基于支持向量机的非线性预测控制技术

探讨了利用支持向量机进行非线性系统辨识的方法,并将支持向量机模型应用到非线性预测控制,提出了基于支持向量机模型的非线性预测控制算法.对一个CSTR反应器的仿真表明,支持向量机在小样本情况下具有良好的非线性建模能力和泛化能力.基于支持向量机的预测控制具有很好的控制性能,为通用非线性控制提供了一种新的控制思路.􀁽     

基于支持向量机的激光焊接过程辨识与控制

激光焊接过程数学模型足一个较强非线性的数学模型,通常的线性辨识方法无法得到它精确的数学模型.支持向量机作为一种新的机器学习方法,具有较强的非线性拟合能力,应用支持向量机非线性系统回归建模方法,辨识出具有典型非线性特性的焊接过程模型,并采用预测控制算法对焊接过程进行控制.实验证明,支持向量机对非线性系统具有很好的拟合效果,基于支持向量机的预测控制具有较好的非线性控制效果.     

基于小波支持向量机的木材干燥控制技术研究

针对木材干燥系统强耦合非线性的特性,提出了一种基于小波最小二乘支持向量机的预测控制方法.讨论了利用小波支持向量机对木材干燥系统进行系统识别的方法,并将辨识模型应用于预测控制算法,实现了木材干燥的自适应控制.仿真结果表明,基于小波支持向量机的预测控制技术具有较好的鲁棒性,对木材干燥系统有很好的实用性.     

基于SVM逆系统的非线性系统广义预测控制

针对工业过程中普遍存在的非线性被控对象,提出了一种基于支持向量机（SVM）逆系统的广义预测控制算法。该方法根据广义预测控制基于预测模型的特点,将基于支持向量机系统辨识的方法应用于逆系统构建和广义预测控制。该方法利用SVM强大的非线性映射能力离线辨识被控非线性系统的α阶逆模型,并将辨识出的逆模型连接在原被控统之前形成一个α阶纯延时伪线性系统。然后采用广义预测控制（GPC）算法实现对构造出的伪线性系统的预测控制。仿真实验表明了该算法的有效性和优越性。     

基于支持向量机的非线性离散动力学系统控制

提出了一种用支持向量机辨识系统状态空间模型的非线性离散动力学系统控制新方法. 在本方法中, 采用最小二乘支持向量机在每一个工作点辨识非线性系统的局部最优线性化模型. 针对该模型, 采用常规的线性控制方法在每个工作点设计局部线性控制器, 并在整个控制任务的每个工作点重复此设计过程.用该方法对两个典型的非线性离散系统采用极点配置技术进行了仿真验证, 结果显示系统对参考输入具有满意的跟踪性能, 证明该方法是有效和可行的.     

工业过程的子空间模型辨识

子空间模型辨识方法(SMI)是一类新兴的直接估计线性状态空间模型的黑箱建模方法,近年来获得了广泛关注.和传统的线性建模方法相比,SMI的优势不仅在于算法本身的简单可靠,也在于它的状态空间表达.本文首先简要介绍了SMI的基本思想以及3种基本算法(N4SID,MOESP,CVA).然后将这类方法应用于一个实际的工业过程建模,同时对3种SMI基本算法和一种传统辨识算法—预测误差方法(PEM)进行了研究对比.     

一类非线性逆系统的加权最小二乘支持向量机辨识方法

文中依据T-S模型的思想,提出了一种加权最小二乘支持向量机辨识算法.它采用模糊c均值(FCM)聚类确定规则数目,通过Gauss型函数将原输入输出空间分成若干子空间,在子空间中使用最小二乘支持向量机(LS-SVM)拟合获得子模型,然后由一个权重机制合成这些子模型,得到系统的模型.文中使用该方法去辨识关键反馈变量难以获得的非线性逆系统.为了得到这类逆系统的有效建模数据,采用了联合逆系统方法.仿真结果表明,加权最小二乘支持向量机辨识方法是有效的,它能够实现这类非线性逆系统的辨识,而且拟合误差平稳,波动幅度小,拟合精度和泛化能力都较好.     

基于支持向量机的一类非线性系统预测控制

针对一类具有输出反馈耦合的离散非线性系统,将过程的非线性部分通过支持向量机转化为全局线性状态空间模型,并在目标函数中引入系统状态的变化,给出一种类似于离散PI最优调节器的新型预测控制器.该方法不需要在线辨识系统参数,因为系统的内模已转换成全局离线模型.由于引入了新的优化目标函数,该控制器的控制效果和鲁棒性优于仅考虑预测输出误差的传统预测控制器.仿真结果表明,它也优于经典离散PI最优调节器.     

基于相关向量机的非线性动态系统辨识

基于具有核函数不用满足Mercer条件、相关向鼍自动确定及核函数少特点的稀疏贝叶斯的相关向量机核学习方法,提出了平滑先验条件约束的相关向量机的学习方法,采用稀疏贝叶斯模型的最大边缘似然算法加快了求解相关向量机的向量,并采取交叉验证法确定其核参数提高了相关向量机辨识的泛化性.该方法避免了支持向量机的非线性系统辨识的模型结构难于确定的问题,与支持向量机辨识方法相比较,辨识的模型结构更简洁.仿真表明,该方法应用于非线性动态系统的辨识,具有良好的效果.     

基于L S-SVM 的非线性预测控制技术

探讨了利用最小二乘支持向量机(LS—SVM)进行非线性系统辨识的方法，LS—SVM用等式约束代替传统支持向量机中不等式约束，求解过程从解QP问题变成解一组等式方程，将得到的LS—SVM模型应用到非线性预测控制，提出了基于LS—SVM模型的非线性预测控制算法，通过CSTR过程仿真表明，最小二乘支持向量机学习速度快，在小样本情况下具有良好的非线性建模和泛化能力，基于LS—SVM的预测控制算法具有很好的控制性能。     

Unscented Kalman filter based nonlinear model predictive control of a LDPE autoclave reactor

This paper presents a multivariable nonlinear model predictive control (NMPC) scheme for the regulation of a low-density polyethylene (LDPE) autoclave reactor. A detailed mechanistic process model developed previously was used to describe the dynamics of the LDPE reactor and the properties of the polymer product. Closed-loop simulations are used to demonstrate the disturbance rejection and tracking performance of the NMPC algorithm for control of reactor temperature and weight-averaged molecular weight (WAMW). In addition, the effect of parametric uncertainty in the kinetic rate constants of the LDPE reactor model on closed-loop performance is discussed. The unscented Kalman filtering (UKF) algorithm is employed to estimate plant states and disturbances. All control simulations were performed under conditions of noisy process measurements and structural plant–model mismatch. Where appropriate, the performance of the NMPC algorithm is contrasted with that of linear model predictive control (LMPC). It is shown that for this application the closed-loop performance of the UKF based NMPC scheme is very good and is superior to that of the linear predictive controller.     

A hybrid model predictive control strategy for nonlinear plant-wide control

A plant-wide control strategy based on integrating linear model predictive control (LMPC) and nonlinear model predictive control (NMPC) is proposed. The hybrid method is applicable to plants that can be decomposed into approximately linear subsystems and highly nonlinear subsystems that interact via mass and energy flows. LMPC is applied to the linear subsystems and NMPC is applied to the nonlinear subsystems. A simple controller coordination strategy that counteracts interaction effects is proposed for the case of one linear subsystem and one nonlinear subsystem. A reactor/separator process with recycle is used to compare the hybrid method to conventional LMPC and NMPC techniques.     

基于FPGA的非线性预测控制器设计

对于非线性程度较高的复杂对象,非线性模型预测控制（NonlinearModelPredictiveControl,NMPC）是一种有效的控制策略。为了实现对这类对象的有效控制,设计了一种基于FPGA（FieldProgrammableGateArray）的非线性预测控制器,该嵌入式控制器具有灵活性和高适应性等特点,能够应用于工业现场控制。为了满足工业控制的可行性和实时性要求,提出了一种序贯二次规划（SQP）算法的改进算法,在FPGA有限的计算资源下,保证每个采样间隔内都能得到NMPC优化问题的可行解。经仿真实验证明,采用非线性预测控制器在计算速度和精度上都能达到较好的性能。     

Experimental application of nonlinear model predictive control: temperature control of an industrial semi-batch pilot-plant reactor

This paper describes the application of nonlinear model predictive control (NMPC) to the temperature control of a semi-batch chemical reactor equipped with a multi-fluid heating/cooling system. The strategy of the nonlinear control system is based on a constrained optimisation problem, which is solved repeatedly on-line by a step-wise integration of a nonlinear dynamic model and optimisation strategy. A supervisory control routine has been developed, based on the same nonlinear dynamic model, to handle automatically the fluid changeovers. Both NMPC and supervisory control have been implemented on a PC and applied to a 16 l batch reactor pilot plant. Experiments illustrate the feasibility of such a procedure involving predictive control and supervisory control.     

A note on stability, robustness and performance of output feedback nonlinear model predictive control

In recent years, nonlinear model predictive control (NMPC) schemes have been derived that guarantee stability of the closed loop under the assumption of full state information. However, only limited advances have been made with respect to output feedback in the framework of nonlinear predictive control. This paper combines stabilizing instantaneous state feedback NMPC schemes with high-gain observers to achieve output feedback stabilization. For a uniformly observable MIMO system class it is shown that the resulting closed loop is asymptotically stable. Furthermore, the output feedback NMPC scheme recovers the performance of the state feedback in the sense that the region of attraction and the trajectories of the state feedback scheme can be recovered to any degree of accuracy for large enough observer gains, thus leading to semi-regional results. Additionally, it is shown that the output feedback controller is robust with respect to static sector bounded nonlinear input uncertainties.     

基于在线支持向量回归的非线性模型预测控制方法

针对非线性模型预测控制中离线模型难以适应非线性对象实时变化的缺点, 提出一种基于在线支持向量回归的非线性模型预测控制方法. 该方法通过在线支持向量回归离线训练与在线学习相结合的方式, 建立具有在线校正特性的预测模型, 同时采用最速下降原理滚动优化非线性模型预测控制的目标函数, 求得多步控制量. 通过对非线性对象的控制结果表明, 所提出方法有效且具有良好的自适应性.

     

Identification and predictive control of a multistage evaporator

A recurrent neural network-based nonlinear model predictive control (NMPC) scheme in parallel with PI control loops is developed for a simulation model of an industrial-scale five-stage evaporator. Input–output data from system identification experiments are used in training the network using the Levenberg–Marquardt algorithm with automatic differentiation. The same optimization algorithm is used in predictive control of the plant. The scheme is tested with set-point tracking and disturbance rejection problems on the plant while control performance is compared with that of PI controllers, a simplified mechanistic model-based NMPC developed in previous work and a linear model predictive controller (LMPC). Results show significant improvements in control performance by the new parallel NMPC–PI control scheme.     

一类非线性系统非切换解析模型预测控制方法研究

针对关系度不确定非线性系统, 基于模型预测控制理论和切换解析非线性模型预测控制(Nonlinear model predictive control, NMPC) 提出了一种非切换的解析NMPC新方法. 论证了在非切换解析NMPC控制律下, 通过坐标变换可以将闭环系统分别在关系度确定和不确定的两个子空间近似为线性系统, 得出非切换解析NMPC使闭环系统稳定的必要条件. 通过仿真实验验证了非切换解析NMPC可以达到很好的响应特性, 无需切换的特征也扩大了其应用范围.     

Nonlinear model predictive control of fed-batch fermentations using dynamic flux balance models

Fed-batch fermentation is an important production technology in the biochemical industry. Using fed-batch Saccharomyces cerevisiae fermentation as a prototypical example, we developed a general methodology for nonlinear model predictive control of fed-batch bioreactors described by dynamic flux balance models. The control objective was to maximize ethanol production at a fixed final batch time by adjusting the glucose feeding rate and the aerobic–anaerobic switching time. Effectiveness of the closed-loop implementation was evaluated by comparing the relative performance of NMPC and the open-loop optimal controller. NMPC was able to compensate for structural errors in the intracellular model and parametric errors in the substrate uptake kinetics and cellular energetics by increasing ethanol production between 8.0% and 14.7% compared with the open-loop operating policy. Minimal degradation in NMPC performance was observed when the biomass, glucose, and ethanol concentration and liquid volume measurements were corrupted with Gaussian white noise. NMPC based on the dynamic flux balance model was shown to improve ethanol production compared to the same NMPC formulation based on a simpler unstructured model. To our knowledge, this study represents the first attempt to utilize a dynamic flux balance model within a nonlinear model-based control scheme.     

A formulation of nonlinear model predictive control using automatic differentiation

An efficient algorithm is developed to alleviate the computational burden associated with nonlinear model predictive control (NMPC). The new algorithm extends an existing algorithm for solutions of dynamic sensitivity from autonomous to non-autonomous differential equations using the Taylor series and automatic differentiation (AD). A formulation is then presented to recast the NMPC problem as a standard nonlinear programming problem by using the Taylor series and AD. The efficiency of the new algorithm is compared with other approaches via an evaporation case study. The comparison shows that the new algorithm can reduce computational time by two orders of magnitude.