基于动态PLS框架的多回路内模控制器设计(英文)

In this paper, a multi-loop internal model control (IMC) scheme in conjunction with feed-forward strategy based on the dynamic partial least squares (DyPLS) framework is proposed. Unlike the traditional methods to decouple multi-input multi-output (MIMO) systems, the DyPLS framework automatically decomposes the MIMO process into a multi-loop system in the PLS subspace in the modeling stage. The dynamic filters with identical structure are used to build the dynamic PLS model, which retains the or-thogonality among the latent variables. To address the model mismatch problem, an off-line least squares method is applied to obtain a set of optimal filter parameters in each latent space. Without losing the merits of model-based control, a simple and easy-tuned IMC structure is readily carried over to the dynamic PLS control framework. In addition, by projecting the measurable disturbance into the latent subspace, a multi-loop feed-forward control is yielded to achieve better performance for disturbance rejection. Simulation re-sults of a distillation column are used to further demonstrate this new strategy outperforms conventional control schemes in servo behavior and disturbance rejection.     

Adaptive soft sensor modeling framework based on just-in-time learning and kernel partial least squares regression for nonlinear multiphase batch processes

Batch processes are characterized by inherent nonlinearity, multiple phases and time-varying behavior that pose great challenges for accurate state estimation. A multiphase just-in-time (MJIT) learning based kernel partial least squares (KPLS) method is proposed for multiphase batch processes. Gaussian mixture model is estimated to identify different operating phases where various JIT-KPLS frameworks are built. By applying Bayesian inference strategy, the query data is classified into a particular phase with the maximal posterior probability, and thus the corresponding JIT-KPLS framework is chosen for online prediction. To further improve the predictive accuracy of the MJIT-KPLS algorithm, a hybrid similarity measure and an adaptive selection strategy are proposed for selecting local modeling samples. Moreover, maximal similarity replacement rule is proposed to update database. A procedure of input variable selection based on partial mutual information is also presented. The effectiveness of the MJIT-KPLS algorithm is demonstrated through application to industrial fed-batch chlortetracycline fermentation process.     

基于高阶偏最小二乘的间歇过程建模

间歇过程的产品与现代人的生活息息相关,而建立可靠的模型是保障间歇过程安全运行的基础。针对间歇过程的数据特点,引入一种新的广义线性回归模型--高阶偏最小二乘（higher order partial least squares,HOPLS）。它与传统的间歇过程建模方法具有本质的不同,三维数据（批次×变量×时间）不需要展开成二维矩阵,而是直接被分解成一组正交的Tucker矩阵之和。通过高阶奇异值分解（high order singular value decomposition,HOSVD）,张量变换和高阶正交迭代（higher order orthogonal iteration,HOOI）找到能同时包含自变量和因变量最大信息的潜向量,与此同时得到对应的负载向量。对于新观测值,通过模型就可以实现对因变量的预测。最后利用PenSim2.0,对青霉素发酵过程进行仿真研究,验证了该间歇过程建模方法的有效性。     

Generalized generic model control of high‐purity internal thermally coupled distillation column based on nonlinear wave theory

A new model‐based control strategy for the internal thermally coupled distillation column (ITCDIC) is presented. Based on the nonlinear wave theory that describes the nonlinear dynamics in the separation processes, a simplified nonlinear wave model is established that concerns both the wave propagation and the profile shape. An advanced controller (WGGMC) is formulated by combining the nonlinear wave model with a generalized generic model control (GGMC). Compared with a conventional generic model controller based on a data‐driven model (TGMC), and another wave‐model based generic model controller (WGMC) developed in our previous work, WGGMC exhibits the best performances in both servo control and regulatory control. Furthermore, WGGMC can handle a very‐high‐purity system of ITCDIC with top product composition of 0.99999, while the other two controllers fail to work. © 2013 American Institute of Chemical Engineers AIChE J, 59: 4133–4141, 2013     

基于时间差分和局部加权偏最小二乘算法的过程自适应软测量建模

工业过程软测量模型常常因为过程的变量漂移、非线性和时变等问题而使得预测性能下降。因此,时间差分已被应用于解决过程变量漂移问题。但是,时间差分框架下的全局模型往往不能很好地描述过程非线性和时变等特性。为此,提出了一种融合时间差分模型和局部加权偏最小二乘算法的自适应软测量建模方法。时间差分模型可以大大减少过程变量漂移的影响,而局部加权偏最小二乘算法作为一种即时学习方法,可以有效解决过程非线性和时变问题。该方法的有效性在数值例子和工业过程实例中得到了有效验证。     

Development of a new soft sensor method using independent component analysis and partial least squares

Soft sensors are used widely to estimate a process variable which is difficult to measure online. One of the crucial difficulties of soft sensors is that predictive accuracy drops due to changes of state of chemical plants. To cope with this problem, a regression model can be updated. However, if the model is updated with an abnormal sample, the predictive ability can deteriorate. We have applied the independent component analysis (ICA) method to the soft sensor to increase fault detection ability. Then, we have tried to increase the predictive accuracy. By using the ICA‐based fault detection and classification model, the objective variable can be predicted, updating the PLS model appropriately. We analyzed real industrial data as the application of the proposed method. The proposed method achieved higher predictive accuracy than the traditional one. Furthermore, the nonsteady state could be detected as abnormal correctly by the ICA model. © 2008 American Institute of Chemical Engineers AIChE J, 2009     

基于最小二乘支持向量机的蜡沉积速率预测

考虑蜡沉积影响因素的复杂性和最小二乘支持向量机在小样本预测方面的优势,基于最小二乘支持向量机预测的原理,通过优化最小二乘支持向量机的参数,建立了蜡沉积速率的预测模型,并对蜡沉积速率进行了预测。结果表明：该方法在样本数量较小时仍具有较高的精度,蜡沉积速率的预测值和实验值的吻合程度较好;最小二乘支持向量机建模时可以得到直观的函数表达式,而神经网络方法却不能得到模型的显式表达式,因此该方法具有明显的优势;应用径向基核（RBF）作为核函数时,不同初值的正则化参数?和核函数宽度?对预测结果具有较大影响,使用时应合理选择。     

Estimation of effluent quality index based on partial least squares stochastic configuration networks

Accurate and reliable measurement of the effluent quality indicators of wastewater treatment plants is the key to successful control and optimization of wastewater treatment plants. Due to the complexity of the operation and the delay of laboratory analysis, it is difficult to achieve real-time control of effluent quality. In order to improve the accuracy and reliability of the estimation, this paper proposes a method of stochastic configuration network based on partial least squares (PLS-SCN). In order to overcome the forecast risk caused by high dimensionality and multicollinearity of the input data, the partial least squares(PLS) is embedded into the stochastic configuration network(SCN) framework replacing the classic ordinary least squares (OLS). The PLS-SCN method extracts the main latent variables that affect the effluent quality from the output of the hidden layer, and enhances the generalization performance through orthogonal projection operations. The simulation results of the effluent quality index of a municipal sewage treatment plant show that the PLS-SCN network has a good input and output relationship, and its performance is better than traditional SCN and PLS, and it can quickly and reliably estimate the sewage quality.     

基于偏最小二乘随机配置网络的污水水质指标估计

准确、可靠地测量污水处理厂的出水水质指标是成功控制和优化污水处理厂的关键。由于现有的离线化验方法存在操作繁复、测量滞后的问题,难以实现水质的实时控制。为了提高估计的准确性和可靠性,提出了一种偏最小二乘的随机配置网络方法 (PLS-SCN)。为了克服输入数据高维度和多重共线性导致的预测风险,将偏最小二乘(PLS)方法嵌入到随机配置网络(SCN)框架中,以代替经典的普通最小二乘(OLS)方法。PLSSCN方法从隐含层输出中提取影响水质指标的主要潜在变量,通过正交投影运算来增强泛化性能。某城市污水处理厂水质指标仿真结果表明,PLS-SCN网络具有良好的输入输出关系,性能优于传统SCN和PLS方法,能够快速、可靠地估计污水水质的质量。     

Ensemble locally weighted partial least squares as a just‐in‐time modeling method

The predictive ability of soft sensors, which estimate values of an objective variable y online, decreases due to process changes in chemical plants. To reduce the decrease of predictive ability, adaptive soft sensors have been developed. We focused on just‐in‐time soft sensors, especially locally weighted partial least squares (LWPLS) regression. Since a set of hyperparameters in an LWPLS model has to be set beforehand and there is only onedataset, a traditional LWPLS model is difficult to accurately predict y‐values in multiple process states. In this study, we propose to combine LWPLS and ensemble learning, and predict y‐values with multiple LWPLS models, whose datasets and sets of hyperparameters are different. The weights of LWPLS models are determined based on Bayes’ theorem, considering their predictive ability. We confirmed that the proposed model has higher predictive accuracy than traditional models through numerical simulation data and two industrial data analyses. © 2015 American Institute of Chemical Engineers AIChE J, 62: 717–725, 2016     

Interpreting the dynamic effect of internal heat integration on reactive distillation columns

In this work,the impact of internal heat integration upon process dynamics and controllability by superposing reactive section onto stripping section,relocating feed locations,and redistributing catalyst within the reactive section is explored based on a hypothetical ideal reactive distillation system containing an exothermic reaction:A + B ←→ C + D.Steady state operation analysis and closed-loop controllability evaluation are carried out by comparing the process designs with and without the consideration of internal heat integration,For superposing reactive section onto stripping section,favorable effect is aroused due to its low sensitivities to the changes in operating condition,For ascending the lower feed stage,somewhat detrimental effect occurs because of the accompanied adverse internal heat integration and strong sensitivity to the changes in operating condition.For descending the upper feed stage,serious detrimental effect happens because of the introduced adverse internal heat integration and strong sensitivity to the changes in operating condition.For redistributing catalyst in the reactive section,fairly small negative influence is aroused by the sensitivity to the changes in operating condition.When reinforcing internal heat integration with a combinatorial use of these three strategies,the decent of the upper feed stage should be avoided in process development.Although the conclusions are derived based on the hypothetical ideal reactive distillation column studied,they are considered to be of general significance to the design and operation of other reactive distillation columns.     

基于实验设计与偏最小二乘逆模型的产品设计方法

现代市场中,同类产品规格多样、更新换代速度加快,快速经济地设定新产品操作条件成为生产过程的迫切需求。本文首先改进实验设计(DOE)优化方法中的响应曲面法(RSM),可随着新实验数据的累积在线更新搜索算法的外延步长;然后提出RSM与偏最小二乘(PLS)逆模型的结合方法,可保证模型具有良好的内插和外延性能,显著减少实验次数、快速找到满足新产品要求的操作条件;最后通过数值仿真以及某化工过程的应用仿真验证了方法的可行性和有效性。     

一类化工过程多变量系统的自适应非线性预测控制

针对化工过程的一类多变量非线性系统,提出了一种自适应非线性预测控制(ANMPC)算法。在采用递归最小二乘法进行预测模型参数在线辨识的基础上,将系统的静态非线性关系用一个反向传播(BP)神经网络稳态模型来表示,通过稳态模型求得的动态增益来进一步校正预测模型的参数。详述了ANMPC控制器设计步骤,通过在一个多变量pH中和过程中的仿真验证了本算法的可行性和有效性。     

Low-gain internal model control PID controller design based on second-order filter

A design method is proposed for low-gain internal model control (IMC) proportional-integral-derivative (PID) controllers based on the second-order filter. The PID parameters are obtained by approximating the feedback form of the IMC controller with a Maclaurin series, in which the second-order filter is applied using the IMC approach to achieve a low-gain PID controller that is suitable for model mismatch problems. Analytical PID tuning rules based on the second-order filter are derived for several common-use process models. The second-order filter is designed from the desired time domain performances of maximum overshoot and settling time. Furthermore, the robustness of the IMC PID controller based on the second-order filter is analyzed, and results show that its robustness performance is better than the first-order filter under certain conditions. Finally, three categories of models divided by the ration of time constant and time delay are presented in the comparative numerical simulations to validate the effectiveness and generality of the proposed PID controller design method.     

Parametric and nonparametric model based control of a packed distillation column

Parametric and nonparametric model based control systems were applied to control the overhead temperature of a packed distillation column separating methanol–water mixture. Experimental and theoretical studies have been done to observe the efficiency and performance of both control systems. Generalized predictive control (GPC) system based on a parametric model has been tried to keep the overhead temperature at the desired set point. First, a parametric model which is controlled auto regressive integrated moving average (CARIMA) was developed and then the parameters of this model were identified by applying pseudo random binary sequence (PRBS) and using Bierman algorithm. After that this model was used to design the GPC system. Tuning parameters of the GPC system have been calculated using the simulation program of the packed distillation column. Using the predicted parameters, experimental and theoretical GPC systems were found very effective in controlling the overhead temperature. Dynamic matrix control (DMC) system based on a nonparametric model has been used to track the overhead temperature of the packed distillation column. For this purpose, a nonparametric model known as the dynamic matrix was determined using the reaction curve method. A step change in heat input to the reboiler was applied to the manipulated variable and the temperature of the overhead product was observed. After that, the dynamic matrix was used to design the DMC system. Several calculations have been done to define the DMC control parameters. The best values of the tuning parameter were used to realize the DMC system for controlling the overhead temperature experimentally and theoretically. In the presence of some disturbances, the DMC system gives oscillation and offset in experimental studies.     

基于自适应模糊神经网络的非线性系统模型预测控制

针对非线性动态系统的控制问题,提出了一种基于自适应模糊神经网络(adaptive fuzzy neural network,AFNN)的模型预测控制(model predictive control, MPC)方法。首先,在离线建模阶段,AFNN采用规则自分裂技术产生初始模糊规则,采用改进的自适应LM学习算法优化网络参数;然后,在实时控制过程,AFNN根据系统输出和预测输出之间的误差调整网络参数,从而为MPC提供一个精确的预测模型;进一步,AFNN-MPC利用带有自适应学习率的梯度下降寻优算法求解优化问题,在线获取非线性控制量,并将其作用到动态系统实施控制。此外,给出了AFNN-MPC的收敛性和稳定性证明,以保证其在实际工程中的成功应用。最后,利用数值仿真和双CSTR过程进行实验验证。结果表明,AFNN-MPC能够取得优越的控制性能。     

基于T-S模糊模型的间歇过程的迭代学习容错控制

间歇过程不仅具有强非线性,同时还会受到诸如执行器等故障影响,研究非线性间歇过程在具有故障的情况下依然稳定运行至关重要。针对执行器增益故障及系统所具有的强非线性,提出一种新的基于间歇过程的T-S模糊模型的复合迭代学习容错控制方法。首先根据间歇过程的非线性模型,利用扇区非线性方法建立其T-S模糊故障模型,再利用间歇过程的二维特性与重复特性,在2D系统理论框架内,设计2D复合ILC容错控制器,进而构建此T-S模糊模型的等价二维Rosser模型,接着利用Lyapunov方法给出系统稳定充分条件并求解控制器增益。针对强非线性的连续搅拌釜进行仿真,结果表明所提出方法具有可行性与有效性。     

基于T-S模糊模型与粒子群优化的非线性预测控制

引言模型预测控制属于一种基于模型的多变量的控制算法,发展至今已在化工过程控制方面得到了广泛的应用[1-5]。状态反馈预测控制[6-8]是模型预测控制技术的一种,基于状态空间模型,采用实测状态     

基于一般模型控制的高纯内部热耦合精馏策略

内部热耦合精馏塔（ITCDIC）是精馏节能控制的一个前沿。本文提出了一种基于一般模型控制（GMC）的内部热耦合精馏塔的先控策略,以解决导致传统线性控制策略难以得到较好控制效果的高纯下内部热耦合精馏塔的非线性。以苯-甲苯物系作为研究实例,对所提出的高纯ITCDIC控制策略进行了详细研究。设定值改变和过程扰动下的控制品质表明了所提出的高纯ITCDIC的GMC控制策略的切实有效性。