UNBIASED IDENTIFICATION OF CONTINUOUS-TIME PARAMETRIC MODELS USING A TIME-WEIGHTED INTEGRAL TRANSFORM

A time-weighted integral transform is presented to identify a continuous SISO or MIMO parametric model based on a single dynamic test under open-loop or closed-loop operation. Moving-horizon algorithms are proposed to obtain unbiased estimates of the model parameters. The off-line algorithm in a least-squares form and the on-line algorithm in a recursive form are provided. An effective technique based on pattern recognition is also developed to determine the system order and time delay from observed data in a simple manner. Furthermore, the proposed method can be easily applied as a model reduction technique that results in an ideal model with delay for any specified order.     

ROBUST NONLINEAR PREDICTIVE CONTROL USING A DISTURBANCE ESTIMATOR

A robust nonlinear predictive control strategy using a disturbance estimator is presented. The disturbance estimator is comprised of two parts: one is the disturbance model parameter adaptation and the other is future disturbance prediction. A linear discrete model is proposed as a disturbance model which is formulated by using process inputs and available process measurements. The recursive least square (RLS) method with exponential forgetting is used to determine the uncertain disturbance model parameters and for the future disturbance prediction, future disturbances projected by the future process inputs are used. Two illustrative examples: a jacketed CSTR as a SISO system: an adiabatic CSTR as a MIMO system, and experimental results of the distillation column control are presented. The results indicate that a substantial improvement in nonlinear predictive control performance is possible using the disturbance estimator.     

Unbiased identification of continuous-time parametric models using a time-weighted integral transform

A time-weighted integral transform is presented to identify a continuous SISO or MIMO parametric model based on a single dynamic test under open-loop or closed-loop operation. Moving-horizon algorithms are proposed to obtain unbiased estimates of the model parameters. The off-line algorithm in a least-squares form and the on-line algorithm in a recursive form are provided. An effective technique based on pattern recognition is also developed to determine the system order and time delay from observed data in a simple manner. Furthermore, the proposed method can be easily applied as a model reduction technique that results in an ideal model with delay for any specified order.     

On improving accuracy of computationally efficient nonlinear predictive control based on neural models

For nonlinear processes the classical model predictive control (MPC) algorithm, in which a linear model is used, usually does not give satisfactory closed-loop performance. In such nonlinear cases a suboptimal MPC strategy is typically used in which the nonlinear model is successively linearised on-line for the current operating point and, thanks to linearisation, the control policy is calculated from a quadratic programming problem. Although the suboptimal MPC algorithm frequently gives good results, for some nonlinear processes it would be beneficial to further improve control accuracy. This paper details a computationally efficient nonlinear MPC algorithm in which a neural model is linearised on-line along the predicted trajectory in an iterative way. The algorithm needs solving on-line only a series of quadratic programming problems. Advantages of the discussed algorithm are demonstrated in the control system of a high-purity ethylene–ethane distillation column for which the classical linear MPC algorithm does not work and the classical suboptimal MPC algorithm is slow. It is shown that the discussed algorithm can give practically the same control accuracy as the algorithm with on-line nonlinear optimisation and, at the same time, the algorithm is significantly less computationally demanding.     

混杂系统基于参数规划的状态反馈预测控制

In this article, state feedback predictive controller for hybrid system via parametric programming is proposed. First, mixed logic dynamic （MLD） modeling mechanism for hybrid system is analyzed, which has a distinguished advantage to deal with the logic rules and constraints of a plant. Model predictive control algorithm with moving horizon state estimator （MHE） is presented. The estimator is adopted to estimate the current state of the plant with process disturbance and measurement noise, and the state estimated are utilized in the predictive controller for both regulation and tracking problems of the hybrid system based on MLD model. Off-line parametric programming is adopted and then on-line mixed integer programming problem can be treated as the parameter programming with estimated state as the parameters. A three tank system is used for computer simulation, results show that the proposed MHE based predictive control via parametric programming is effective for hybrid system with model/olant mismatch, and has a potential for the engineering applications.     

Optimization‐Based Nonlinear Centralized Controller Tuning of Liquid‐Liquid Extraction Processes

Abstract

The control problem of an agitated contactor is considered in this work. A Scheibel extraction column is modeled using the non‐equilibrium backflow mixing cell model. Model dynamic analysis shows that this process is highly nonlinear, thus the control problem solution of such a system needs to tackle the process nonlinearity efficiently. The control problem of this process is solved by developing a multivariable nonlinear control system implemented in MATLAB?. In this control methodology, a new controller tuning method is adopted, in which the time‐domain control parameter‐tuning problem is solved as a constrained optimization problem. A MIMO (multi‐input multi‐output) PI controller structure is used in this strategy. The centralized controller uses a 2×2 transfer function and accounts for loops interaction. The controller parameters are tuned using an optimization‐based algorithm with constraints imposed on the process variables reference trajectories. Incremental tuning procedure is performed until the extractor output variables transient response satisfies a preset uncertainty which bounds around the reference trajectory. A decentralized model‐based IMC (internal model control) control strategy is compared with the newly developed centralized MIMO PI control one. Stability and robustness tests are applied to the two algorithms. The performance of the MIMO PI controller is found to be superior to that of the conventional IMC controller in terms of stability, robustness, loops interaction handling, and step‐change tracking characteristics.     

Empirical strategies for open-loop on-line optimization

Two empirical strategies for open-loop on-line optimization are developed as alternatives to the use of mechanistic process models. These strategies are based on on-line identification of dynamic multi-input single-output (MISO) and multi-input multi-output (MIMO) models. The steady state gain of these models provides information for steady state optimization. Desirability functions, originally developed for multi-objective optimization, are utilized as objective function modifiers for constrained on-line optimization. The integration of dynamic model identification and desirability functions results in an on-line optimizer which combines fast optimizing speed with the ability to predict future encroachments on constraint boundaries. Corrections to the search direction are based on these predictions, reducing the probability of actual constraint violation. The optimization strategies are tested by simulation on nonlinear multivariable interacting systems at two levels of complexity: a CSTR supporting a multiple reaction and a fluid catalytic cracker. Both methods were effective in avoiding violation of constraints but the MIMO strategy required fewer steps to reach an optimum and was less prone to generate a nonfeasible optimization step.     

Closed-loop step response identification of integrating and unstable processes

Motivated by the fact that integrating and unstable processes are usually operated in a closed-loop manner for safety and economic reasons, this paper proposes a systematic closed-loop identification method based on step response test to facilitate closed-loop system operation and on-line optimization. To avoid jeopardizing the closed-loop system stability of such a process, guidelines are given for proper implementation of a closed-loop step test for model identification. By introducing a damping factor to the closed-loop step response for realization of the Laplace transform in frequency domain, a frequency response estimation algorithm is developed in terms of the closed-loop control structure used for identification. Accordingly, three model identification algorithms are derived analytically in frequency domain to obtain the widely used low-order process models of first-order-plus-dead-time (FOPDT) and second-order-plus-dead-time (SOPDT). To enhance fitting accuracy for a higher order process, in particular for a specified frequency range interested to control design and on-line tuning, a weighted least-squares fitting algorithm is also given based on the estimated process frequency response points. Illustrative examples from the recent literature are used to demonstrate the effectiveness and merits of the proposed identification algorithms.     

A study on the adaptive predictive control method for multivariable bilinear processes

A predictive control method for multivariable bilinear processes is derived based on ARMA model. To identify bilinear process models, we use simple equation error method extended to multivariable system. We can obtain the adaptive predictive controller for multivariable bilinear processes by incorporation of the identification algorithm. Offset compensator is provided to correct for the effects of unmeasured disturbances and model inaccuracies. A filter with a singled parameter is used to correct for the effects of an incorrect model. Results of simulation on multivariable bilinear processes show that the proposed control method has satisfactory performance.     

ON-LINE OPTIMIZATION OF DISTILLATION COLUMNS IN SERIES

This paper presents an on-line optimization algorithm for distillation columns which consists of a steepest descent technique based on a simple model of product recovery with on-line estimation of the critical model parameter. The algorithm is developed for a single column and then for a two-column train. A dynamic programming approach is used to reduce the optimization problem to a single parameter search for each column in the train. The resulting algorithm is simple and computer resource requirements are small. The algorithm has been successfully used in two industrial applications, one consisting of two columns in series and the other of three columns in series.     

一种基于混合建模技术的MIMO软测量建模方法

针对复合肥产品中几种养分含量需要同时预报的一类多输入/多输出(MIMO)软测量建模问题,提出一种基于混合建模方法的复合肥养分含量MIMO软测量模型。该混合模型首先对几个不能实时测量的关键辅助变量采用基于限定记忆部分最小二乘算法的数据驱动建模方法建立自适应软测量模型,然后采用简化机理模型实时计算三种养分含量。基于实际工业过程数据的仿真结果表明,所建模型运算速度快、预测精度高,可以满足复合肥养分含量在线预报的要求。     

OPERATING REGIME DECOMPOSITION USING NEURAL NETWORKS

This article addresses the problem of identification of a nonlinear process operating over a wide range of conditions. The global space is divided into multiple local regimes, a nonlinear model is developed for each regime, and a quadratic programming-based algorithm is used to ensure smooth transition between the regimes on-line. The use of nonlinear models as opposed to linear models reduces the number of local regimes needed. Neural networks are used to model these regimes because of their strong ability to capture nonlinearity, and their combination with the switching algorithm improves transient performance. The performance of the method is demonstrated on an exothermic CSTR and a pH neutralization process.     

Operating regime decomposition using neural networks

This article addresses the problem of identification of a nonlinear process operating over a wide range of conditions. The global space is divided into multiple local regimes, a nonlinear model is developed for each regime, and a quadratic programming-based algorithm is used to ensure smooth transition between the regimes on-line. The use of nonlinear models as opposed to linear models reduces the number of local regimes needed. Neural networks are used to model these regimes because of their strong ability to capture nonlinearity, and their combination with the switching algorithm improves transient performance. The performance of the method is demonstrated on an exothermic CSTR and a pH neutralization process.     

On-line estimation of vapour pressure of stabilized gasoline via ANN’s

The use of artificial neural network based model for the on-line estimation of the Reid Va-por Pressure of stabilized gasoline in a stabilizer after the stripper-reabsorber in the fluid catalyticcracking unit is investigated.The quadratic basis function network(QBFN)which uses a simplequadratic function instead of sigmoid function typically used in back-propagation network is em-ployed.180 sets of historical operation data have been selected for training and testing the QBFN.To overcome the local minimum point which often occurs during the training phase,a new algorithmcombining the simulated annealing approach with the improved GDR has been applied.Furthermore,the developed model has been installed on-line in a refinery for on-line testing.Thetesting results show that the model is sufficiently accurate and it can be used on site as an on-lineanalyzer.     

On-line implementation of a multichannel estimator

In this work, the on-line implementation of an alternative algorithm for state-parameter estimation is discussed. A real-time multichannel estimator is developed based on the idea of parallel processing. It can be applied to a multiparameter phenomenon to estimate the individual values accurately and with a minimum of computing effort. Each parameter can be dealt with separately and when desired. This technique is applied to data obtained from experiments to evaluate heat-transfer parameters in a natural convention process. Results are compared with those obtained from well-known empirical correlations.     

基于定量反馈理论的多变量解耦预测控制及其在CSTR化学过程中的应用

1 INTRODUCTION Many multi-input and multi-output (MIMO) sys- tems worldwide are regarded as linear invariants, but there are still some difficulties in controlling these systems. The challenges arise from the need to achieve both robust stability and control performance when the plants to be controlled are highly uncer- tain[1―3]. Quantitative feedback theory (QFT) is a fre- quency domain design technique[4], which is perhaps the only known method that deals with highly uncer- tain pla…     

基于ISR-UKF的PET固相缩聚过程参数和状态估计

连续式PET固相缩聚移动床反应器具有显著的分布参数特性,由于建模简化或过程时变等原因使得所建模型参数不精确,导致反应器状态的估计失真。首先采用正交配置方法离散PET固相缩聚过程的偏微分方程模型,然后基于改进的平方根不敏Kalman滤波算法(ISR-UKF),设计自适应联合估计器,同时获得参数和状态估计值。实验结果表明,参数估计结果合理,状态估计精度较高且稳定性好;并获得了频率因子、活化能近似值和有效系数随反应温度动态变化的规律,表明所提出的联合估计器能获得较好的实际应用效果。     

基于混合建模技术的复合肥养分含量MIMO软测量模型

In compound fertilizer production, several quality variables need to be monitored and controlled simultaneously. It is very diifficult to measure these variables on-line by existing instruments and sensors. So, soft-sensor technique becomes an indispensable method to implement real-time quality control. In this article, a new model of multi-inputs multi-outputs （MIMO） soft-sensor, which is constructed based on hybrid modeling technique, is proposed for these interactional variables. Data-driven modeling method and simplified first principle modelingmethod are combined in this model. Data-driven modeling method based on limited memory partial least squares（LM-PLS） al.gorithm is used to build soft-senor models for some secondary variables.then, the simplified first principle model is used to compute three primary variables on line. The proposed model has been used in practicalprocess; the results indicate that the proposed model is precise and efficient, and it is possible to realize on line quality control for compound fertilizer process.     

Pole-placement self-tuning control of nonlinear Hammerstein system and its application to pH process control

By taking advantage of the separation characteristics of nonlinear gain and dynamic sector inside a Hammerstein model, a novel pole placement self tuning control scheme for nonlinear Hammerstein system was put forward based on the linear system pole placement self tuning control algorithm. And the nonlinear Hammerstein system pole placement self tuning control (NL-PP-STC) algorithm was presented in detail. The identification ability of its parameter estimation algorithm of NL-PP-STC was analyzed, which was always identifiable in closed loop. Two particular problems including the selection of poles and the on-line estimation of model parameters, which may be met in applications of NL-PP-STC to real process control, were discussed. The control simulation of a strong nonlinear pH neutralization process was carried out and good control performance was achieved.     

基于混合建模技术的复合肥养分含量MIMO软测量模型

In compound fertilizer production, several quality variables need to be monitored and controlled simultaneously. It is very difficult to measure these variables on-line by existing instruments and sensors. So, soft-sensor technique becomes an indispensable method to implement real-time quality control. In this article, a new model of multi-inputs multi-outputs (MIMO) soft-sensor, which is constructed based on hybrid modeling technique, is proposed for these interactional variables. Data-driven modeling method and simplified first principle modeling method are combined in this model. Data-driven modeling method based on limited memory partial least squares (LM-PLS) algorithm is used to build soft-senor models for some secondary variables; then, the simplified first principle model is used to compute three primary variables on line. The proposed model has been used in practical process; the results indicate that the proposed model is precise and efficient, and it is possible to realize on line quality control for compound fertilizer process.