algorithm (GA) parameters is a difficult problem, and if not addressed adequately, solutions of good quality are unlikely to be found. A number of approaches have been developed to assist in the calibration of GAs, however there does not exist an accepted method to determine the parameter values. In this paper, a GA calibration methodology is proposed based on the convergence of the population due to genetic drift, to allow suitable GA parameter values to be determined without requiring a trial-and-error approach. The proposed GA calibration method is compared to another GA calibration method, as well as typical parameter values, and is found to regularly lead the GA to better solutions, on a wide range of test functions. The simplicity and general applicability of the proposed approach allows suitable GA parameter values to be estimated for a wide range of situations.     

基于一种新的基因操作策略的改进遗传算法

提出一种新的基因操作策略,该策略利用单纯形法的思想产生新样本,将遗传算法寻优的随机性与传统算法寻优的方向性有机地结合在一起.仿真结果表明,将改进的遗传算法用于训练神经网络辨识器,可提高收敛速度和模型拟合精度.     

Neural network based modeling and optimization of deep drawing – extrusion combined process

A combined deep drawing–extrusion process is modeled with artificial neural networks (ANN’s). The process is used for manufacturing synchronizer rings and it combines sheet and bulk metal forming processes. Input–output data relevant to the process was collected. The inputs represent geometrical parameters of the synchronizer ring and the outputs are the total equivalent plastic strain (TEPS), contact ratio and forming force. This data is used to train the ANN which approximates the input-output relation well and therefore can be relied on in predicting the process input parameters that will result in desired outputs provided by the designer. The complex method constrained optimization is applied to the ANN model to find the inputs or geometrical parameters that will produce the desired or optimum values of TEPS, contact ratio and forming force. This information will be very hard to obtain by just looking at the available historical input–output data. Therefore, the presented technique is very useful for selection of process design parameters to obtain desired product properties.     

Genetic algorithm optimization of a planar slot array using full wave method‐of‐moments analysis

Waveguide‐fed slot arrays in standing wave mode have been employed successfully in space based remote sensing radars because of their high efficiency, ease of deployment and their ability to withstand the radiation environment. Although the bandwidth requirement in such systems is minimal, at Ka band and above manufacturing tolerances in the order of 1 mil (25 μm), achieved in the dip brazing process, may affect their performance. To produce designs that are less sensitive to manufacturing tolerance, genetic algorithm (GA) optimization is employed in conjunction with a full wave analysis utilizing the method‐of‐moments solution to the pertinent integral equations of slot apertures of a planar array. In this work, a single 8 × 10 sub‐array of an interferometric antenna, proposed previously for a planetary mapping application, was investigated. The array was first designed by the Elliott's procedure and subsequently the design parameters were perturbed by GA optimization using the moment method analysis. The fitness parameter is a weighted function of return loss and gain over a number of frequencies in the operating band. A matching waveguide section consisting of inductive irises is also optimized using GA and mode matching technique. Optimum designs producing nearly constant gain and good return loss over 6% bandwidth are found to be less sensitive to manufacturing tolerance than the initial Elliott design. © 2013 Wiley Periodicals, Inc. Int J RF and Microwave CAE, 2013.     

Application of an evolutionary algorithm for parameter optimization in a gully erosion model

Herein we demonstrate how to use model optimization to determine a set of best-fit parameters for a landform model simulating gully incision and headcut retreat. To achieve this result we employed the Covariance Matrix Adaptation Evolution Strategy (CMA-ES), an iterative process in which samples are created based on a distribution of parameter values that evolve over time to better fit an objective function. CMA-ES efficiently finds optimal parameters, even with high-dimensional objective functions that are non-convex, multimodal, and non-separable. We ran model instances in parallel on a high-performance cluster, and from hundreds of model runs we obtained the best parameter choices. This method is far superior to brute-force search algorithms, and has great potential for many applications in earth science modeling. We found that parameters representing boundary conditions tended to converge toward an optimal single value, whereas parameters controlling geomorphic processes are defined by a range of optimal values.     

Local bifurcations analysis of a state-dependent delay differential equation

In this paper, a first-order equation with state-dependent delay and a nonlinear right-hand side is considered. The conditions of the existence and uniqueness of the solution of the initial value problem are supposed to be executed.The task is to study the behavior of solutions of the considered equation in a small neighborhood of its zero equilibrium. The local dynamics depends on real parameters, which are coefficients of the right-hand side decomposition in a Taylor series.The parameter, which is a coefficient at the linear part of this decomposition, has two critical values that determine the stability domain of the zero equilibrium. We introduce a small positive parameter and use the asymptotic method of normal forms in order to investigate local dynamics modifications of the equation near each two critical values. We show that the stability exchange bifurcation occurs in the considered equation near the first of these critical values, and the supercritical Andronov–Hopf bifurcation occurs near the second of these values (provided the sufficient condition is executed). Asymptotic decompositions according to correspondent small parameters are obtained for each stable solution. Next, a logistic equation with state-dependent delay is considered to be an example. The bifurcation parameter of this equation has the only critical value. A simple sufficient condition of the occurrence of the supercritical Andronov–Hopf bifurcation in the considered equation near a critical value has been obtained as a result of applying the method of normal forms.     

Theoretical and evolutionary parameter tuning of neural oscillators with a double-chain structure for generating rhythmic signals

A neural oscillator with a double-chain structure is one of the central pattern generator models used to simulate and understand rhythmic movements in living organisms. However, it is difficult to reproduce desired rhythmic signals by tuning an enormous number of parameters of neural oscillators. In this study, we propose an automatic tuning method consisting of two parts. The first involves tuning rules for both the time constants and the amplitude of the oscillatory outputs based on theoretical analyses of the relationship between parameters and outputs of the neural oscillators. The second involves an evolutionary tuning method with a two-step genetic algorithm (GA), consisting of a global GA and a local GA, for tuning parameters such as neural connection weights that have no exact tuning rule. Using numerical experiments, we confirmed that the proposed tuning method could successfully tune all parameters and generate sinusoidal waves. The tuning performance of the proposed method was less affected by factors such as the number of excitatory oscillators or the desired outputs. Furthermore, the proposed method was applied to the parameter-tuning problem of some types of artificial and biological wave reproduction and yielded optimal parameter values that generated complex rhythmic signals in Caenorhabditis elegans without trial and error.