About
48
Publications
2,673
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
124
Citations
Introduction
Publications
Publications (48)
In this paper, a decentralized Proportional-Integral-Derivative control with an inverted decoupling structure is designed and implemented in an industrial diesel hydrotreating unit. The hydrotreating unit consists of two reactors with two catalytic beds each. The temperature control of each catalytic bed is disturbed by fluctuations in the precedin...
In this paper, an iterative procedure is proposed to design decentralized robust controllers for multivariable systems, based on the effective transfer function (ETF). The uncertainties in the process are taken into account using the multiplicative model, represented by a weight transfer function matrix. An H∞ optimization is applied to the ETF of...
In this work, connections between the areas of systems identification and deep learning are explored. A method of nonlinear systems identification is proposed based on deep neural networks and Hammerstein models. The proposed approach is formulated from the separation of the linear and non-linear parts of the models. Thus, the used network structur...
Multivariable (MIMO) processes are commonly found in the industry. The coupling between the different loops makes the controller design difficult. To solve the problem of controlling these processes, MIMO proportional integral (PI) controller design methods have been found in the literature. In this paper, the review and the experimental applicatio...
Many controller tuning techniques are model-based. For multivariable processes, the model identification is a difficult task. On the other hand, these processes are often found in the industry. Thus, data-driven control techniques have been developed. In this article, the VRFT and OCI tuning techniques for the multivariable PI controller are review...
Proportional integral derivative (PID) controllers are the most used in practice for regulatory control. This is due to the good performance achieved by this controller for a variety of processes. However, about 60% have performance issue. This problem can be more evident in multivariable processes, due to the coupling between the loops. One way to...
In this paper, an iterative method to identify first-order plus dead-time (FOPDT) models is proposed. The proposed method uses two frequency points as equality constraints to estimate the model parameters using the least-squares identification algorithm. The time delay term is approximated by a second-order Taylor expansion. The iterative method is...
Effective decoupling is not always achieved due to unavoidable model mismatch or change in the operating point. Thus, evaluation and redesign of the decoupler may improve performance. The redesign of decouplers is a subject that received little attention in the literature. In this paper, a methodology to evaluate and redesign simplified decouplers...
A Proportional-Integral (PI) controller frequency domain retuning technique for square multivariable process is presented in this paper. Assuming knowledge of the current PI controller, the controller gains increments are computed in such that the new closed-loop matches with a desired reference model. This is performed using only the closed-loop r...
This paper presents the application of the time and frequency domain data-driven methods to the iterative decentralized PID tuning of MIMO process. Each single loop is sequentially excited to generate data required for retuning. The initial PID tuning gains are adjusted in order to match the defined decentralized MIMO reference model. The proposed...
In this paper, a multi-input multi-output proportional integral derivative controller tuning technique using frequency-domain data is proposed. The problem is formulated using loop shaping performance specifications and the bounds on the control signal magnitude as constraints. The result is a convex-concave optimization problem. This is convexifie...
Neste trabalho, abordam-se duas técnicas clássicas de controle PI baseadas em dados. O objetivo é revisar e comparar as técnicas Iterative Feedback Tuning (IFT) e Correlation-based Tuning (CbT) em um contexto de prática laboratorial. A escolha do modelo de referência é realizada utilizando uma abordagem IMC (Internal Modal Control). Essas técnicas...
Neste artigo, uma estrutura de controle Proporcional-Integral-Derivativo (PID) descentralizado com desacoplamento invertido e projetado e implementado em uma unidade industrial de Hidrotratamento (HDT) de Diesel. A unidade HDT é composta por dois reatores com dois leitos catalíticos cada. O controle de temperatura de cada leito é perturbado por flu...
No controle feedforward, os modelos do processo e da perturbação podem resultar em um problema onde o compensador ideal é não-realizável. Sua função de transferência pode ser imprópria, possuir polos instáveis ou atrasos não-realizáveis. Neste trabalho é apresentada uma técnica de sintonia para controladores feedforward. O método consiste na soluçã...
In this paper, data-driven proportional–integral–derivative (PID) controller retuning is addressed. Time and frequency domain data collected from a closed-loop experiment is used to evaluate the actual closed-loop and to retune the PID controller. Assuming knowledge of the current PID controller, optimal controller gains increments are computed in...
The inverted decoupler is usually designed from estimated process models. These models are usually of reduced order, so that effective decoupling is not always achieved. The redesign of decouplers is a subject that received little attention in literature. In this paper, the methodology for decoupler evaluation and redesign in open-loop is reviewed....
In this paper, inverted decoupling control for two-input/two-output (TITO) processes is considered. Knowing that the design of the decouplers is based on estimated models of the processes, adjustments in the decoupling parameters may be necessary. Thus, a methodology of evaluation and redesign of the inverted decoupler is proposed. Simulation examp...
This paper is about data-driven PID iterative tuning technique using frequency and time domain data from the initial controller parameters. The data is obtained using relays experiments and the PID tuning gains are adjusted in order to match the defined reference model and frequency response specializations using a constraint iterative optimization...
This paper is about a description and implementation of a relay switched by output limits defined by an user that excites the system for some defined conditions. After the end of the experiment, the strategy makes the process tend to a condition similar to the initial. The generated data can be used for parameter identification for ITD and IFOTD mo...
This paper is about modeling and control of a XY positioning table. It is described the construction and instrumentation of a double axis positioning system. It is developed a general electromechanical model, followed by the identification of simplified models for control purposes. At last, it is proposed a PID control strategy, tuned by Internal M...
It is presented a state-space realization approach for identification of a pilot scale heat exchanger plant. The proposed methodology is a computational efficient way to estimate state-space models from experimental data with pulsed input signals. It is specially suitable for short data batches.
In this paper a methodology is presented to assess the performance of IMC PI controllers from closed-loop response data for a chosen setpoint signal. Time and frequency domains tools are used to assess the closed-loop performance. Simulation examples illustrate the methodology.
Considerando o interesse recente em pesquisas sobre controle com desacoplamento para sistemas TITO, o objetivo deste artigo é apresentar valorosa revisão bibliográfica sobre esse tópico e associado a isso apresentar estudo comparativo entre algumas das principais técnicas de projeto do desacoplador para sistemas TITO. Nesse contexto, três técnicas...
In this paper a proposal for redesign PI controllers with Symmetrical Optimum (SO) based design is presented. The method uses an integration plus first-order model, identified from two frequency points obtained from combined relay experiment. The proposal for the identification of the used model is also presented. The main idea is to apply a double...
In this paper it is proposed a procedure for closed-loop evaluation and PI controller redesign using a relay-based experiment. The experiment data is used to evaluate the actual closed-loop. It is evaluated using classical robustness measures. A first-order plus dead-time(FOPDT) process model is estimated using the relay data. The controller redesi...
In this paper the decoupling control of two-inputs two-outputs (TITO) processes using a simplified dynamic decou- pler is considered. A frequency-domain identification technique that uses a relay experiment to estimate a first-order plus dead- time (FOPDT) TITO model is presented. The resulting FOPDT TITO model is accurate around at the frequency o...
PID controllers are widely used in the process control industry due to their simplicity and ease of re-tuning on-line. A software package has been developed for system identification and PID control loop tuning by a partnership between UFCG and CENPES-PETROBRAS. This system implements literature and industry standard PID tuning algorithms and devel...
In this paper it is described a flexible laboratory-scale quadruple-tank coupled system designed for Control Education and Research purposes. The quadruple-tank process presented is a flexible structure for experimental implementation and evaluation of different control strategies in a multivariable non-linear process. The structure allows the impl...
First-Order plus Time-Delay (FOPTD) models are extensively used to approximate systems in order to tune PID controllers. Several step response based identification techniques which focus on the time response have been developed. In order to recover a model which matches the time response while capturing the system dynamics around frequencies of int...
First-Order plus Time-Delay (FOPTD) models are extensively used to approximate systems in order to tune PID controllers. Several estimation techniques have been developed and the obtained models applied to controller design. Most estimation techniques are based on least-squares and excitation such as step, square waves and pseudo-random signals. In...
A laboratory-scale temperature system for Control and Automation courses is presented. The system is a coupled two-input two-output process which use peltier modules for temperature control. This paper describes the laboratory-scale system in details. A linear dynamic model of the thermoelectric cooler including the heat sink and the cooling-load h...
In this work the identification of first-order plus dead-time models from a relay experiment is considered. The relay excitation is applied to the closed-loop. Alternative techniques for identification are examined and simple algorithms are proposed for dealing with the dead-time. Simulation examples are used to illustrate the techniques.