Improved system identification method for Hammerstein-Wiener processes

We propose a new system identification method for Hammerstein-Wiener processes, in which an input static nonlinear block, a linear dynamic block, and an output static nonlinear block are connected in a series. The proposed method can estimate the model parameters in a very simple way without solving the full-dimensional nonlinear optimization problem by activating the process with a specially designed test signal, composed of a relay feedback signal, a binary signal and a multi-step signal. The proposed method analytically identifies the output nonlinear static function and the input nonlinear static function from the relay signal and the multi-step signal, respectively. The linear dynamic subsystem is identified from the relay feedback signal and the binary signal with existing well-established linear system identification methods. We demonstrate with a simple example that the proposed method can be successfully applied to identify the Hammerstein-Wiener-type nonlinear process.     

A self-tuning control method for Wiener nonlinear systems and its application to process control problems

Many chemical processes can be modeled as Wiener models, which consist of a linear dynamic subsystem follow-ed by a static nonlinear block. In this paper, an effective discrete-time adaptive control method is proposed for Wiener nonlinear systems with uncertainties. The parameterization model is derived based on the inverse of the nonlinear function block. The adaptive control method is motivated by self-tuning control and is derived from a modified Clarke criterion function, which considers both tracking properties and control efforts. The un-certain parameters are updated by a recursive least squares algorithm and the control law exhibits an explicit form. The closed-loop system stability properties are discussed. To demonstrate the effectiveness of the obtained results, two groups of simulation examples including an application to composition control in a continuously stirred tank reactor (CSTR) system are studied.     

Extremum seeking control using a partial sum of input-output product

Recent extremum seeking control that uses a continuous perturbation and the integral feedback of perturbation- output product is based on a static nonlinear process. The method can be applied to dynamic nonlinear processes for tracking and maintaining the optimal operating points. It has several tuning parameters, such as the integral controller gain and the magnitude and frequency of the continuous perturbation signal. The frequency of the continuous perturbation signal should be low enough to ensure the time-scale separation between the real-time optimization and the process dynamics for the closed-loop stability. However, for some processes, fast perturbations are preferred because they can be attenuated easily in subsequent processes such as buffers and storages. For this, we propose an extremum seeking control method where the partial sum of perturbation-output product is used for a faster squarewave perturbation. Simulations for two processes of parallel competing reactions have been given, and a simple liquid level system to test extremum seeking control methods is suggested.     

Two-degree-of-freedom output feedback controllers for discrete-time nonlinear systems

A discrete-time, model-based output feedback control structure for nonlinear processes is developed in the present work. The structure makes use of a closed-loop observer, while at the same time it guarantees that the overall feedback controller possesses integral action. An algebraic transformation is applied on the observer states to insure that the input/output gain of the observer matches the model upon which the static state feedback control law is based. The resulting control algorithm is a two-degree-of-freedom control law, in the sense that the output and the set point are processed in different ways. The control structure is shown not only to have the same properties as the standard model-state feedback structure, but also that it emerges from a model algorithmic control framework. Finally, a simulation example using an exothermic CSTR operating at an open-loop unstable steady state is used to evaluate the closed-loop performance of the proposed method.     

Dynamic output feedback covariance control of stochastic dissipative partial differential equations

In this work, we develop a method for dynamic output feedback covariance control of the state covariance of linear dissipative stochastic partial differential equations (PDEs) using spatially distributed control actuation and sensing with noise. Such stochastic PDEs arise naturally in the modeling of surface height profile evolution in thin film growth and sputtering processes. We begin with the formulation of the stochastic PDE into a system of infinite stochastic ordinary differential equations (ODEs) by using modal decomposition. A finite-dimensional approximation is then obtained to capture the dominant mode contribution to the surface roughness profile (i.e., the covariance of the surface height profile). Subsequently, a state feedback controller and a Kalman-Bucy filter are designed on the basis of the finite-dimensional approximation. The dynamic output feedback covariance controller is subsequently obtained by combining the state feedback controller and the state estimator. The steady-state expected surface covariance under the dynamic output feedback controller is then estimated on the basis of the closed-loop finite-dimensional system. An analysis is performed to obtain a theoretical estimate of the expected surface covariance of the closed-loop infinite-dimensional system. Applications of the linear dynamic output feedback controller to both the linearized and the nonlinear stochastic Kuramoto-Sivashinsky equations (KSEs) are presented. Finally, nonlinear state feedback controller and nonlinear output feedback controller designs are also presented and applied to the nonlinear stochastic KSE.     

FEEDBACK LINEARIZATION DESIGNS FOR UNCERTAIN NONLINEAR TIME-DELAY PROCESSES: A COMPARATIVE STUDY

This work concerns the phenomena in which the feedback linearization control is applied to uncertain nonlinear time-delay processes. Under the I/O linearization algorithm, both nonlinear controllers are used to stabilize the closed-loop system with transformed delay inputs. When the effect of input perturbations can converge to zero or asymptotically vanish, these nonlinear feedback designs with only an adjustable parameter can directly improve the tracking performance. The simple linearizing controller can directly regulate the system output at unstable operating point. Combined with deadtime compensation the nonlinear predictive controller with the aid of appropriate state prediction is valid for the real process in the presence of large time delay. Finally, via computer simulation and test of control ability of both feedback control designs the useful comparative results are presented.     

非线性Hammerstein模型预测控制策略及其在pH中和过程中的应用

所有实际工业过程都包含一定程度的非线性,如pH中和过程由于其本身的强非线性是工业过程控制中具有挑战性的难题,但至今为止仍缺乏有效的非线性控制方法。将基于差分方程模型的模型预测控制策略(model predictive control,MPC)推广到包含一个静态非线性多项式函数和一个线性差分方程动态环节的非线性Hammerstein系统,详细描述了基于静态非线性多项式函数的最优控制作用求解方法,提出了一套新的非线性Hammerstein MPC 控制策略(nonlinear Hammerstein predictive control,NLHPC)。pH中和过程控制仿真和控制实验表明,NLHPC的控制结果好于工业上常用的非线性 PID(nonlinear PID,NL-PID)控制器。     

Fractional order integrator for the relay feedback identification of a process Nyquist point in the third quadrant

A fractional order integrator can be used for the relay feedback identification of a process Nyquist point in the third quadrant, and to implement the fractional order integrator, it is often approximated by integer order systems. Here, instead of the usual rational transfer function approximation of the fractional order integrator in the relay feedback system, a simple analytic method which utilizes the on-off characteristics of relay output is proposed. Simulation results show that the proposed method can find process Nyquist points in the third quadrant without worrying about the approximation errors and ranges of the fraction order integrator.     

Two-degree-of-freedom output feedback controllers for nonlinear processes

In this work, a nonlinear output feedback control algorithm is proposed, in the spirit of model-state feedback control. The structure provides state estimates using a process model, the measured output, and the residual between the model output and the measured output. These estimates will track the process states at a rate determined by a set of tunable parameters. An algebraic transformation of the state estimates is incorporated in the control structure to ensure that the input/output gain of the observer matches the model upon which the static state feedback control law is based. The transformed states are then used in the control law. This leads to a controller of minimal order possessing integral action. The control structure is shown to have the same properties as the standard model-state feedback structure. The resulting algorithm is a two-degree of freedom control law, in the sense that the control action is not a function of the error only, but the output and the set point are processed in different ways. Finally, a simulation example using an exothermic CSTR operating at an open-loop unstable steady state is used to demonstrate the closed-loop performance of the proposed method.     

Relay feedback identification for processes under drift and noisy environments

Relay feedback identification methods are widely used to find the process ultimate information and tune proportional‐integral‐derivative controllers. The conventional relay feedback method has several disadvantages, which include poor estimates of the process ultimate information for low‐order processes, chattering of relay for noisy environments, and asymmetric relay responses for constant biases or slow drifts in the process outputs. Methods to mitigate each of the above disadvantages are available. However, a systematic method to treat all of them has not been studied yet. Here, simple relay feedback methods that resolve these problems by introducing band‐pass filters in the feedback loop are proposed. The high‐pass filter part in band‐pass filter removes a constant bias or low frequency drift, and the low‐pass filter part removes high frequency noise and high‐order harmonic terms in the relay feedback oscillation, resulting better estimates of the process ultimate information. Because filters used for the proposed methods are able to reject constant biases, the process steady state gains can be estimated without disturbing the relay feedback oscillations and first order plus time delay (FOPTD) models can be obtained by combining the process steady state gains with the relay oscillation information. © 2010 American Institute of Chemical Engineers AIChE J, 2011     

Dynamic Modelling of a Polymer Electrolyte Membrane Fuel Cell Stack by Nonlinear System Identification

Real‐time models of polymer electrolyte membrane fuel cell (PEMFC) stacks with high accuracy are required, e.g. for the design of controllers or online diagnosis tools. By using physical and chemical laws representing the processes in a PEMFC stack, very detailed, but computationally complex models can be retrieved. In this paper, a nonlinear dynamic model obtained by system identification is proposed for PEMFC stacks. The model structure in this contribution is based on a modular concept and is divided into a static and a dynamic part. The static part represents the stationary points and the dynamic part describes the deviation from these stationary points due to changes in the input signals. Both parts can be modelled by different methods. A characteristic map and a neural network (NN) are proposed for the static part. For the dynamic part, transfer functions and a linear state‐space model retrieved by canonical variate analysis (CVA) are investigated.     

Robust controller synthesis for multivariable nonlinear systems with unmeasured disturbances

This work concerns robust controller synthesis using the differential geometric concepts for minimum phase nonlinear systems with unmeasurable disturbances. A pseudo-linearization of the disturbance model at the input-output linearization stage is applied to yield a linear subsystem for controller design. Based on this linear model, a multi-loop controller framework is implemented, whereby μ-synthesis is used to design off-line robust controller in the outer loop while state feedback is implemented in the inner loop. Through proper selection of weights, the outer robust controller is explicitly designed to address both uncertainty and disturbance rejection whereas the inner controller is used for on-line static state feedback. Numerical simulations are used to illustrate robustness of the controller for multi-input multi-output temperature control in two non-isothermal continuous stirred tank reactors in series.     

Process identification method using relay feedback and backward integrals

We propose a new process identification method that combines the two methods of the relay feedback to activate the process and the backward integrals to estimate the model parameters. Novel deviation variables are introduced to incorporate the case that the initial part of the process is unsteady-state without sacrificing the dynamic information included in the initial part, while the previous approaches assign zero-weighting to the initial parts, resulting in loss of the dynamic information included in the initial part. The final cyclic-steady-state part of the process input and output data is chosen as the reference of the deviation variables. The proposed method can estimate the model parameters analytically by using the backward integrals and the least squares method.     

Arma neuron networks for modeling nonlinear dynamical systems

A new neuronal structure, the ARMA neuron, is proposed here. These new neurons are designed for modeling nonlinear dynamics often encountered in chemical engineering processes. They are an extension of standard neurons which are used for static process modeling. These new neurons contain internal input/output dynamic structure and can model dynamic non-linearities in a flexible manner. A nonlinear output transformation is used here as opposed to a linear version used earlier (Krishnapura and Jutan, 1993). New algorithms for training networks comprised of the new ARMA neurons are developed using the backpropagation approach. The ARMA neurons are used to model both simulated and experimental nonlinear dynamic processes, including an industrial fluidized bed reactor.     

Dynamic gray‐box modeling for on‐line monitoring of polymer extrusion viscosity

Melt viscosity is a key indicator of product quality in polymer extrusion processes. However, real time monitoring and control of viscosity is difficult to achieve. In this article, a novel “soft sensor” approach based on dynamic gray‐box modeling is proposed. The soft sensor involves a nonlinear finite impulse response model with adaptable linear parameters for real‐time prediction of the melt viscosity based on the process inputs; the model output is then used as an input of a model with a simple‐fixed structure to predict the barrel pressure which can be measured online. Finally, the predicted pressure is compared to the measured value and the corresponding error is used as a feedback signal to correct the viscosity estimate. This novel feedback structure enables the online adaptability of the viscosity model in response to modeling errors and disturbances, hence producing a reliable viscosity estimate. The experimental results on different material/die/extruder confirm the effectiveness of the proposed “soft sensor” method based on dynamic gray‐box modeling for real‐time monitoring and control of polymer extrusion processes. POLYM. ENG. SCI., 2012. © 2012 Society of Plastics Engineers     

Approximation of the optimal minimum-phase output for control of nonlinear nonminimum-phase processes

Nonminimum-phase parts are better removed in the feedback loop like the time delay term. For this, Wright and Kravaris [1992] proposed the concept of optimal minimum-phase output to control nonlinear nonminimumphase processes. However, their optimal minimum-phase output has no analytic solutions for processes with more than three state variables. Here, methods for analytic minimum-phase outputs approximating the optimal ones are proposed, having no limitations in the number of state variables. The proposed methods provide analytic solutions for processes with three state variables and simple numerical solutions for those with more state variables.     

Nonlinear inferential control for an exothermic packed-bed reactor: geometric approaches

This article deals with the low-order output regulator designs for a commercial-scale packed-bed reactor. Under the static interpolation-based technique for data reconciliation mechanism, the optimum-based two-input control scheme by exploiting the secondary output information can almost attenuate measurable disturbances on the primary output. The steady-state approach not only reduces the critical hot spot and/or thermal runaway but it also dominates the desirable conversion rate in the exit. Moreover, the piece-wise iterative procedure connected with feasible programming evolution for control computation is proposed such that the flexible output regulator with the aid of the ‘intelligent’ algorithm can effectively improve the transient performance. Simulation results have shown that the non-distributed, piece-wise feedback control scheme turns out to be robust against unknown perturbations.     

Globally stable control systems for processes with input multiplicities

A nonlinear process with input multiplicity has two or more input values for a given output at the steady state, and the process steady state gain changes its sign as the operating point changes. A control system with integral action will be unstable when both signs of the process gain and the controller integral gain are different, and its stability region will be limited to the boundary where the process steady state gain is zero. Unlike processes with output multiplicities, feedback controllers cannot be used to correct the sign changes of process gain. To remove such stability limitation, a simple control system with parallel compensator is proposed. The parallel compensator can be easily designed based on the process steady state gain information and tuned in the field. Using the two time scale method, the stability of proposed control systems for processes with input multiplicities can be checked.     

FEEDBACK LINEARIZATION DESIGNS FOR UNCERTAIN NONLINEAR TIME-DELAY PROCESSES: A COMPARATIVE STUDY

This work concerns the phenomena in which the feedback linearization control is applied to uncertain nonlinear time-delay processes. Under the I/O linearization algorithm, both nonlinear controllers are used to stabilize the closed-loop system with transformed delay inputs. When the effect of input perturbations can converge to zero or asymptotically vanish, these nonlinear feedback designs with only an adjustable parameter can directly improve the tracking performance. The simple linearizing controller can directly regulate the system output at unstable operating point. Combined with deadtime compensation the nonlinear predictive controller with the aid of appropriate state prediction is valid for the real process in the presence of large time delay. Finally, via computer simulation and test of control ability of both feedback control designs the useful comparative results are presented.     

Assessment of controller performance: a relay feedback approach

Assessment of control loop performance is one of the important tasks to be carried out in a plant, regardless of the control strategy. The present work utilizes the shape information from single relay feedback test to assess the performance of a feedback system. The process considered is a typical first-order plus dead time process and the controller used is PI type of controller. An ideal relay is introduced in the feedback loop before the controller. The shape of the relay output characterizes the performance of the controller. The mismatch in the integral time can also be observed from the shape. Based on the shape information, analytical expressions for relay feedback responses are derived and straightforward procedures are evolved to assess the performance of the controller. The proposed scheme assesses the controller performance and computes the new tuning parameters, in case retuning of the controller is necessary. The robustness of the method is also tested for second-order plus dead time as well as for higher order systems. More importantly the present approach employs only one relay test for (1) assessment and (2) retuning of the controller. It provides a reliable way that is compatible even to a non-expert operator to assess the controller performance.