广义回归神经网络的改进及在延迟焦化建模中的应用

广义回归神经网络（GRNN）具有明确的概率意义,其参数大多能自动确定,仅光滑因子参数需优化估值.采用优进遗传算法（EGA）,将确定性与随机性寻优操作相融合,实现了高效全局搜优,它所基于的优进策略包括设计Powell寻优算子、改进交叉算子、自适应地调整交叉率和变异率等.以推广能力作为优化目标,所建的GRNN有很强的非线性拟合能力和优良的预报性能,将其成功地为延迟焦化过程建模,与径向基网络（RBFN）等相比,显示了明显的优势.     

CGA-RBFN模型及其在丙烯产率预测中的应用

利用混沌变量的遍历性和不规则性,将其引入遗传算法,可提高其全局搜优的性能;采用混沌遗传算法(CGA)训练径向基函数网(RBFN),并均衡地兼顾网络的拟合与预报能力,恰当地设计适应度函数,由此建成的CGA-RBFN模型,其预测能力与稳健性都有提高。将其应用于烃类热裂解丙烯预测,效果良好,与传统方法相比有明显的优越性。     

RBF-CSR-EHCGA模型及其在脉冲萃取中的应用

径向基函数-循环子空间回归（RBF-CSR）是一种有效的非线性网络模型,以高斯条为基函数,性能更优,但其参数多,且难以选定,将显著影响模型性能.为此,本文提出基于优进策略的混合编码遗传算法（EHCGA）,以不同的方式为各类参数编码,并引入确定性的Powell算子,提高全局搜优效率.EHCGA算法以模型预报性能为目标,优选参数,以此建立RBF-CSR-EHCGA模型,它的预报精度高、稳定性良好.已成功应用于回收己内酰胺的脉冲萃取过程建模,效果良好,明显优于其他网络模型,也优于近似机理模型.     

基于RBFN-PLSR方法的CO2提纯塔模型

CO2提纯塔出口浓度与其影响因素之间存在复杂的非线性关系,传统的线性回归和非线性回归方法难以建立起它们之间准确的关系模型。本文运用RBFN的最佳函数逼近性能,结合PLSR的空间变换方法,建立了CO2提纯塔模型。交叉验证表明,所建模型平均拟合相对误差为0．0063%,平均预报相对误差为0．1210％,该模型可用于提纯塔出口CO2浓度的预测。     

用改进型RBF网络进行催化剂活性估值

引言Chen等［1]提出了一种径向基函数网络（RBFN）离线学习方法．事实上,RBFN的特点使它更适合于在线过程辨识,Pottmann等［2］提出了一种基于正交规则和统计检验的分段回归算法来设计和训练RBFN,吕强等［3]提出了一种采用递推K-均值族和Kalman滤波的混合算法（KKRBFN）．为了提高RBFN递推训练算法的精度和实时性,本文提出了一种基于递推K-均值簇和改进递推折息法的混和算法的RBFN;并将该RBFN算法用于某固定床反应器催化剂活性系数的估计,以验证所提出RBFN学习算法的有效性．l改进型RBFN的学习策略径向基函数网络可…     

发酵动力学建模的新方法

提出一种基于径向基函数网络(RBFN)和偏最小二乘回归(PLSR)技术的发酵动力学建模新方法.通过对谷氨酸发酵过程中菌体生长模型为例进行研究,研究结果证明了该方法的有效性.     

RBF-MCSR方法用于二甲苯异构化装置的建模

推导了多因变量循环子空间回归 (MCSR)算法 ,并将MCSR集成于径向基网络 (RBFN)的输出端 ,由此提出了RBF—MCSR方法 ,它能表达复杂的非线性关系 ,而且在更为宽广的解空间内选取具有简明的解析形式的最优模型。将该法应用于二甲苯异构化装置 ,效果良好 ,与现有的RBF -PLSR比较 ,显示出其非线性建模的优势     

基于互补神经网络级联模型的聚丙烯MFR预报

提出了一种基于集成数据处理的,由高斯基-自适应复合基函数构成的互补径向基函数(RBF)神经网络系统和隐马尔科夫模型(HMM)的聚丙烯熔体流动速率(MFR)预报方法。首先构造HES-KDE-TVW集成数据处理方法,挖掘建模数据规律;然后构造自适应复合基函数,搭建互补的RBF神经网络预报模型;最后引入HMM对聚丙烯生产过程中的随机误差进行估计。经过工厂实际数据检验,模型在精度、泛化性及可靠性方面具有较好的综合性能。此种建模方法能为聚丙烯生产过程中牌号切换和质量控制提供一种备选的指导方案。     

群智能优化LSSVM最优聚丙烯熔融指数预报

提出了群智能优化AC_ICPSO(ant colony and immune clone particle swarm optimization)算法,融合蚁群算法与粒子群算法进行动态群体搜索,设计交叉算子和变异算子、群体多次编码、迭代选择等,来提高数据搜索的范围、精度和收敛的效率,避免早熟,降低算法的复杂度。然后利用AC_ICPSO方法对最小二乘支持向量机预报模型(LSSVM)进行参数寻优,得到最优的AC_ICPSO_LSSVM预报模型。以实际聚丙烯生产的熔融指数预报作为实例进行研究,结果表明所提出的AC_ICPSO_LSSVM方法有效,具有良好的预报精度。     

自适应粒子群优化算法在聚丙烯熔融指数预报上的应用

针对丙烯聚合生产控制中聚丙烯熔融指数在线测量的控制要求,以及过程变量间相关性高的特点,提出一种基于自适应粒子群优化算法和径向基函数神经网络的聚丙烯熔融指数预报新方法。该方法采用变参数的自适应粒子群优化算法提高优化算法的效率和收敛性,并且融合了主成分分析、统计建模以及智能优化方法,从而降低了预报模型的复杂度。提出了一种基于径向基函数神经网络的统计预报模型的参数优化和结构优化方法。使用该统计模型对工厂实际生产过程进行预报,并与国内外相关研究报道相比较,表明了本文所提出的预报方法的有效性和更高的准确性。     

AUTOMATICALLY CONFIGURING RADIAL BASIS FUNCTION NEURAL NETWORKS FOR NONLINEAR INTERNAL MODEL CONTROL

A nonlinear internal model control (NIMC) strategy based on automatically configuring radial basis function networks (RBFN) is proposed for single-input single-output (SISO) systems of relative degree greater than unity. The automatic configuration and training of the RBFN is carried out employing hierarchically-self-organizing-learning algorithm, which eliminates a predefined network structure, with closed-loop input-output data generated for a series of setpoint changes using PI controller. Simulation studies with automatically configuring RBFN for isothermal polymerization reactor control demonstrate the superior performance of the proposed control strategy with automatically configuring RBFN over PI control for setpoint tracking as well as disturbance rejection.     

基于混合粒子群优化的异类多模型软测量方法研究与应用

基于异类组合预测模型可提高模型的预测精度及鲁棒性的思想,提出一种基于混合粒子群优化的异类多模型非线性组合软测量建模的新方法。即先分别用混合粒子群优化的径向基函数神经网络、最小二乘支持向量机及部分最小二乘算法对训练集训练得出子模型,然后将具有性能互补性的三个子模型的输出作为反向传播网络的输入得到最后结果。用混合粒子群优化的方法来选取径向基函数神经网络和最小二乘支持向量机的模型参数,该方法克服了常用的交叉验证法耗时与盲目性问题。三层反向传播网络具有无限逼近特性,使得整个组合预测模型具有更好的泛化能力和预报精度。将其应用于汽油调合系统中研究法辛烷值的预测,仿真结果表明,该方法是可行且有效的。     

改进的全息搜索策略及其在化工优化中的应用

引言 基于样本数据估计反应动力学参数是常见的化工优化问题.参数估计的通用规则是偏差最小化,许多经典的序贯类寻优方法都可用于这一目的,例如Powell共轭梯度法、模式搜索法、变度量法等.     

Dynamic recurrent radial basis function network model predictive control of unstable nonlinear processes

A multistep model predictive control (MPC) strategy based on dynamically recurrent radial basis function networks (RBFNs) is proposed for single-input single-output (SISO) control of uncertain nonlinear processes. The control system consists of two automatically configured RBFNs, a trained network representing the plant model and a network with on-line learning to function as controller. The automatic configuration and learning of the networks is carried out by using a hierarchically self-organizing learning algorithm. This control strategy is structurally simple and computationally efficient since a single output node of each RBFN is configured to provide multistep predictions for plant output and controller. The performance of the proposed RBFNMPC strategy is evaluated by applying to two unstable nonlinear chemical processes, a chemical reactor and a biochemical reactor, and also a stable polymerization reactor. Further, the results of the RBFNMPC is compared with similar RBFN model based control strategies and also with well tuned PID/PI controller. The results show the better performance of the proposed RBFNMPC for the control of open-loop unstable nonlinear processes that exhibit multiple steady-state behavior.     

Modeling of the mixing of monodisperse particles using a stationary DEM-based Markov process

This paper provides a discussion on the modeling of granular mixing using Markov chain theory. Previous papers on this topic are either based on restrictive underlying assumptions about the flow structure or are limited to a small number of states. In this paper, a generalized approach for the construction of a multidimensional state space first-order Markov chain that represents the mixing of monodisperse particles is introduced. The transition probability matrix is computed directly using results obtained from a discrete element model. This work shows that, if accurate measurements of the state of the system are available, the associated Markov operator leads to a good estimate of the particle dynamics in the system.     

Modified nonlinear generalized ridge regression and its application to develop naphtha cut point soft sensor

A novel modified nonlinear generalized ridge regression (MNGRR) is proposed to model highly nonlinear system. MNGRR applies nonlinear transformation for independent variables to expand independent variable space. Then, the generalized ridge regression (GRR), which employs a modified differential evolution (MDE) to obtain the optimal ridge parameters according to the predicting error, is applied to remove the multicollinearity among the expanded variables, and thus the model that can describe complex nonlinear system and has good predicting ability is obtained. In practice, MNGRR is applied to develop naphtha 95% cut point soft sensor due to the existence of highly nonlinear relationship between process variables and naphtha 95% cut point in atmosphere distillation unit and the fact that few on-line hardware sensors are available and these are also difficult to maintain. Satisfactory results were obtained. The comparison results show that the performance of MNGRR is better than line regressions, nonlinear ordinary least squares regression and nonlinear traditional ridge regression. Further, MDE uses an adaptive mutation operator to overcome the premature and enhance the probability of obtaining the global optimal solution. The comparison results demonstrated that MDEs on-line and off-line performances are all superior to those of traditional DE (TDE), the probability of obtaining the global optimal solution is larger than that of TDE, and that the parameter sensitivity degree of MDE is lower than that of TDE.     

A CONTRIBUTION TO THE MATHEMATICAL MODELING OF BUBBLY/SLUG FLOW REGIME TRANSITION

A new mathematical modeling approach has been applied to the analysis of bubbly vapor/liquid flows. In particular, an integro-differential equation has been formulated which describes the bubble size distribution function. Various moments of this equation yield important two-phase flow parameters, such as the bubble number density, the mean bubble radius, and the interfacial area density. The steady-state distribution function has been numerically evaluated and an approximate analytical solution has been constructed. It was found that the model appears to be inherently capable of predicting the bubble to slug flow regime transition.