Comparing MIMO Process Control Methods on a Pilot Plant

Boeira, Emerson Christ; Bordignon, Virgínia; Eckhard, Diego; Campestrini, Lucíola

doi:10.1007/s40313-018-0387-6

Cited by 7 publications

(3 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…where and are given by (4) and . The MPC objective is to minimize a quadratic cost function (6) with Rs being a column vector containing the future setpoints and a matrix that weighs ΔU.…”

Section: Model Predictive Controlmentioning

confidence: 99%

“…Controllers that make use of evolutive mechanisms [3], neural networks [4] a model free discrete time neural network control is designed for the trajectory tracking of a kind of nonlinear processes. The introduced control has three main characteristics: (1 or model prediction [5,6], for example, can be implemented easily using Python, as the language offers various facilities to manipulate generic data, like easy to use lists and dictionaries, generators, comprehensions, and offers good support to object-oriented and functional programming. Advanced control techniques like state-feedback [7] or even those with intensive computational requirements like fuzzy-control 8 and model predictive control [9,10] can then be easily implemented.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Affordable Control Platform with MPC Application

Sousa¹,

Leite²,

Scola³

2018

STUD INFORM CONTROL

View full text Add to dashboard Cite

This paper presents a control platform developed to interface with various hardware, allowing the design and rapid implementation even of advanced controllers, both on academic and industrial systems. The code of the controllers is written in the open-source Python language, facilitating the translation of code usually written in commercial software. The proposed platform can use from Arduinos to Programmable Logic Computers (PLCs). Beyond the research and tests on industrial facilities, the simplicity of the proposed platform allows its use also for educational and training purposes. Therefore, the proposed platform can help students focus on system analysis and control theory instead of hardware interfacing issues, while using low cost hardware. Developed in a client-server scheme, the platform can run in affordable computers while taking advantage of high-level mathematical and graphical tools available in Python language, allowing rapid implementation of advanced controllers. The use of this platform is illustrated with an implementation of a model predictive control (MPC) of a level control in a laboratory-scale process. A PLC is used to take the level measures, to dispatch control signals, and also for interlocking secure tasks. The controller runs on a Raspberry Pi computer that communicates with the PLC through an ethernet link.

show abstract

“…where and are given by (4) and . The MPC objective is to minimize a quadratic cost function (6) with Rs being a column vector containing the future setpoints and a matrix that weighs ΔU.…”

Section: Model Predictive Controlmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Affordable Control Platform with MPC Application

Sousa¹,

Leite²,

Scola³

2018

STUD INFORM CONTROL

View full text Add to dashboard Cite

show abstract

“…Furthermore, the centralized controller can easily be tuned even using the knowledge of the steady-state matrix alone. Extended discussions on the need for centralized control have been presented in published works. − …”

Section: Introductionmentioning

confidence: 99%

Nonsquare Multivariable Chemical Processes: A Hybrid Centralized Control Proposal

Revelo

Herrera

Camacho

et al. 2020

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

This paper presents a hybrid centralized control scheme for a nonsquare multivariable process. The proposed approach combines the Smith predictor, gain-scheduling methodology, and Davison method with the Particle Swarm Optimization (PSO) algorithm, all of which are combined to solve a nonsquare control-system problem that compensates for the multiple and different time delays and process nonlinearities. We call this fusion a hybrid control scheme. The Davison method does not provide a fine-tuning methodology for the centralized controller; therefore, the PSO method is added. This optimization method yields the best values for δ and ε, improving the process response with a smoother controller action, with a trade-off between the performance and robustness of the proposed controller. This method is applied to a reactor−separator−recycle (R−S−R) plant. These process types are characterized as being subjected to strong interactions among its variables and present strong nonlinearities. The R−S−R plant is modeled using the identification method based on the reaction curve, from which its equivalent transfer function (ETF) is determined. ETF represents a multivariable system with multiple time delays. In the current proposal, the nonlinearities that are present in the R−S−R system are compensated using the gain-scheduling strategy. The simulation tests verify the performance of the proposed controller, which is performed in MatLab. This controller is compared with a proportional integral (PI)-centralized controller and Smith delay compensator. All controllers are tuned using PSO.

show abstract