Alessandro Brambilla

Optimal operation of a separation plant using artificial neural networks

Abstract An ANN (Artificial Neural Network) based optimizer has been built and applied to the separation of a gas flow coming out of a hydrocracking reactor. The ANN based optimizer is shown to be able to achieve a precision similar to the one of an optimizer based on a first principle model with a very significant decrease in computational time. A central point of the development of the ANN models lies in the choices related to the training procedure; suggestions about the criteria which should lead such choices have been given.

Non linear standard regulators

Abstract The problem of designing Non Linear Standard Regulators (NLSR) is faced for uncertain systems with time delay. Some recent rules for robust tuning of conventional controllers are recalled. The favourable characteristics of the Variable Structure Control (VSC) approach to the design of NLSR for system without delay are shown by a comparison of performance with Standard Regulators in the presence of large parameter variations. The VSC approach is then extended, by means of a suitable modification of the Smith predictor control structure, to systems with time delay. A control law, able to give acceptable performance for different process parameters, is then proposed and simulation results are discussed.

Tuning of Conventional Controllers for Robust Performance

Abstract By adopting an appropriate tuning, conventional controllers can give results quite comparable with advanced controllers for problems of robust performance of relevant interest in process control. The control objective is to guarantee the fastest time response compatible with a small amount of overshoot, accounting for the presence of uncertainty in time delay and in the order of the model assumed for design. By means of a straightforward procedure, expressions are obtained for controllers parameters which permit to point out the approximations introduced by the conventional structure and leave only one parameter for tuning. Correlations for the tuning parameter are then given in order to achieve the desired robust performance specifications and a comparison with advanced controllers is presented.

Design of double filter IMC controller for ill-conditioned distillation columns

Abstract A direct method for robust design of double filter IMC controllers is presented. With respect to the general scheme, already proposed to control ill-conditioned plants, this method is based on a simple filter structure depending on fewer parameters. Guidelines for their selection are indicated; in many cases performance is mostly affected by a single parameter, thus making the controller tuning a much easier task to perform.

RIGOROUS SIMULATION AND MODEL PREDICTIVE CONTROL OF A CRUDE DISTILLATION UNIT

Abstract This paper describes the application of a widely-used commercial multivariable predictive controller to a rigorously simulated crude distillation process. After describing the main process and controller features, it is shown how the two simulation and control environments can be interfaced together. A number of simulation results of typical product quality changes and crude switches are presented. The final goal of this paper is to demonstrate how rigorous dynamic simulators can be effectively used to reduce the costs of Advanced Process Control projects by shortening model identification, controller design and commissioning phases.

The effect of vapour mixing on efficiency of large diameter distillation plates

Abstract Liquid and vapour flow maldistribution has been experimentally detected in large diameter distillation plates. Theoretical models developed to predict plate efficiency taking into account similar phenomena generally assume completely mixed vapour. This paper analyses the effect of vapour mixing on the plate efficiency for different values of stripping factor, point efficiency and liquid mixing degree and shows the remarkable influence of vapour mixing in presence of liquid maldistribution. Moreover a criterion for the prediction of the plate efficiency is suggested.

MULTIVARIABLE SUBSPACE IDENTIFICATION AND PREDICTIVE CONTROL OF A HEAT-INTEGRATED SUPERFRACTIONATOR RIGOROUS MODEL

Abstract The control of a heat-integrated superfractionator is addressed in this paper. A multivariable subspace identification method is proposed to overcome the difficulties associated to the large time constants of the process and identify a linear process model, upon which a constrained predictive controller is developed. The effectiveness of the proposed identification and control algorithm is shown by means of closed-loop rigorous dynamic simulation results.

ANALYSIS AND DESIGN OF CASCADE CONTROL SCHEMES FOR DISTILLATION

Abstract Different issues related to the introduction of a cascade loop in dual composition control of distillation columns are addressed. In particular it is pointed out how static and dynamic interactions can be reduced by the introduction of the cascade and a measure of this reduction is defined. Two other disturbance-specific parameters are defined in order to address the choice of the loop to which cascade is added and of the pilot tray. Applications to different columns show that the use of the defined parameters can guide advantageously in the analysis and design of cascade control for distillation.

A Critical Comparison of Linear and Nonlinear Property Estimators in Inferential Control

Abstract In this paper, a comparison of linear and nonlinear estimators with particular emphasis to the closed-loop properties of the resulting inferential control scheme is presented. The concept of closed-loop “consistency” is introduced as an effective criterion for choosing the auxiliary variables. An estimator is consistent if it guarantees low closedloop steady-state offset in the true unmeasured controlled variables. By means of a case study of a high purity distillation column, a number of issues that can arise in inferential control are emphasized, and their implications on the closed-loop stability are discussed. It is shown that the use of some nonlinear estimators, which in general guarantee a superior precision, may be inappropriate because of the presence of zero gains and gain inversions that can lead the closed-loop system to instability. Moreover, in multi-input multi-output (MIMO) systems it is possible that the estimator requires the auxiliary variables to reach values that are not reachable by the actual plant.

Estimator design with PLS model for consistent control of refinery main fractionators

Abstract In this paper the problem of designing product quality inferentials for refinery main fractionations is addressed by using the PLS regression. A simulated crude distillation unit is chosen as case study, and several linear steady-state estimators are designed and compared in terms of accuracy and consistency, i.e. the estimator ability of guaranteeing low closed-loop offset. The paper shows the importance of the auxiliary measurement choice in order to build an effective inferential control scheme. Moreover, it shows that the use of only temperature measurements is not sufficient to guarantee an acceptable estimator performance. Additional measurements as the operating pressure and internal flows have been used to improve the estimator accuracy and consistency.