基于小波网络的非线性动态过程建模和在环氧氯丙烷生产过程中的应用

A type of wavelet neural network, in which the scale function is adopted only,is proposed in this paper for non-linear dynamic process modelling.Its network size is decreased significantly and the weight coefficients can be estimated by a linear algorithm.The wavelet neural network holds some advantages supeiior to other types of neural networks.First, its network structure is easy to specify based on its theoretical analysis and intuition.Secondly, network training does not rely on stochastic gradient type techniques and avoidd the problem of poor convergence or undesirable local minima.The excellent statistic properties of the weight parameter estimations can be proven here.Both theoretical analysis and simulation study show that the identification method is robust and reliable. Furthermore,a hybrid network structure incorporating first-principle knowledge and wavelet network is developed to solve a commonly existing problem in chemical production processes.Applications of the hybrid network to a practical production process demonstrates that model generalisation capability is significantly improved.     

Neural network intelligent system for the on -line optimization in chemical plants

A strategy of developing on-line optimization intelligent systems based on combiningflowsheeting simulation and optimization package with artificial neural networks(ANN)is presented inthis paper.A number of optimization cases for a certain chemical plant are obtained off-line byusing PROCESS-Ⅱ or other flowsheeting programming with optimization.Then,taking these cases astraining examples,we establish a neural network systems which can be used on-line as an optimizer toobtain setpoints from input data sampled from distributed control system through gross error detectionand data reconciliation procedures.Such an on-line optimizer possesses two advantages over nonlinearprogramming package:first of all,there is no convergence problem for the trained ANN to be usedonline;secondly,the frequency for setpoints updating is not limited because only algebraic calculationrather than optimization is required to be carried out on-line.Here two key problems ofimplementing ANN approaches to the on-line optimization ar     

A simulation of diesel hydrotreating process with real component method

Computer simulation is a good guide and reference for development and research on petroleum refining process-es. Traditionally, pseudo-components are used in the simulation, in which their physical properties are estimated by empirical relations and cannot be associated with actual chemical reactions, as no molecular structure is avail-able for pseudo-components. This limitation can be overcome if real components are used. In this paper, a real component based method is proposed for the simulation of a diesel hydrotreating process by using the software of Unisim Design. This process includes reaction units and distillation units. The chemical reaction network is established by analyzing the feedstock. The feedstock is characterized by real components, which are obtained based on true boiling point curve. Simulation results are consistent with actual data.     

输入训练神经网络的维数约简算法及其在化工过程建模中的应用

Many applications of principal component analysis （PCA） can be found in dimensionality reduction. But linear PCA method is not well suitable for nonlinear chemical processes. A new PCA method based on improved input training neural network （IT-NN） is proposed for the nonlinear system modelling in this paper. Momentum factor and adaptive learning rate are introduced into learning algorithm to improve the training speed of IT-NN. Contrasting to the auto-associative neural network （ANN）, IT-NN has less hidden layers and higher training speed. The effectiveness is illustrated through a comparison of IT-NN with linear PCA and ANN with experiments. Moreover, the IT-NN is combined with RBF neural network （RBF-NN） to model the yields of ethylene and propylene in the naphtha pyrolysis system. From the illustrative example and practical application, IT-NN combined with RBF-NN is an effective method of nonlinear chemical process modelling.     

酶膜耦合反应器中反冲过程的神经网络模拟

In the enzymatic membrane reactor for separating casein hydrolysate, backflushing technology has been used to decrease the fouling of the membrane. Predication of the backflushing efficiency poses a complex non-linear problem as the system integrates enzymatic hydrolysis, membrane separation and periodic backflushing together. In this paper an alternative artificial neural network approach is developed to predict the backflushing efficiency as a function of duration and interval. A contour plot of backflushing performance is presented to model these effects, and the backflushing conditions have been optimized as duration of 10 s and interval of 10 min using this neural network. Also, simple neural networks are established to predict the time evolution of flux before and after backflushing. The results predicted by the models are in good agreement with the experimental data, and the average deviations for all the cases are well within ±5%. The neural network approach is found to be capable of modeling the backflushing with confidence.     

Fault diagnosis for distillation process based on CNN–DAE

Distillation is the most widely used operation for liquid mixture separation in the chemical industry. It is of great importance to detect and diagnose faults in distillation process. Due to the strong feedback and coupling of processes in a distillation column, it is difficult to use deep auto-encoders(DAEs) alone to achieve good results in detecting and diagnosing faults, in terms of accuracy and efficiency. This paper proposes a hybrid fault-diagnosis model based on convolutional neural networks(CNNs) and DAEs, by integrating the powerful capability of CNN in feature extraction and of DAE in classification. A case study was carried out with the distillation process of depropanization. It is shown that the proposed hybrid model is of good performance compared to other models, in terms of the accuracy of fault detection in such a process. Also, with the increase of structural layers of the CNN–DAE model, the diagnostic accuracy will be improved, with an optimal accuracy of 92.2%.     

DOBD Algorithm for Training Neural Network: Part I. Method

Overfitting is one of the important problems that restrain the application of neural network. The traditional OBD (Optimal Brain Damage) algorithm can avoid overfitting effectively. But it needs to train the network repeatedly with low calculational efficiency. In this paper, the Marquardt algorithm is incorporated into the OBD algorithm and a new method for pruning network－the Dynamic Optimal Brain Damage (DOBD) is introduced. This algorithm simplifies a network and obtains good generalization through dynamically deleting weight parameters with low sensitivity that is defined as the change of error function value with respect to the change of weights. Also a simplified method is presented through which sensitivities can be calculated during training with a little computation. A rule to determine the lower limit of sensitivity for deleting the unnecessary weights and other control methods during pruning and training are introduced. The training course is analyzed theoretically and the reason why DOBD algorithm can obtain a much faster training speed than the OBD algorithm and avoid overfitting effectively is given.     

运用人工神经网络关联二元混合物中气液相平衡

In this paper, a back propagation artificial neural network （BP-ANN） model is presented for the simultaneous estimation of vapour liquid equilibria （VLE） of four binary systems viz chlorodifluoromethan-carbondioxide, trifluoromethan-carbondioxide, carbondisulfied-trifluoromethan and carbondisulfied-chlorodifluoromethan. VLE data of the systems were taken from the literature for wide ranges of temperature （222.04-343.23K） and pressure （0.105 to 7.46MPa）. BP-ANN trained by the Levenberg-Marquardt algorithm in the MATLAB neural network toolbox was used for building and optimizing the model. It is shown that the established model could estimate the VLE with satisfactory precision and accuracy for the four systems with the root mean square error in the range of 0.054-0.119. Predictions using BP-ANN were compared with the conventional Redlich-Kwang-Soave （RKS） equation of state, suggesting that BP-ANN has better ability in estimation as compared with the RKS equation （the root mean square error in the range of 0.115-0.1546）.     

PVC生产过程故障传感器的检测与重构

Based on principal component analysis, this paper presents an application of faulty sensor detection and reconstruction in a batch process, polyvinylchloride (PVC) making process. To deal with inconsistency in process data, it is proposed to use the dynamic time warping technique to make the historical data synchronized first,then build a consistent multi-way principal component analysis model. Fault detection is carried out based on squared prediction error statistical control plot. By defining principal component subspace, residual subspace and sensor validity index, faulty sensor can be reconstructed and identified along the fault direction. Finally, application results are illustrated in detail by use of the real data of an industrial PVC making process.     

基于支持向量机回归的天然气负荷预测研究

Natural gas load forecasting is a key process to the efficient operation of pipeline network. An accurate forecast is required to guarantee a balanced network operation and ensure safe gas supply at a minimum cost. Machine learning techniques have been increasingly applied to load forecasting. A novel regression technique based on the statistical learning theory, support vector machines (SVM), is investigated in this paper for natural gas shortterm load forecasting. SVM is based on the principle of structure risk minimization as opposed to the principle of empirical risk minimization in conventional regression techniques. Using a data set with 2 years load values we developed prediction model using SVM to obtain 31 days load predictions. The results on city natural gas short-term load forecasting show that SVM provides better prediction accuracy than neural network. The software package natural gas pipeline networks simulation and load forecasting (NGPNSLF) based on support vector regression prediction has been developed, which has also been applied in practice.     

一种新的选择性神经网络集成方法及其在PTA中的应用

神经网络集成可以显著提高神经网络的泛化性能。传统的集成方法中大都采用将训练的所有网络直接进行组合的方式形成集成网络,而实际上这些网络可能具有一定的相关性。为此,选择性神经网络集成成为目前研究的热点,它能够进一步提高集成网络的泛化性能。本文提出了一种利用网络权值计算网络模型之间差异度的新的选择性神经网络集成方法DWSEN。UCI数据测试表明,与流行的集成方法Bagging和Boosting比较,本方法有着更好的泛化能力和稳定性。将DWSEN应用于精对苯二甲酸（PTA）溶剂系统脱水塔装置的建模过程,结果显示,利用该方法训练得到的集成模型具有更好的泛化性能,能够较好地模拟生产运行过程。     

基于基团贡献神经网络集成法估算有机物常压凝固点

基于基团贡献法应用人工神经网络对有机物的常压凝固点进行了估算,输入参数为有机物的基团数和表征异构体的参数,输出为常压凝固点.分析了采用最速梯度下降法的BP算法在训练过程中产生误差饱和情况的原因,采用在隐含层节点中加入误差饱和预防函数用来防止误差饱和情况的出现.仿真结果表明,所采用的方法能有效地减小网络在误差表面陷入低谷的可能性和提高网络的收敛速率.采用神经网络集成法建立了神经网络集成模型,通过仿真合理选择隐含层节点数和采用交叉验证法用于防止BP网络的过度训练,增强了网络的泛化能力.估算结果表明,所建立的神经网络集成模型,其网络有良好的稳定性和预测精度,207个样本估算的绝对平均相对误差为8.62%.     

基于径向基神经网络的聚丙烯熔融指数预报

引　言在化工生产中大部分生产流程具有非线性、大时滞、结构复杂等特性 ,而且生产变量之间存在着不同程度的耦合与关联 .前馈神经网络由于具有强大的拟合非线性函数的能力 ,已成为生产指标预测的有力工具[1] .其中径向基 (radialbasisfunction ,RBF)神经网络相对于神经网络BP     

Study and Application of Fault Prediction Methods with Improved Reservoir Neural Networks

Time-series prediction is one of themajor methodologies used for fault prediction. Themethods based on recurrent neural networks have been widely used in time-series prediction for their remarkable non-liner mapping ability. As a new recurrent neural network, reservoir neural network can effectively process the time-series prediction. However, the ill-posedness problemof reservoir neural networks has seriously restricted the generalization performance. In this paper, a fault prediction algorithm based on time-series is proposed using improved reservoir neural networks. The basic idea is taking structure risk into consideration, that is, the cost function involves not only the experience risk factor but also the structure risk factor. Thus a regulation coefficient is introduced to calculate the outputweight of the reservoir neural network. As a result, the amplitude of outputweight is effectively controlled and the ill-posedness problemis solved. Because the training speed of ordinary reservoir networks is naturally fast, the improved reservoir networks for time-series prediction are good in speed and generalization ability. Experiments on Mackey-Glass and sunspot time series prediction prove the effectiveness of the algorithm. The proposed algorithm is applied to TE process fault prediction. We first forecast some timeseries obtained from TE and then predict the fault type adopting the static reservoirs with the predicted data. The final prediction correct rate reaches 81%.     

优化增量型随机权神经网络及应用

针对传统增量型随机权神经网络(I-RVFLNs)存在网络参数难以优化确定、模型收敛速度慢和结构复杂的问题,提出一种优化增量型随机权神经网络算法,即O-I-RVFLNs。与传统I-RVFLNs不同,所提O-I-RVFLNs算法首先设定了一个期望的建模残差向量,然后在每次新增隐层节点时,选择可以达到或小于此节点期望残差的输入权值和偏置作为该节点的输入参数,进而提高网络的收敛速度。除此之外,考虑到算法在不断迭代更新过程中建模误差越来越小,下降趋势越来越不明显的问题,将各指标参数相邻两次迭代均方根误差的差值考虑在算法终止条件内,并借鉴统计过程控制中的西电规则制定了相应的算法收敛判定准则。最后,基于UCI能效数据和实际高炉工业数据,对所提O-I-RVFLNs算法进行了验证和应用。结果表明,相对于其他RVFLNs算法,所提算法建立的数据模型能够获得更紧凑的网络结构以及更好的泛化性能和预测精度。     

基于深度集成支持向量机的工业过程软测量方法

基于支持向量机（SVM）的软测量建模方法已经在工业过程控制领域得到广泛应用,然而传统支持向量机直接针对原始测量变量建立模型,未能充分挖掘数据的内在特征信息以提高预测精度。针对该问题,本文提出一种基于深度集成支持向量机（DESVM）的软测量建模方法。该方法首先利用深度置信网络（DBN）来对数据进行深层次的信息挖掘,提取出数据的内在特征,然后引入基于Bagging算法的集成学习策略,构建基于深度数据特征的集成支持向量机模型,以提升软测量预测模型的泛化能力。最后通过数值系统和真实工业数据对方法进行应用分析,结果表明本文提出的方法能够有效提升支持向量机软测量模型的预测精度,能够更好地预测过程质量指标的变化。     

Artificial neural network-based modeling of pressure drop coefficient for cyclone separators

In order to build the complex relationships between cyclone pressure drop coefficient (PDC) and geometrical dimensions, representative artificial neural networks (ANNs), including back propagation neural network (BPNN), radial basic functions neural network (RBFNN) and generalized regression neural network (GRNN), are developed and employed to model PDC for cyclone separators. The optimal parameters for ANNs are configured by a dynamically optimized search technique with cross-validation. According to predicted accuracy of PDC, performance of configured ANN models is compared and evaluated. It is found that, all ANN models can successfully produce the approximate results for training sample. Further, the RBFNN provides the higher generalization performance than the BPNN and GRNN as well as the conventional PDC models, with the mean squared error of 5.84 × 10^?4 and CPU time of 120.15 s. The result also demonstrates that ANN can offer an alternative technique to model cyclone pressure drop.     

基于优化神经网络的高拱坝力学参数反演

高拱坝力学性能参数变化规律复杂,使用人工智能算法进行预测已经成为反演参数的重要手段。使用遗传算法对神经网络进行优化来检验优化后算法的性能,并比较不同算法应用于参数反演中预测结果的精度。根据某高拱坝运行期变形监测数据,分别使用RBF神经网络和遗传算法优化的BP(GA-BP)神经网络对不同水位工况下的坝段分区混凝土弹性模量进行反演。基于反演结果进行有限元正分析计算,将所得结果与实测数据进行对比,检验反演精度和效率。结果表明:GA-BP网络的最大预测误差为1.8%,相比于RBF网络预测精度提高了约50%。使用神经网络进行拱坝力学参数反演实用性好,优化后的神经网络比传统BP神经网络在计算精度和效率两方面均有明显改进,且GA-BP神经网络反演比RBF神经网络反演精度更高。     

基于改进Bagging算法的高斯过程集成软测量建模

为提高对工况复杂的工业过程进行软测量建模的模型精度和泛化能力,提出了一种基于改进Bagging算法的高斯过程集成软测量建模方法。该算法采用高斯过程回归算法建立集成学习模型的基学习器,并在Bagging算法对训练样本重采样生成基学习器训练子集的基础上,采用基于正则化互信息的特征排序指标进行基学习器的输入特征抽取,实现有监督的特征扰动,从而改善学习器的差异度。待测样本进行软测量估计时,根据各高斯过程基学习器输出的方差自适应地选择基学习器进行集成输出。采用工业双酚A生产装置反应器的现场数据建模仿真,结果表明该方法是有效的。     

高聚物流变性能预测与研究

根据影响高聚物流变性能的各因素之间的关系,运用Matlab软件建立一个人工神经网络。对5种高聚物材料进行实验;用实验数据对网络进行训练,并对高聚物流变性能进行预测;将预测和实测的结果进行分析对比。结果表明,用人工神经网络的方法预测高聚物的流变性能是可行的,预测精度较高,平均误差小于5%。