Prediction of the plasma distribution using artificial neural network

In this work,an artificial neural network (ANN) model is established using a back-propagation training algorithm in order to predict the plasma spatial distribution in an electron cyclotron resonance (ECR)—plasma-enhanced chemical vapor deposition (PECVD) plasma system. In our model, there are three layers:the input layer, the hidden layer and the output layer. The input layer is composed of five neurons: the radial position, the axial position, the gas pressure, the microwave power and the magnet coil current. The output layer is our target output neuron: the plasma density. The accuracy of our prediction is tested with the experimental data obtained by a Langmuir probe, and ANN results show a good agreement with the experimental data. It is concluded that ANN is a useful tool in dealing with some nonlinear problems of the plasma spatial distribution.     

基于SOM和RBF网络的软测量方法研究

针对氢粉碎过程中钕铁硼粉碎状态不可知,为有效预测合金的反应状态,提出了一种基于自组织特征映射(SOM)神经网络和径向基函数(RBF)神经网络结合构建的网络模型。在该模型中,SOM神经网络作为聚类网络,采用无教师学习算法对输入样本进行自组织分类,并将分类中心及其对应的权值向量传递给RBF神经网络,作为径向基函数的中心;RBF神经网络作为基础网络,采用高斯函数作为径向基函数实现从输入到隐含层的非线性映射,输出层则采用有教师学习算法训练网络的权值,从而实现输入层到输出层的线性映射。并以钕铁硼氢粉碎过程合金中氢含量为检测对象,运用上述方法在MATLAB平台上建立了合金中氢含量预测模型,并完成了仿真验证。     

Characterisation of the two artificial neural plasma density with network models

This paper establishes two artificial neural network models by using a multi layer perceptron algorithm and radial based function algorithm in order to predict the plasma density in a plasma system. In this model, the input layer is composed of five neurons： the radial position, the axial position, the gas pressure, the microwave power and the magnet coil current. The output layer is the target output neuron： the plasma density. The accuracy of prediction is tested with the experimental data obtained by the Langmuir probe. The effectiveness of two artificial neural network models are demonstrated, the results show good agreements with corresponding experimental data. The ability of the artificial neural network model to predict the plasma density accurately in an electron cyclotron resonance-plasma enhanced chemical vapour deposition system can be concluded, and the radial based function is more suitable than the multi layer perceptron in this work.     

Modeling of Thermal Conductivity of Concrete with Vermiculite by Using Artificial Neural Networks Approaches

In this article, the thermal conductivity of concrete with vermiculite is determined and also predicted by using artificial neural networks approaches, namely the radial basis neural network and multi-layer perceptron. In these models, 20 datasets were used. For the training set, 12 datasets (60%) were randomly selected, and the residual datasets (8 datasets, 40%) were selected as the test set. The root mean square error, the mean absolute error, and determination coefficient statistics are used as evaluation criteria of the models, and the experimental results are compared with these models. It is found that the radial basis neural network model is superior to the other models.     

A spintronic memristive circuit on the optimized RBF-MLP neural network

A radial basis function network (RBF) has excellent generalization ability and approximation accuracy when its parameters are set appropriately. However, when relying only on traditional methods, it is difficult to obtain optimal network parameters and construct a stable model as well. In view of this, a novel radial basis neural network (RBF-MLP) is proposed in this article. By connecting two networks to work cooperatively, the RBF's parameters can be adjusted adaptively by the structure of the multi-layer perceptron (MLP) to realize the effect of the backpropagation updating error. Furthermore, a genetic algorithm is used to optimize the network's hidden layer to confirm the optimal neurons (basis function) number automatically. In addition, a memristive circuit model is proposed to realize the neural network's operation based on the characteristics of spin memristors. It is verified that the network can adaptively construct a network model with outstanding robustness and can stably achieve 98.33% accuracy in the processing of the Modified National Institute of Standards and Technology (MNIST) dataset classification task. The experimental results show that the method has considerable application value.     

尖楔前体飞行器FADS系统的径向基网络建模及验证

文章研究了针对一种用于尖楔外形的嵌入式大气数据传感（flush air data sensing，FADS）系统的解算模型及精度.首先基于飞行包络及CFD数据建立了FADS系统的测压孔选取标准；然后基于径向基函数（radial basis function，RBF）的人工神经网络建模技术构建了FADS系统的网络解算模型；最后给出了模型的测试误差，分析了气动延时效应、位置误差等误差源模型对算法精度的影响，并给出了网络模型的预测精度.结果表明，针对尖楔外形测压孔配置特征，基于RBF的人工神经网络算法解算精度较好，攻角、侧滑角、Mach数及静压的网络输出预测值与真实值吻合较好，输出的测试误差（绝对值）分别小于0.25°，0.5°，0.05及250 Pa.结果同时表明神经网络建模技术在尖楔前体飞行器FADS系统中的有效性.      

Application of artificial neural networks to X-ray fluorescence spectrum analysis

X-ray fluorescence (XRF) is widely applied as a mature nondestructive testing method, and appropriate improvement of quantitative analysis methods can improve the accuracy of XRF. Artificial neural network is an intelligent information processing system, its developments and application in XRF are reviewed, and representative models (back propagation, radial basis function, genetic algorithm artificial neural network, and others) are discussed in more details in overfitting, generalization, and algorithm efficiency. Potential directions of developing artificial neural network applied in XRF are proposed in this review as a further study.     

Modeling of a hemispherical inductively coupled plasma using neural network

In this study, a hemispherical inductively coupled plasma (HICP) was modeled by using a neural network called a radial basis function network (RBFN). The prediction performance of RBFN models were optimized by using a genetic algorithm. Using a Langmuir probe, experimental data were collected from the HICP equipment of 10 turns. For a systematic modeling, plasma discharge was characterized by using a statistical experiment. The process parameters involved include a radio frequency source power, pressure, position of probe tip, and Cl₂ flow rate. The plasma characteristics modeled include plasma density and electron temperature. From the optimized models, 3D plots were generated to explore parameter effects. Plasma density (or electron temperature) was the most strongly dependent on the tip position. The effect of source power on plasma density was almost independent of Cl₂ flow rate. The effect of pressure was inclined to slightly decrease plasma density. Unlike in other plasma sources, electron temperature was little affected by pressure. The effect of Cl₂ flow rate of increasing electron temperature was the most significant under higher plasma density.     

Application of functional link artificial neural network for mitigating nonlinear effects in coherent optical OFDM

Artificial Neural Networks are developed as an important technique for equalization and have been widely used to mitigate the nonlinear effects in coherent optical systems. For the compensation of nonlinearities in coherent optical orthogonal frequency division multiplexing technique, the most popular artificial neural network model is a multilayer perceptron (MLP), as it is able to perform complex mapping between input and output spaces with significant success. However due to the complexity of multilayer perceptron nonlinear equalizer (MLP-NLE) model training of neural network is difficult. To overcome computational complexity issues of MLP-NLE, a single neuron based functional link artificial neural network nonlinear equalizer (FLANN-NLE) has been developed in this paper. Better performance of an equalizer is attributed to the usage of aPSO-BP algorithm for training the FLANN-NLE. The proposed FLANN-NLE surpasses the existing works both in terms of Q-Factor and computational complexity. For a fiber length of 1000 km and at launch power of ?6 dBm, the improvement in Q-Factor is approximately equal to 3.3 and 1 dB in contrast to the previously reported values of approximately 3 and 0.7 dB at bit rate of 40 and 80 Gbps respectively.     

Perceptron-like computation based on biologically-inspired neurons with heterosynaptic mechanisms

Perceptrons are one of the fundamental paradigms in artificial neural networks and a keyprocessing scheme in supervised classification tasks. However, the algorithm they provideis given in terms of unrealistically simple processing units and connections andtherefore, its implementation in real neural networks is hard to be fulfilled. In thiswork, we present a neural circuit able to perform perceptron’s computation based onrealistic models of neurons and synapses. The model uses Wang-Buzsáki neurons withcoupling provided by axodendritic and axoaxonic synapses (heterosynapsis). The maincharacteristics of the feedforward perceptron operation are conserved, which allows tocombine both approaches: whereas the classical artificial system can be used to learn aparticular problem, its solution can be directly implemented in this neural circuit. As aresult, we propose a biologically-inspired system able to work appropriately in a widerange of frequencies and system parameters, while keeping robust to noise and error.     

基于深度神经网络的HL-2A等离子体水平位移研究

基于门控循环单元(GRU)的神经网络,构建预测模型的网络拓扑结构,训练和测试了HL-2A装置等离子体水平位移系统响应模型。测试结果显示了该模型对43%的样本数据的拟合度超过80%。把该网络模型作为被控对象,使用基于径向基函数(RBF)神经网络的模型参考自适应控制(MRAC)算法,设计了一个HL-2A等离子体水平位移的MRAC系统。仿真结果显示,该控制系统的输出响应能快速地跟踪各种输入参考信号,控制器能够较好地控制等离子体的水平位移并具有强的抗扰动能力。     

基于枸杞红外光谱人工神经网络的产地鉴别

应用红外光谱技术,快速鉴别枸杞药材产地。利用傅里叶变换红外光谱, 测定了45个来自青海省不同产地的枸杞样品的红外光谱。以常规预处理方法和小波变换对红外光谱原始数据进行了预处理。对比常用的窗口移动平滑预处理、标准正态变换以及多元散射校正, 小波变换是一种有效实用的光谱预处理方法。为了提高神经网络的训练速度, 在利用人工神经网络建立模型之前, 通过小波变换的方法对光谱变量进行了压缩, 同时对建立的模型的相关参数进行了详细的讨论。结果表明,红外光谱数据压缩到原来的1/8,其分析精度与原始光谱数据基本相当。以压缩了的光谱数据作为反向传播(BP)网络的输入变量,产地类别作为神经网络的输出变量,建立3层人工神经网络。其中隐含层神经元个数为5个,输出层神经元个数为1个。隐层的传递函数是tansig,输出层传递函数是purelin,网络训练函数trainlm,权阈值的学习函数是learngdm。net.trainParam.epochs=1 000,net.trainParam.goal=0.001。对10个未知枸杞产地类别进行了预测,预测结果准确率达100%。实验表明, 建立的模型能够正确地对枸杞样品快速地进行产地鉴别。红外光谱法结合人工神经网络可作为中药材产地分类鉴别的一种新的现代化方法。     

基于NW型小世界人工神经网络的污水出水水质预测

为了预测污水处理出水水质,针对污水处理过程具有多变量、非线性、时变性、严重滞后的特点,提出了基于NW型小世界人工神经网络的污水处理出水水质预测模型。首先根据污水处理系统确定模型输入输出变量个数,然后建立了多层前向小世界神经网络模型,并对网络模型的隐层结构进行了优化研究。借助污水处理过程的历史数据进行了仿真研究,结果表明：和同规模的多层前向人工神经网络相比,小世界神经网络对污水出水水质预测具有较高精度和收敛速度,为污水出水水质的实时预测提供了一种有效的新方法。     

基于深度稀疏学习的土壤近红外光谱分析预测模型

提出一种基于深度稀疏学习的土壤近红外光谱分析预测模型。首先,使用稀疏特征学习方法对土壤近红外光谱数据进行约简,实现土壤近红外光谱内容的稀疏表示;然后采用径向基函数神经网络以稀疏表示特征系数为输入,以所测土壤成分为输出,分别建立土壤有机质、速效磷、速效钾的非线性预测模型。结果表明用该模型预测土壤有机质的含量是可行的,但对土壤速效磷和速效钾含量的预测还需对模型做进一步的优化。     

基于模糊递归小波神经网络的葡萄酒品质预测

针对葡萄酒品质预测模型难以建立的问题,提出一种基于模糊递归小波神经网络的葡萄酒品质预测模型。利用葡萄酒物理化学指标和品酒师打分作为模型的输入输出,采用梯度下降算法在线学习隶属函数层中心、宽度和小波函数平移因子、伸缩因子、自反馈权重因子以及输出层权值。仿真实验时,首先利用Mackey-Glass混沌时间序列进行了性能测试,然后利用UCI数据集葡萄酒品质数据对所建立的品质预测模型进行了验证。结果显示,与多层感知器、径向基函数神经网络等传统前馈神经网络相比,构建的模糊递归小波神经网络品质预测模型具有更高的预测精度,更加适合于葡萄酒的品质预测。     

基于BP神经网络的多传感器信息融合研究

BP网络是应用最广的一种人工神经网络,将BP神经网络应用到压力检测领域的温度等非线性补偿,具有重要的实用价值,对压力检测精度的改进效果显著。从传感器信息融合的角度看,神经网络就是一个融合系统。通过对神经网络基本理论的阐述,针对研究对象将BP神经网络原理与多传感器信息融合技术有机集合起来,提出了基于BP神经网络的二传感器信息融合模型及改进算法,建立了BP神经网络训练标准样本库,并对该网络模型进行主要技术指标的测试和仿真工作,测试结果表明构建的模型及其改进算法能很好地满足了高精度压力检测仪的指标要求。     

近红外光谱的北方寒地土壤含水率预测模型研究

我国北方寒地温差大,土壤温差对近红外光谱测量土壤墒情有较大影响。针对这一问题,以北方寒地土壤为研究对象,探究大范围温度胁迫下（-20～40 ℃）土壤的近红外光谱与土壤不同含水率之间的关系预测模型方法。选取黑龙江八一农垦大学农学院试验基地中的黑土,经烘干、过筛等操作处理后配置含水率范围在15%～50%内八种不同湿度的土壤样品,建立北方寒地土壤大范围温度胁迫下土壤的近红外光谱信息与含水率之间的定量预测模型。在全波段光谱数据的基础上,结合五种不同光谱信号预处理方法,采用BP神经网络算法、优化支持向量机算法（SVM）、高斯过程算法（GP）三种智能算法建立北方寒地土壤近红外光谱与含水率的预测模型并验证模型的效果。利用69组数据进行训练建模, BP神经网络相关参数设置为学习速率0.05,最大训练次数设置为5 000,隐层单元数确定为20;SVM采用径向基函数,并利用leave-one-out cross validation确定了最佳惩罚参数为0.87,使模型预测的准确性提高;高斯过程算法内部采用马顿核。模型的定量评估采用决定系数(R2)和均方根误差（RMSE）。结果表明,在建立的全部BP神经网络模型中,效果最佳的为S_G-BP神经网络模型,模型的R2为0.960 9,RMSE为2.379 7;在SVM模型中SNV-SVM模型的效果最好,模型的R2为0.991 1,RMSE为1.081 5;在GP模型中S_G-GP模型的效果最好,模型的R2为0.928,RMSE为3.258 1,综上基于SNV预处理的SVM模型训练效果最优。利用剩余的35组光谱数据作为预测集验证模型性能,经模型对比分析发现基于SVM算法的预测模型效果优于其他两种算法,其中基于S_G的SVM模型效果最优,其预测模型的R2和差RMSE分别为0.992 1和0.736 9。综合建模集与预测集的参数最终确定基于S_G的SVM模型为最佳模型。此模型可以作为大范围温度胁迫条件下（寒地）的土壤含水率有效预测方法,为设计优化适宜寒地便携式近红外土壤含水率快速测量仪提供科学依据。     

DVD光学头的RBF神经网络自适应PID控制器设计

为了使三维光存储技术的应用水平得到提高,以DVD伺服技术、双光子吸收技术为基础组建了一套信息存储系统。针对DVD光学读取头系统,采用RBF神经网络自适应PID控制器进行控制,充分利用RBF神经网络的自学习和全局非线性逼近能力,在线调整修正PID控制器的3个参数,使其达到一种最优控制,并通过MATLAB软件进行了计算机仿真。由仿真结果可以得出:通过应用RBF神经网络自适应PID控制算法,系统单位阶跃响应的调整时间为0.25 s,并使系统的超调量降低到几乎为零。     

基于SFLA优化BP神经网络的语音信号分类

BP神经网络是一种多层前馈网络,数据经过网络的输入层、隐含层逐层处理后,由输出层进行输出,通过和期望输出的对比进行反向传播,调整网络参数使输出不断逼近期望输出;在使用BP神经网络对语音特征信号进行分类的过程中,会出现BP神经网络易陷入局部最优解、学习收敛速度慢的问题;针对此问题提出一种基于SFLA优化BP神经网络权值和阀值的方法,引入SFLA算法优化网络权值和阀值,利用SFLA优化后的BP网络模型进行语音特征信号分类;仿真结果表明,经SFLA优化后的BP神经网络与未优化的神经网络相比,不仅训练速度快, 而且误差小,语音特征信号分类的正确率平均提高1.31%。     

基于粒子群优化支持向量机的太阳电池温度预测

建立通用而精确的太阳电池热模型对光伏系统的建模、输出功率与转换效率的损失分析至关重要. 基于复杂的太阳电池温度机理, 分别研究了太阳电池温度的稳态热模型(steady state thermal model, SSTM)和支持向量机(support vector machines, SVM) 方法建立的精确预测热模型. 首先, 基于空气温度、太阳辐射强度、风速3个最主要因素与太阳电池温度的近似线性关系, 在已有SSTM的基础上, 建立并校正了太阳电池的SSTM并采用差分进化算法提取模型的未知参数. 其次, 为提高SVM的模型预测精度, 采用粒子群优化(particle swarm optimization, PSO) 算法对SVM的核参数和惩罚因子进行动态寻优, 在确定输入/输出样本集并划分训练集和测试集的基础上, 建立了基于粒子群优化支持向量机(PSO-SVM)的太阳电池温度精确预测热模型. 最后, 搭建实验平台, 在实验操作过程中减弱空气湿度、太阳入射角和热迟滞效应等因素对太阳电池温度的耦合. 通过实验对比表明, 建立的预测热模型性能可靠、全面、简洁, 其参数寻优算法优于遗传算法和交叉校验法, 模型预测精度优于反向传播神经网络(back propagation neural network) 和SSTM.