Long Range Predictive Control of Nonlinear Processes Based on Recurrent Neuro-Fuzzy Network Models

A recurrent neuro-fuzzy network-based nonlinear long range model predictive control strategy is proposed in this paper. The process operation is partitioned into several fuzzy operating regions. Within each region, a local linear model is used to model the process. The global model output is obtained through the centre of gravity defuzzification. Based upon a neuro-fuzzy network model, a nonlinear model-based predictive controller can be developed by combining several local linear model-based predictive controllers which usually have analytical solutions. This strategy avoids the time consuming numerical optimisation procedure, and the uncertainty in convergence to the global optimum which are typically seen in conventional nonlinear model-based predictive control strategies. Furthermore, control actions obtained based on local incremental models contain integration actions which can nat-urally eliminate static control offsets. The technique is demonstrated by an application to the modelling and control of liquid level in a water tank.     

Modelling and Multi-Objective Optimal Control of Batch Processes Using Recurrent Neuro-fuzzy Networks

In this paper, the modelling and multi-objective optimal control of batch processes, using a recurrent neuro-fuzzy network, are presented. The recurrent neuro-fuzzy network, forms a "global" nonlinear long-range prediction model through the fuzzy conjunction of a number of "local" linear dynamic models. Network output is fed back to network input through one or more time delay units, which ensure that predictions from the recurrent neuro-fuzzy network are long-range. In building a recurrent neural network model, process knowledge is used initially to partition the processes non-linear characteristics into several local operating regions, and to aid in the initialisation of corresponding network weights. Process operational data is then used to train the network. Membership functions of the local regimes are identified, and local models are discovered via network training. Based on a recurrent neuro-fuzzy network model, a multi-objective optimal control policy can be obtained. The proposed technique is applied to a fed-batch reactor.     

Modeling and Optimal Control of Batch Processes Using Recurrent Neuro-Fuzzy Networks

A recurrent neuro-fuzzy network based strategy for batch process modeling and optimal control is presented in this paper. The recurrent neuro-fuzzy network allows the construction of a “global” nonlinear long-range prediction model from the fuzzy conjunction of a number of “local” linear dynamic models. In this recurrent neuro-fuzzy network, the network output is fed back to the network input through one or more time delay units. This particular structure ensures that predictions from a recurrent neuro-fuzzy network are long-range or multi-step-ahead predictions. Long-range predictions are particularly important for batch processes where the interest lies in the product quality and quantity at the end of a batch. To enhance batch process control and monitoring, a model capable of predicting accurately the product quality/quantity at the end of a batch is required. Process knowledge is used to initially partition the process nonlinear characteristics into several local operating regions and to aid in the initialization of the corresponding network weights. Process input output data is then used to train the network. Membership functions of the local regimes are identified and local models are discovered through network training. An advantage of this recurrent neuro-fuzzy network model is that it is easy to interpret. This helps process operators in understanding the process characteristics. The proposed technique is applied to the modeling and optimal control of a fed-batch reactor.     

Multiple fuzzy model-based temperature predictive control for HVAC systems

In this paper, a multiple model predictive control (MMPC) strategy based on Takagi–Sugeno (T–S) fuzzy models for temperature control of air-handling unit (AHU) in heating, ventilating, and air-conditioning (HVAC) systems is presented. The overall control system is constructed by a hierarchical two-level structure. The higher level is a fuzzy partition based on AHU operating range to schedule the fuzzy weights of local models in lower level, while the lower level is composed of a set of T–S models based on the relation of manipulated inputs and system outputs correspond to the higher level. Following this divide-and-conquer strategy, the complex nonlinear AHU system is divided into a set of T–S models through a fuzzy satisfactory clustering (FSC) methodology and the global system is a fuzzy integrated linear varying parameter (LPV) model. A hierarchical MMPC strategy is developed using parallel distribution compensation (PDC) method, in which different predictive controllers are designed for different T–S fuzzy rules and the global controller output is integrated by the local controller outputs through their fuzzy weights. Simulation and real process testing results show that the proposed MMPC approach is effective in HVAC system control applications.     

Real-time control of a water-gas shift reactor by a model-based fuzzy gain scheduling technique

A model-based fuzzy gain scheduling technique is proposed. Fuzzy gain scheduling is a form of variable gain scheduling which involves implementing several linear controllers over a partitioned process space. A higher-level rule-based controller determines which local controller is executed. Unlike conventional gain scheduling, a controller with fuzzy gain scheduling uses fuzzy logic to dynamically interpolate controller parameters near region boundaries based on known local controller parameters. Model-based fuzzy gain scheduling (MFGS) was applied to PID controllers to control a laboratory-scale water-gas shift reactor. The experimental results were compared with those obtained by PID with standard fuzzy gain scheduling, PID with conventional gain scheduling, simple PID and a nonlinear model predictive control (NMPC) strategy. The MFGS technique performed comparably to the NMPC method. It exhibited excellent control behaviour over the desired operating space, which spanned a wide temperature range. The other three PID-based techniques were adequate only within a limited range of the same operating space. Due to the simple algorithm involved, the MFGS technique provides a low cost alternative to other computationally intensive control algorithms such as NMPC.     

Diagnostic neural adaptive control of drifting systems

A hierarchical network of neural network planning and control is employed to successfully accomplish a task such as grasping in a cluttered real world environment. In order for the individual robot joint controllers to follow their specific reference commands, information is shared with other neural network controllers and planners within the hierarchy. Each joint controller is initialized with weights that will acceptably control given a change in any of several crucial parameters across a broad operating range. When increased accuracy is needed as parameters drift, the diagnostic node fuzzy supervisor interprets the controller network's diagnostic outputs and transitions the weights to a closest fit specificchild controller. Future reference commands are in turn influenced by the diagnostic outputs of every robot joint neural network controller. The neural network controller and diagnostics are demonstrated for linear and nonlinear plants.     

机械手臂的多模型预测控制

针对机械手臂的非线性特点,提出了基于隶属度函数的多模型预测控制方法。该方法首先根据机械手臂的特点,选择合适的调度变量,将机械手臂的工作空间划分为若干个工作子空间,在每个子空间内的平衡点处对机械手臂进行线性化处理,得到相应的线性子模型,从而得到机械手臂的多模型表示;其次针对每个线性子模型设计局部预测控制器,使其在相应的子空间内达到控制要求;最后选择梯形隶属度函数与局部预测控制器进行加权求和,获得全局多模型预测控制器,以对机械手臂进行控制。仿真结果表明,当机械手臂的工作条件在大范围内变化时,全局多模型预测控制器的控制性能远优于常规PD控制器,达到了预期的控制目的。     

Developing a robust model predictive control architecture through regional knowledge analysis of artificial neural networks

Chemical processes are nonlinear. Model based control schemes such as model predictive control are highly related to the accuracy of the process model. For a highly nonlinear chemical system, it is clear to implement a nonlinear empirical model, such as artificial neural network model, should be superior to a linear model such as dynamic matrix model. However, unlike linear systems, the accuracy of a nonlinear empirical model strongly depends on its original data or training data based on how the model is built up. A regional-knowledge index is proposed in this study and applied in the analysis of dynamic artificial neural network models in process control. New input patterns that imply extrapolations and thus unreliable prediction by an artificial neural network model can be recognized from a significant decrease in the regional-knowledge index. To tackle the extrapolation problem and assure stability of the control system, we propose to run a neural adaptive controller in parallel with a model predictive control. A coordinator weights the outputs of these two controllers to make the final control decision. The present state of the controlled process and the model fitness to the present input pattern determine the weightings of the controller's output. The proposed analysis method and the modified model predictive control architecture have been applied to a neutralization process and excellent control performance is observed in this highly nonlinear system.     

Realization of robust nonlinear model predictive control by offline optimisation

This paper describes a new method for increasing the computational efficiency of nonlinear robust model-based predictive control. It is based on the application of neuro-fuzzy networks and improves the computation efficiency by arranging the online optimisation to be done offline. The offline optimisation is realized by offline training a neuro-fuzzy network, consisting of zero-order T–S fuzzy rules, which is designed to approximate the input–output relationship of a robust model-based predictive controller. The design and the training of the neuro-fuzzy network are described, and the corresponding control algorithm is developed. Experiment results performed on the temperature control loop of an experimental air-handling unit (AHU) demonstrate the effectiveness of this approach.     

基于神经网络非线性模型的多级工作点阶跃响应扩展DMC预测控制

利用前馈神经网络建立对象的非线性预测模型,在不同工作点做阶跃响应,建立其局部线性模型,用隶属函数将局部线性模型加权得到全局线性模型,全局线性模型用于滚动优化,非线性模型用于预测系统输出和校正线性模型,实现非线性预测控制,仿真结果表明该方法控制效果良好,可满足实时要求。     

输入采用广义梯形隶属函数的两维最简模糊控制器结构分析

总结了应用最为广泛的三角形和梯形隶属函数的共同特点, 明确定义了一种将以上两种隶属函数作为特例的广义梯形 (Generalized trapezoid-shaped, GTS) 隶属函数, 推导了输入变量采用 GTS 隶属函数的 I 类和 II 类两维最简模糊控制器的解析式. 基于此, 深入研究了模糊控制器的解析结构, 并证明了这两类模糊控制器等价于一种变结构的非线性 (或线性) PI 控制器与相应的非线性 (或定常) 控制偏置之和, 并且在其输入论域上是单调递增、连续且有界的. 最后, 将该类控制器应用于倒立摆控制系统, 通过仿真证明了其有效性, 同时揭示了此类控制器是一种更一般化的模糊控制器.     

Neuro-fuzzy Model-based Control

The paper deals with the Neuro-fuzzy model-based control and its application. Various types of the fuzzy logic and neural-net-based nonlinear autoregressive models with exogenous variables are reviewed with respect to the model error. Two types of model-based neuro-fuzzy control – a cancellation controller and a predictive controller are reviewed – and the robustness issues of such control are discussed. Finally, the application of the proposed design method to a laboratory scale heat exchanger is given.     

基于多LPV 模型的调度离线鲁棒预测控制

针对一类具有大工作区域和快时变特性的约束非线性系统, 采用多个线性参数时变(LPV) 模型近似描述原非线性系统. 对于各LPV 模型, 设计基于参数独立Lyapunov 函数的局部离线预测控制器. 构造各局部控制器间的切换策略, 在保证切换稳定性的同时, 使相互重叠的稳定域覆盖期望的工作区域. 仿真结果表明, 相比于已有的调度预测控制方法, 所提出的方法不仅能够保证控制输入在给定的约束范围内, 而且在局部控制器切换次数少的情况下, 获得良好的控制性能.

     

基于向量Lyapunov函数方法的模糊控制系统稳定性分析和设计

采用模糊动态模型对连续时间非线性系统进行模糊控制,对闭环模糊系统的稳定性进行分析,并给出系统化的控制器设计程序,在一系列局部模型通过模糊隶属函数连接得到的连续的全局模型中,全面考虑其它关联子系统对标称线性系统的摄动,并利用向量Lyapunov函数的概念和方法,得到了闭环模糊系统稳定的充分条件;仿真例子验证了该设计方法的正确性。     

Efficient scheduled stabilizing output feedback model predictive control for constrained nonlinear systems

In this note, we present a computationally efficient scheduled output feedback model predictive control (MPC) algorithm for constrained nonlinear systems with large operating regions. We design a set of local output feedback predictive controllers with their estimated regions of stability covering the desired operating region, and implement them as a single scheduled output feedback MPC which on-line switches between the set of local controllers and achieves nonlinear transitions with guaranteed stability. The algorithm is illustrated with a highly nonlinear continuous stirred tank reactor process.     

Multiple Neural Control Strategies Using a Neuro-Fuzzy Classifier

The paper deals with the control of complex dynamic systems. The main objective is to partition the whole operational system domain in local regions using an incremental neuro-fuzzy classifier in order to achieve multiple neural control strategies for the considered system. In our case, this approach is applied to a greenhouse operating during one day. Therefore, banks of neural controllers and direct neural local models are made from different partitioned greenhouse behaviors and two multiple neural control strategies are proposed to control the greenhouse. The selection of the suitable controller is accomplished by computing the minimal output error between desired and direct neural local models outputs in the case of the first control strategy and from a supervisor block containing the considered neuro-fuzzy classifier in the case of the second control strategy. Simulation results are then carried out to show the efficiency of the two control strategies.     

A Neuro-Fuzzy Approach to Identify Lettuce Growth and Greenhouse Climate

A hybrid neuro-fuzzy approach called the NUFZY system, which embeds fuzzy reasoning into a triple-layered network structure, has been developed to identify nonlinear systems. A set of membership functions at the input layer is partially linked with a layer of rules, using pre-set parameters. By means of a simplified centroid of gravity defuzzification method, the output becomes linear in the weights. Therefore, very fast estimation of the weight parameters can be achieved by using the orthogonal least squares (OLS) method, which also provides a method to efficiently remove the redundant fuzzy rules from the prototype rule base of the NUFZY system. In this paper, the NUFZY system is applied to identify lettuce growth and greenhouse temperature from real experimental data.Results show that the NUFZY model with the fast OLS training can perform quite well in predicting both lettuce growth and greenhouse temperature. In contrast to the mechanistic modeling procedures, the neuro-fuzzy approach offers an easier route and a fast way to build the nonlinear mapping of inputs and outputs. In addition, the resulting internal network structure of the NUFZY system is a self-explanatory representation of fuzzy rules. Under this frame, it is a perspective that one is able to incorporate the human knowledge in this approach, and, hopefully, to deduce any interpretable rules that describe the systems' behavior.     

Using adaptive neuro-fuzzy inference system for hydrological time series prediction

Conventionally, the multiple linear regression procedure has been known as the most popular models in simulating hydrological time series. However, when the nonlinear phenomenon is significant, the multiple linear will fail to develop an appropriate predictive model. Recently, intelligence system approaches such as artificial neural network (ANN) and neuro-fuzzy methods have been used successfully for time series modelling. In most instances for neural networks, multi layer perceptrons (MLPs) that are trained with the back-propagation algorithm have been used. The major shortcoming of this approach is that the knowledge contained in the trained networks is difficult to interpret. Using neuro-fuzzy approaches, which enable the information that is stored in trained networks to be expressed in the form of a fuzzy rule base, would help to overcome this issue. In the present study, a time series neuro-fuzzy model is proposed that is capable of exploiting the strengths of traditional time series approaches. The aim of this article is to investigate the potential of a neuro-fuzzy system with a Sugeno inference engine, considering different numbers of membership functions. Three rivers have been selected and daily prediction for them was applied. For better judgment, outcomes of the network have been compared to an autoregressive model.     

A method for design of a hybrid neuro-fuzzy control system based onbehavior modeling

It is known that control signals from a fuzzy logic controller are determined by a response behavior of a controlled object rather than its analytical models. That implies that the fuzzy controller could yield a similar control result for a set of plants with a similar dynamic behavior. This idea lends to modeling of a plant with unknown structure by defining several types of dynamic behaviors. On the basis of dynamic behavior classification, a new method is presented for the design of a neuro-fuzzy control system in two steps: 1) we model a plant with unknown structure by choosing a set of simplified systems with equivalent behavior as “templates” to optimize their fuzzy controllers off-line; and 2) we use an algorithm for system identification to perceive dynamic behavior and a neural network to adapt fuzzy logic controllers by matching the “templates” online. The main advantage of this method is that convergence problem can be avoided during adaptation process. Finally, the proposed method is used to design neuro-fuzzy controllers for a two-link manipulator     

Control design approaches for nonlinear systems using multiple models

It is difficult to realize control for some complex nonlinear systems operated in different operating regions. Based on developing local models for different operating regions of the process, a novel algorithm using multiple models is proposed. It utilizes dynamic model bank to establish multiple local models, and their membership functions are defined according to respective regions. Then the nonlinear system is approximated to a weighted combination of the local models. The stability of the nonlinear system is proven. Finally, simulations are given to demonstrate the validity of the proposed method.