Nonlinear model predictive control for distributed parameter systems by time–space-coupled model reduction

Nonlinear high-dimensional distributed parameter systems (DPSs) described by sets of parabolic partial different equations (PDEs) exhibit a dominant, low-dimensional slow behavior that can be captured using model reduction. A time–space-coupled model reduction architecture combining encoder–decoder networks with recurrent neural networks (RNNs) was presented in our previous work, for modeling the spatiotemporal dynamics of DPSs without recourse to the governing equations. In this work, we further understand the stability of the training dynamics of the deep architecture by using the Lyapunov exponents (LEs). Subsequently, we construct nonlinear model predictive control (MPC) formulations for the DPS based on the learned, dimensional-reduced model. We use a path-integral optimal control algorithm for MPC implementation to avoid any analytic derivatives of the dynamics. The effectiveness of integration of a deep neural network-based model with MPC is demonstrated in a tubular reactor with recycle cases. The results of the simulation also show that the LE can serve as a readout of training stability for the learned dynamical model.     

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。     

A Comparative Study of Combined Feedforward/Feedback Model Predictive Control for Nonlinear Systems

Model predictive control (MPC) provides a natural framework to realize feedforward and feedback control for nonlinear systems where the effect of disturbances (DVs) cannot be separated from that of manipulated variables (MVs). This study examines the performance of MPC with measured DVs as partial inputs of the model used, which is termed as combined feedforward/feedback MPC (CMPC) in contrast to conventional MPC using a model without input of any measured DV. In the simulation of a pH process, we demonstrate the clear superiority of CMPC over MPC. In the experiment with a bench‐scale ethanol and water distillation column, CMPC and MPC using artificial neural network (ANN) models are applied to the dual temperature control problem. External recurrent neural networks (ERNs) with and without a measured DV (feed rate of the column) as their partial input are built and employed in the experiment, with a result that inclusion of the measured DV in the model makes CMPC perform significantly better than MPC. To strengthen practical experience in applying ANN‐based MPC, a detailed procedure of the experiment is also documented.     

Online machine learning modeling and predictive control of nonlinear systems with scheduled mode transitions

This work develops a model predictive control (MPC) scheme using online learning of recurrent neural network (RNN) models for nonlinear systems switched between multiple operating regions following a prescribed switching schedule. Specifically, an RNN model is initially developed offline to model process dynamics using the historical operational data collected in a small region around a certain steady-state. After the system is switched to another operating region under a Lyapunov-based MPC with suitable constraints to ensure satisfaction of the prescribed switching schedule policy, RNN models are updated using real-time process data to improve closed-loop performance. A generalization error bound is derived for the updated RNN models using the notion of regret, and closed-loop stability results are established for the switched nonlinear system under RNN-based MPC. Finally, a chemical process example with the operation schedule that requires switching between two steady-states is used to demonstrate the effectiveness of the proposed RNN-MPC scheme.     

Toward a low cost and high performance MPC: The role of system identification

The high cost of model predictive control (MPC) technology has hampered its wide application in process industries beyond the refining/petrochemical industry. This work strives to increase the efficiency of MPC deployment. First, a semi-automatic MPC system is introduced. It consists of three modules: an MPC module, an online identification module and a control monitor module. The goal of the MPC technology is twofold: (1) to considerably reduce the cost of MPC commissioning and maintenance and (2) to increase control performance. System identification plays important roles in all the three parts of the MPC system. In the identification module, the so-called ASYM method of identification is used. It is demonstrated with an industrial application. In the control module, adaptive disturbance model identification is developed for improving control performance; in the monitor module, a method of model error detection method is developed. Industrial applications and simulations are used to demonstrate the ideas. Finally, we comment on some industrial needs on MPC research and development.     

Machine learning-based predictive control of nonlinear processes. Part I: Theory

This article focuses on the design of model predictive control (MPC) systems for nonlinear processes that utilize an ensemble of recurrent neural network (RNN) models to predict nonlinear dynamics. Specifically, RNN models are initially developed based on a data set generated from extensive open-loop simulations within a desired process operation region to capture process dynamics with a sufficiently small modeling error between the RNN model and the actual nonlinear process model. Subsequently, Lyapunov-based MPC (LMPC) that utilizes RNN models as the prediction model is developed to achieve closed-loop state boundedness and convergence to the origin. Additionally, machine learning ensemble regression modeling tools are employed in the formulation of LMPC to improve prediction accuracy of RNN models and overall closed-loop performance while parallel computing is utilized to reduce computation time. Computational implementation of the method and application to a chemical reactor example is discussed in the second article of this series.     

Machine learning-based distributed model predictive control of nonlinear processes

This work explores the design of distributed model predictive control (DMPC) systems for nonlinear processes using machine learning models to predict nonlinear dynamic behavior. Specifically, sequential and iterative DMPC systems are designed and analyzed with respect to closed-loop stability and performance properties. Extensive open-loop data within a desired operating region are used to develop long short-term memory (LSTM) recurrent neural network models with a sufficiently small modeling error from the actual nonlinear process model. Subsequently, these LSTM models are utilized in Lyapunov-based DMPC to achieve efficient real-time computation time while ensuring closed-loop state boundedness and convergence to the origin. Using a nonlinear chemical process network example, the simulation results demonstrate the improved computational efficiency when the process is operated under sequential and iterative DMPCs while the closed-loop performance is very close to the one of a centralized MPC system.     

一种基于新型蚁群算法的聚丙烯熔融指数预报模型

聚丙烯熔融指数的实时预报非常重要却十分困难,提出了一种经过新型蚁群算法优化后的PCA-RBF神经网络方法进行熔融指数预报。PCA将原始数据从高维空间映射到低维空间,剔除冗余信息和提取过程特征;RBF神经网络则用来拟合输入与输出之间的非线性关系;最后用适用于连续空间寻优问题的新型蚁群算法对RBF神经网络权值进行优化。实际生产数据的研究结果,表明了所提出的熔融指数预报模型的准确性和可靠性。     

Model predictive control of condensate recycle process in a cogeneration power station: Controller design and numerical application

A model predictive control (MPC) system has been developed for application to the condensate recycle process of a 300 MW cogeneration power station of the East-West Power Plant, Gyeonggido, Korea. Unlike other industrial processes where MPC has been predominantly applied, the operation mode of the cogeneration power station changes continuously with weather and seasonal conditions. Such characteristic makes it difficult to find the process model for controller design through identification. To overcome the difficulty, process models for MPC design were derived for each operation mode from the material balance applied to the pipeline network around the concerned process. The MPC algorithm has been developed so that the controller tuning is easy with one tuning knob for each output and the constrained optimization is solved by an interior-point method. For verification of the MPC system before process implementation, a process simulator was also developed. Performance of the MPC was investigated first with a process simulator against various disturbance scenarios.     

Chemical process fault diagnosis based on a combined deep learning method

The study on fault detection and diagnosis (FDD) of chemical processes has always been the top priority of the chemical process safety. In this paper, a fault diagnosis method combining the deep convolutional with the recurrent neural network (DCRNN) is proposed. In this method, the data from chemical processes are input to the deep convolutional neural network (DCNN) to extract features in spatial domains, and then, the features are fused into the bidirectional recurrent neural network (BRNN). Due to the powerful capabilities of DCNN to extract features in spatial domains and the sensitivity to time series of RNN, the combined method can adaptively learn the dynamic information of the raw data in both spatial and temporal domains and has unique advantages in multivariate chemical processes. The application of the DCRNN model in the Tennessee Eastman (TE) process demonstrates the high accuracy of this proposal in identifying abnormal conditions for the chemical process, compared with the traditional fault identification algorithms of deep learning.     

Model-based control of a molten carbonate fuel cell (MCFC) process

To improve availability and performance of fuel cells, the operating temperature of molten carbonate fuel cells (MCFC) stack should be strictly maintained within a specified operation range, and an efficient control technique should be employed to meet this objective. While most modern control strategies are based on process models, many existing models of MCFC are not ready to be applied in synthesis and operation of control systems. In this study, we developed an auto-regressive moving average (ARMA) model and machine learning methods of least squares support vector machine (LS-SVM), artificial neural network (ANN) and partial least squares (PLS) for the MCFC system based on input-output operating data. The ARMA model showed the best tracking performance. A model predictive control method for the operation of MCFC system was developed based on the proposed ARMA model. The control performance of the proposed MPC methods was compared with that of conventional controllers using numerical simulations performed on various process models including an MCFC process. Numerical results show that ARMA model based control provides improved control performance compared to other control methods.     

基于径向基函数网络的MH/Ni电池建模及容量预测

引　言近年来 ,随着汽车的迅速发展和大量普及 ,它所造成的尾气污染问题也日益突出 .电动车的发展可以有效地解决燃油汽车的污染排放问题 .MH/Ni电池是一种无污染的“绿色能源” ,它具有高比能量、高比功率、长寿命及安全性好等特点 ,是电动车用动力型电池的首选 .在动力型电池     

A new algorithm for developing dynamic radial basis function neural network models based on genetic algorithms

A new method for extracting valuable process information from input–output data is presented in this paper. The proposed methodology produces dynamical radial basis function (RBF) neural network models based on a specially designed genetic algorithm (GA), which is used to auto-configure the structure of the networks and obtain the model parameters. The new RBF network training technique formulates a complete optimization problem, which includes the network structure into the set of free variables that are used to minimize the prediction error. This is a different approach compared with the local search methods employed by other structure selection mechanisms, which are often trapped to local minima. Another advantage of the proposed method is that only one run of the algorithm is required to obtain the optimal network structure, in contrast to the standard RBF training techniques, where the produced model is selected by trial and error. The effectiveness of the method is illustrated through the development of dynamical models for two sets of data: simulated data from a Continuous Stirred Tank Reactor (CSTR) and true data collected from a Kamyr digester, which is a rather complicated reactor used in the pulp and paper industry.     

基于剪接系统的遗传算法RBF网络建模方法

A splicing system based genetic algorithm is proposed to optimize dynamical radial basis function （RBF） neural network, which is used to extract valuable process information from input output data. The novel RBF network training technique includes the network structure into the set of function centers by compromising between the conflicting requirements of reducing prediction error and simultaneously decreasing model complexity. The effectiveness of the proposed method is illustrated through the development of dynamic models as a benchmark discrete example and a continuous stirred tank reactor by comparing with several different RBF network training methods.     

基于主成分分析的输出集成反馈网络及其在化工动态过程建模中的应用

针对主成分分析和反馈神经网络的优点，提出基于主成分分析的输出集成反馈网络建模方法，并对训练算法作了推导，在动态化工过程建模中取得满意的效果。     

基于模型预测控制及智能寻优的水泥粉磨优化控制

本文提出一种基于运行状态软测量和成本软约束的多变量模型预测控制(MPC)。MPC控制与传统的专家经验控制和模糊控制相比,通过模型对系统工艺参数的预测,不断地学习更新模型,更符合水泥粉磨大时延、多工况的特性。应用中通过对水泥粉磨装置的阶跃响应实验,建立多变量预测控制模型,解决水泥粉磨系统生产过程的不确定性。在此基础上,通过增量学习和机器学习找到最优运行参数,使水泥粉磨的MPC控制一直保持在最优工况。     

Monitoring and retuning of low-level PID control loops

In this work, we focus on the problem of monitoring and retuning of low-level proportional-integral-derivative (PID) control loops used to regulate control actuators to the values computed by advanced model-based control systems like model predictive control (MPC). We consider the case where the real-time measurement of the actuation level is unavailable, and thus PID controller monitoring has to be achieved on the basis of process state measurements. A fault detection and isolation (FDI) method involving process models and real-time process measurements is used to monitor the PID control loops and compute appropriate residuals. Once poor tuning is detected and isolated, a PID tuning method based on the estimated transfer function of the control actuator is applied to the isolated, poorly functioning PID controller. An example of a non-linear reactor–separator process operating under MPC with low-level PID controllers regulating the control actuators is used to demonstrate the approach.     

基于互信息和自组织RBF神经网络的出水BOD软测量方法

针对污水处理过程出水生化需氧量（biochemical oxygen demand，BOD）难以实时准确测量的问题，提出了一种基于互信息和自组织RBF神经网络的软测量方法对出水BOD进行预测。首先，使用基于互信息的方法提取相关特征参量作为软测量模型的输入变量；其次，设计一种基于误差校正-敏感度分析的自组织RBF神经网络，使用改进的Levenberg-Marquardt（LM）算法对网络进行训练以提高训练速度；最后将软测量模型应用于UCI公开数据集及实际的污水处理过程，实验结果表明该软测量模型结构紧凑，训练时间相对较短，预测精度有所提高，能够对出水BOD实现快速准确预测。     

基于稀疏局部嵌入深度卷积网络的冷水机组故障诊断方法

针对于冷水机组提出一种基于稀疏局部嵌入深度卷积网络（sparsely local embedding network,SLENet）的故障诊断方法。采用稀疏局部嵌入方法代替卷积核,对输入数据进行特征选择,避免了复杂的训练和调参过程。另外采用空间金字塔最大池化作为网络的输出层,减少了网络的输出维数和分类器的计算量。针对美国采暖、制冷与空调工程师学会提供的冷水机组的典型故障数据进行分类,结果表明,该方法相比深度卷积网络（CNN）和支持向量机（SVM）方法具有更高的故障诊断精度。