Steady-state real-time optimization using transient measurements

Krishnamoorthy, Dinesh; Foss, Bjarne A.; Skogestad, Sigurd

doi:10.1016/j.compchemeng.2018.03.021

Cited by 70 publications

(49 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…One approach could be to combine trust region [17] and MA [18,19,20], reducing thus the distance between two consecutive iterates, therefore maintaining the next inputs in a region whereby the modified model is the most reliable. Methods enabling the estimation of the plant steady state during the transient operation have also been proposed [21,22,23,24], with emphasis on the convergence rate to the plant optimal operating conditions. This is indeed helpful for plant feasibility since the prediction of the steady state of the plant can be frequently updated and potentially improved during the transient, hopefully leading to better predictions of potential violations of the constraints and better decisions to be taken during the transient.…”

Section: Introductionmentioning

confidence: 99%

Output modifier adaptation with filter-based constraints

Papasavvas

François

2020

Journal of Process Control

View full text Add to dashboard Cite

Modifier adaptation (MA) and output modifier adaptation (MAy) are iterative model-based real-time optimization (RTO) algorithms that have the proven ability to drive plants to their optimal operating condition upon convergence despite disturbances and modeling uncertainty, provided the model at hand satisfies model adequacy conditions. But there is no guarantee that constraints are satisfied before convergence. In this article, an improvement of the formulation of MA and MAy is proposed that is proven to bring significant improvements w.r.t. these two limitations-model adequacy and feasibility of iterates. While standard MA or MAy suggests to perform optimization and filtering sequentially, it is proposed to integrate the input filtering stage in the modified model-based optimization problem by means of additional filter-based constraints. The corresponding approach, labeled "KMAy", is (i) proven to preserve constraint qualification despite additional constraints, (ii) proven to preserve the property of MA methods to converge to the true plant optimal inputs, (iii) proven to significantly relax the model adequacy condition-leading it to be independent of the constraints of the optimization problem, (iv) shown to increase the chances of converging from the safe side of the plant constraints and (v) shown to support the choice of input filtering, instead of output or modifier filtering, if the input filter is appropriately chosen. A method for the automatic selection of the largest filter gain with the five aforementioned assets, while minimizing the filter-induced conservatism, is also proposed. The performances of KMAy with and without adaptive gain are successfully illustrated by means of the optimization of a benchmark simulated chemical reactor.

show abstract

Section: Introductionmentioning

confidence: 99%

Output modifier adaptation with filter-based constraints

Papasavvas

François

2020

Journal of Process Control

View full text Add to dashboard Cite

show abstract

“…The Hybrid Real-Time Optimization (HRTO) is a modification of traditional RTO, it was proposed by Skogestad, (2018) to handle some of the limitations of the traditional RTO found in the industry. This technique tackles the challenge of steady-state wait time and eliminates the steady-state detection step of the SRTO mentioned in section 2.1 above.…”

Section: Hybrid Real-time Optimizationmentioning

confidence: 99%

“…Using statistical or heuristic tools, a condition for which a steady state can be said to be attained is set and the plant measurements are tested to detect if the steady state condition has been closely fulfilled. Once the process is believed to be operating close enough to this steady state condition the parameter In the implementation of the 'hybrid real time optimization' proposed by SKOGESTAD, (2018), the parameter estimation (information updating) step is the novelty of the technique. In this approach the information update is carried out using online parameter estimation tools.…”

Section: The Traditional Rto (Static Rto/mpa)mentioning

confidence: 99%

Integration of Real Time Optimization with Model Predictive Control applied to a Gas-lift System: A comparative study.

Shamaki¹

View full text Add to dashboard Cite

There is a rising need for more practical, efficient and sustainable techniques for improving industrial system operation in the face of a highly competitive market. The integration of real time optimization (RTO) and model predictive control (MPC) is a classical approach applied in the industry for improving processes. In this work, we analyze the application of the two most common RTO and MPC integration strategiesone-layer and two-layerto a gas-lifted system. We analyzed the performance of the economic cost function and efficiency in handling disturbance for each strategy, as well as consider practical industrial application. In the two-layer strategy, an upper economic optimization layer uses a rigorous nonlinear steady state model to compute the optimal process decision variables and send to the controller as an optimizing target which then computes the optimal control actions to achieve these targets. For this strategy, the hybrid RTO (HRTO) technique is implemented in the upper layer and an established controller-infinite horizon MPC with zone control-in the lower dynamic control layer. The hybrid RTO stems from the modification of the traditional static RTO found in the industry, to deal with the steady state wait time challenge. This is achieved by updating the optimizer with dynamic information rather than the static used in the traditional RTO. The zone control strategy allows the controller to focus on reaching a desired input target supplied by the optimization layer if the outputs are kept within their specified zones and constraints are respected. In the one-layer strategy, the gradient of the economic cost function is included in the controller cost function to be considered when computing the manipulated variable used to achieve optimal process operation. It was proposed with the main aim of practical industrial application. The two strategies were applied to a gas-lifted system and their results are compared and discussed considering economic objective. The results show that the IHMPC can reach the desired input targets despite abrupt disturbances of the uncertain parameter while keeping the outputs within the desired zone. Therefore, the HRTO can efficiently work with the IHMPC implemented in achieving optimal operation under uncertainties interfering as disturbance. It also shows that the onelayer strategy gives similar results to the two-layer strategy, implying that it can also achieve similar economic objective. However, the two-layer strategy using HRTO technique is more efficient in handling the disturbances.

show abstract

“…A probabilidade de sucesso na operação de uma unidade depende do alto grau de conhecimento das restrições do processo, do comportamento estacionário dos modelos matemáticos do processo e dos parâmetros, além do comportamento dinâmico dos sistemas de controle (CUTLER e PERRY, 1983). Em artigo de Forbes et al (1993) (FORBES et al, 1994, DARBY et al, 2011, KRISHNAMOORTHY et al, 2018.…”

Section: -Revisão Bibliográficaunclassified

“…Uma unidade petroquímica é composta por uma grande quantidade de processos e equipamentos, cada qual com características e objetivos diversos, que tornam a sua integração uma tarefa complexa. A operação de forma rentável e segura atendendo objetivos econômicos, de desempenho, ou ambientais, entre outros em escalas de tempo diferentes, pode demandar grande esforço em sua execução (KRISHNAMOORTHY et al, 2018). Na totalidade dos processos de uma unidade, o atendimento a todos os objetivos e restrições que o processo e o ambiente do negócio impõem, simultaneamente, é um desafio.…”

Section: -Introduçãounclassified

Otimização em linha (RTO) de rede de trocadores de calor de uma refinaria de petróleo.

Ikarimoto¹

View full text Add to dashboard Cite

Uma planta petroquímica, composta por processos e equipamentos com características diversas, caracteriza-se pela complexidade na integração e na operação de forma rentável e segura, atendendo a objetivos econômicos, de desempenho e ambientais. A implementação de ferramentas automatizadas de controle e otimização do processo tem um papel relevante no auxílio ao atendimento a tais objetivos da planta. Nesse contexto, a otimização RTO, abreviação do inglês para Real-time Optimization, é um algoritmo computacional que utiliza métodos matemáticos de otimização e os aplica ao processo químico real em que se busca o objetivo econômico de operação, respeitando-se suas características e restrições físicas e termodinâmicas. A otimização RTO é estruturada na forma clássica em: coleta e tratamento de dados, teste dos dados quanto à condição de estado estacionário, reconciliação de dados e estimação de parâmetros, otimização e realimentação de estado ótimo ao processo real. A aplicação da otimização RTO a processos de utilidades, em particular com o objetivo de maximizar a eficiência energética, é amplamente abordada em estudos acadêmicos. Dois estudos em reconciliação de dados e otimização foram realizados, a partir da reprodução de artigos de Narasimham e Jordache (2000) (reconciliação de dados) e Jäschke e Skogestad (2015) (otimização), utilizando-se os softwares IMPL® e MATLAB®. Posteriormente, a otimização RTO foi implementada e aplicada em uma rede de trocadores de calor (HEN-Heat Exchanger Network) da refinaria Alberto Pasqualini (REFAP-PETROBRAS), localizada na cidade de Canoas-RS. O estudo visou analisar a viabilidade da sua aplicação na otimização da recuperação energética de correntes quentes intermediárias do processo. Foram utilizados os softwares MATLAB®, para a etapa de detecção de estado estacionário, e GAMS® nas etapas de reconciliação de dados, estimação de parâmetros e otimização. O teste de detecção de estado estacionário F-modificado de Cao e Rhinerhat mostrou-se aplicável a dados em linha, após a seleção adequada dos dados a serem coletados e sintonia das constantes do método. As etapas de reconciliação de dados e estimação de parâmetros foram realizadas simultaneamente (DRPE-Data Reconciliation and Parameter Estimation), sendo estimados os coeficientes globais de transferência de calor e vazões de fluido quente por cada trocador de calor. Finalmente, a otimização foi realizada utilizando-se dois problemas de otimização diferentes: maximizando-se a troca térmica total da HEN, e minimizando-se a diferença das temperaturas de saída do fluido frio de cada ramal de trocadores de calor, tal como seria feito se fosse aplicado o controle auto otimizado SOC-Self-Optimizing Control. A partir dos resultados de ambas as otimizações, o potencial de aumento do aproveitamento energético da rede de trocadores de calor ficou entre 5,1% a 5,7%. Palavras-chave: Otimização em tempo real. Eficiência energética. Rede de trocadores de calor.

show abstract

Steady-state real-time optimization using transient measurements

Cited by 70 publications

References 41 publications

Output modifier adaptation with filter-based constraints

Output modifier adaptation with filter-based constraints

Integration of Real Time Optimization with Model Predictive Control applied to a Gas-lift System: A comparative study.

Otimização em linha (RTO) de rede de trocadores de calor de uma refinaria de petróleo.

Contact Info

Product

Resources

About