Argimiro R. Secchi

Federal University of Rio de Janeiro | UFRJ · COPPE - Chemical Engineering Program

In this note, we ensure that the dislocation hyperbolic augmented Lagrangian algorithm converges to a global minimizer, we assuming nonconvexity assumptions. The subproblem generated by this algorithm is solved with the DIRECT algorithm. Finally, we present computational experiments to show the good performance of the proposed algorithm.

Gas-lift is a strategy to enhance oil production from oil wells by reducing the hydrostatic pressure of the fluid. Efficient modeling of this process is a key point for the oil and gas industry to maximize oil output while minimizing gas consumption. This study introduces a novel hybrid model for the gas-lift process, using the Universal Differenti...

In this work, two model-based controllers were developed, one based on a nonlinear model-based controller (NMPC) using a population balance model (PBM) and another using a machine learning approach based on a neural network inverse model-based controller (NNIMC). The performance of the two model-based controllers was compared for different scenario...

Real-time optimization (RTO) methodologies have become essential for optimal process operation in the oil and gas industries. Typically, RTO is based on a steady-state model (steady-state real-time optimization - SSRTO) and operates as a closed-loop optimizer. However, this technique can result in suboptimal policies due to steady-state waiting and...

An attractive method for steady-state real-time optimization (RTO) using transient measurements consists of a persistent parameter adaptation throughout a dynamic parameter estimation followed by a static economic optimization. The method is called RTO with persistent parameter adaptation (ROPA) or Hybrid RTO (HRTO). Although such a method avoids t...

This paper addresses how physical knowledge can improve machine learning in process control. A data-driven tracking control framework using physics-informed neural networks (PINNs) and deep reinforcement learning (DRL) is proposed for dynamical systems, which is particularly important when iterations or repetitions of data collection experiments ar...

Economic dynamic real-time optimization using reinforcement learning approaches is investigated. This task is reformulated as a Markov decision process, and a deep deterministic policy gradient algorithm is used for its resolution. The resulting controller performs well applied to a simulation example of a CSTR affected by process disturbances. In...

RESUMO-Crystallization is a unit operation crucial for the many industries and plays a crucial role in the pharmaceutical industry, since the products majorly consist of solid materials of high purity, being present in almost all of these processes. This work presents an experimental investigation based of the crystallization of praziquantel (PZQ)...

Emission monitoring systems in gas turbines are essential to maintain the energy supply to society. However, gas turbines must be designed to reduce emissions of CO and NOx. One way to monitor emissions is to use a soft sensor , which is a machine learning model. In this study, a Random Forest-based soft sensor was developed to predict CO and NOx e...

No presente trabalho, foi desenvolvido um modelo substituto para o mo-delo fenomenológico proposto previamente utilizado para a simulação do digestor contínuo de polpação Kraft. Foram comparados os desempenhos da rede neuronal MLP-MultiLayer Perceptron e de algoritmos de árvores de decisão, e das abor-dagens com uma e múltiplas saídas. A combinação...

Reinforcement learning (RL) arises from the set of machine learning techniques that are interesting for data-based process control purposes. Many authors discuss the advantages of using RL techniques as alternatives or complements to classical control frameworks, such as model predictive control (MPC), to tackle their drawbacks. The most popular...

Crystallization is a separation and purification process crucial to the pharmaceutical and food industries. In this work, experimental data were used to develop neural network models for the inverse dynamics of batch cooling crystallization of potassium sulfate (K2SO4). The experimental data used to train the NN model came from a series of batch co...

This work proposes an iterative algorithm to fit a multiphase flow model to measured data from offshore fields to be used in real-time applications. Due to the best compromise between accuracy and computational speed, the Drift-Flux Model was used in the proposed procedure to describe the multiphase flow behavior inside a real offshore production p...

The COVID-19 global pandemic is still affecting the world, even considering vaccine applications in most countries, especially due to new variant outbreaks and the possibility that they may present immunological escape. Therefore, mass testing is relevant in infection monitoring and restriction policy evaluations, making low-cost and easy-to-use te...

This paper reviews real-time optimization from a reinforcement learning point of view. The typical control and optimization system hierarchy depend on the layers of real-time optimization , supervisory control, and regulatory control. The literature about each mentioned layer is reviewed, supporting the proposal of a benchmark study of reinforcemen...

The hydrogen bonding network analysis of softwood lignin is relevant to designing novel technologies to overcome the recalcitrance of plant biomass in the industrial deconstruction and manufacturing of lignin-based carbon fibers. In this work we examine, by atomistic simulations, the hydrogen bonding network in guaiacyl-rich lignin and guaiacyl-typ...

Modern design methodologies combine emerging data science tools with modeling, simulation, optimization, control, and design algorithms. One process of great interest where those modern methodologies can be applied is the enantioseparation of praziquantel (PZQ) using the simulated moving bed (SMB) process or its variants. The pure enantiomers of PZ...

This paper presents a literature review of reinforcement learning (RL) and its applications to process control and optimization. These applications were evaluated from a new perspective on simulation-based offline training and process demonstrations, policy deployment with transfer learning (TL) and the challenges of integrating it by proposing a f...

In this work an integrated simulation-optimization model to determine the optimal economic operating point of an industrial site for natural gas processing with multiple process units and feedstock. Natural gas industrial facilities have the intrinsic characteristic of dealing with dynamic scenarios, with no control over raw inlet gas, which might...

One of the difficulties in practical implementations of the classic Real-Time Optimization (RTO) strategy is the integration between optimization and control layers, mainly due to the differences between the models used in each layer, which may result in unreachable setpoints coming from optimization to the control layer. In this context, Economic...

Praziquantel (PZQ) is a racemic mixture prescribed for Schistosomiasis disease treatment. However, only the enantiomer (R)-PZQ presents demonstrated efficiency against infirmity; the (S)-PZQ causes sickness and vomiting due to its bitter taste. The simulated moving bed chromatography system (SMB) is a well-established separation process applied in...

COVID-19 pandemic response with non-pharmaceutical interventions is an intrinsic control problem. Governments weigh social distancing policies to avoid overload in the health system without significant economic impact. The mutability of the SARS-CoV-2 virus, vaccination coverage, and mobility restriction measures change epidemic dynamics over time....

Biomass ethanol production presents shortcomings related to enzymes cost and efficiency of xylose fermentation. Commercial enzymes used in second-generation ethanol (2GE) plants are produced off-site/offshore, with the additional formulation, transportation and storage costs. Alternatively, lower cost cellulases can be produced on-site. This study...

Crystallization is a separation and purification process relevant to industrial sectors, such as pharmaceuticals. The maximum possible recovery of solute amount is one of its goals, and the temperature profile is crucial to achieve this. In this work, neural networks-based models were developed to predict the solute concentration of a batch crystal...

Real-time optimization (RTO) is a model-based technique to drive process operation towards its optimal condition according to an objective function while respecting constraints. However, whenever the used model presents structural uncertainty, the classic two-step approach fails to find the true optimum condition. Recently, a Modifier Adaption (MA)...

Methanol is an important product in chemical industries, having many applications: solvent, fuel and mainly being a feedstock for a large number of industrial processes. As the search for a more sustainable synthesis process grows, the methanol synthesis loop from synthesis gas was studied. In this work, an optimization based on an environmental ob...

Reinforcement learning (RL) is an area of machine learning (ML), in which an agent learns a specific task through direct interaction with its environment with the objective of maximizing the expected cumulative reward. Based on the algorithm, an RL agent learns from the consequences of its actions, rather than being explicitly taught. In the field...

The interest in the simulated moving bed (SMB) technology lies in its variants. Some of them that have a high potential to increase the performance in enantioseparations are the ModiCon and the ModiCon+VariCol. These variants are based on the modulation of feed concentration and a combination of feed concentration and length of zones modulations. T...

In this paper, we present a novel methodology for nonlinear dynamic analysis of chemical processes that are posed as differential–algebraic equations (DAE) systems. With the proposed approach, for the first time, high-index systems, which are often the result of computer-aided modeling, can be treated “as is,” i.e., without the need for model refor...

Black oil delumping, also known as a stream conversion method, converts a black oil wellstream into a compositional wellstream. This procedure ensures consistent flowrate allocations and monitoring of well’s performance. This method requires volumetric oil and gas flowrates given in well-test reports, an equation of state model, and additional blac...

Lipases are enzymes that, in aqueous or non-aqueous media, act on water-insoluble substrates, mainly catalyzing reactions on carboxyl ester bonds, such as hydrolysis, aminolysis, and (trans)esterification. Yarrowia lipolytica is a non-conventional yeast known for secreting lipases and other bioproducts; therefore, it is of great interest in various...

A tuning procedure for a model predictive controller (MPC) is presented for multi-input multi-output systems. The approach consists of two steps based on a hybrid method: the goal attainment method and a variable neighborhood search. In the first step, the weights of the MPC objective function are obtained, minimizing the square error between the c...

An appropriate techno-economic assessment of biorefineries is essential for consolidating a bio-based economy. Herein, the effects of scale and seasonality on the design of a sugarcane to ethanol biorefinery were assessed using an extensive process simulation and economic modeling methodology. Four sub-processes were simulated for eight different s...

In this paper we present a study for the development of a predictive emission monitoring system (PEMS) through hybrid modeling. The system was built by coupling a thermodynamic equilibrium model based on the minimization of the Gibbs energy and machine learning (ML) models, such as artificial neural networks (ANN) with different architectures, supp...

COVID-19 pandemic response with non-pharmaceutical interventions is an intrinsic control problem. Governments balance social distancing policies to avoid overload on health system without major economic impact. A control strategy requires reliable predictions to be efficient on long-term. SARS-CoV-2 mutability, vaccination coverage and time-varying...

Reinforcement Learning (RL) arises from the set of Machine Learning techniques that are interesting for data-based process control purposes. Many authors discuss the advantages of coupling RL techniques to classical control frameworks, such as model predictive control (MPC), to tackle their drawbacks. In this work, an RL based on radial basis funct...

RESUMO-A supersaturação é uma das mais importantes variáveis em processos de cristalização, correspondendo a força motriz desse processo. O controle da supersaturação é uma estratégia amplamente utilizada nos processos de cristalização, utilizado como forma de balancear a ocorrência simultânea do crescimento de cristais e nucleação primária ou secu...

RESUMO-Os efeitos da pandemia de COVID-19 continuaram afetando o mundo no ano de 2021, de modo que a realização de testes em massa para diagnóstico da doença é útil no monitoramento do número de infecções e na tomada de decisão com relação a medidas restritivas. Testes de fácil aplicação e baixo custo são de vital importância. Portanto, o objetivo...

RESUMO-A abordagem de otimização em tempo real (RTO) híbrida combina o melhor de ambas as metodologias estática e dinâmica, isto é, a etapa de estimação de parâmetros ocorre fazendo uso de modelos dinâmicos e medidas transientes, enquanto que a etapa de otimização é realizada através de modelos estacionários. Trabalhos recentes indicam que esta met...

RESUMO-O conceito de Digital Twin (DT) compreende a reprodução digital de um componente, produto ou sistema real. Neste trabalho, o objetivo foi investigar o emprego de um DT na avaliação dinâmica do sistema de despressurização de emergência em uma unidade real de hidrotratamento de diesel-HDT, utilizando para isso o software Aspen HYSYS V10. O DT...

In processes with slow dynamics or subjected to frequent disturbances, the detection of a steady-state operation might be rare, which hinders the application of the classic RTO. To overcome this issue, the Hybrid RTO (HRTO) is a proposition where a dynamic observer substitutes the Steady-State Detection (SSD) and the static parameter estimation ste...

The rise of new digital technologies and their applications in several areas pushes the process industry to update its methodologies with more intensive use of mathematical models—commonly denoted as digital twins—and artificial intelligence (AI) approaches to continuously enhance operational efficiency. In this context, Real-time Optimization (RTO...

During middle distillates hydroprocessing, the relationship between feed properties, operating conditions and the deactivation phenomenon is crucial to achieve fuels specification. In this work, an accelerated deactivation methodology was employed to in short periods observe and study the deactivation phenomenon. For the stabilization step, 100 hou...

The cellulose dissolution is an essential pretreatment process for the chemical conversion of lignocellu-losic biomass into biofuels. Here, the dissolution of a 36-chain Ib cellulose model with a hexagonal cross-section (M36HCS) in water is analyzed by Molecular Dynamics (MD) with CHARMM36/TIP3P (C36/TIP3P) force field using gradual heating at 25 M...

The VariCol and ModiCon processes are two variants of the simulated moving bed (SMB) process, characterized by the modulation of the length of zones of the chromatographic column train and the feed concentration. These features give more flexibility than the conventional operation, leading to essential improvements in the separation and purificatio...

A correction to this paper has been published: https://doi.org/10.1007/s43153-021-00091-9

The flow rate values reported in real time, for an oil and gas production unit, refer to the total amount produced by that unit. On the other hand, data referring to the flow rate of each producing well in real time are usually not available, due to the difficulty in implementing flow measurement devices. These individual flows are determined by pr...

The flow rate values reported in real time, for an oil and gas production unit, refer to the total amount produced by that unit. On the other hand, data referring to the flow rate of each producing well in real time are usually not available, due to the difficulty in implementing flow measurement devices. These individual flows are determined by pr...

First-generation sugarcane ethanol production is a well-established technology. However, second-generation ethanol is not yet consolidated in industries. Under these circumstances, this work presents a multi-objective optimization of this process focusing on economic and environmental objectives. The impact of the Brazilian decarbonization program...

Regression analysis constitutes an important tool for investigating the effect of explanatory variables on response variables. When outliers and bias errors are present, the weighted least squares estimator can perform poorly. For this reason, alternative robust techniques have been studied in several areas of science. However, often these differen...

The natural gas produced in primary separation is passed through a compression system to be pressurized and conditioned before being sent to its final destination. The operation of that system needs to be safe and efficient to avoid equipment damage and reduce energy consumption. The stable and secure operation of the equipment in a compression sys...

New materials for the anode of solid oxide fuel cells are needed, aiming at greater flexibility in the use of fuels and increased performance. Y-doped SrTiO3 (YST) has shown great potential for this application due to its mixed electrical (ionic and electronic) conductivity. In this work, undoped (ST) and doped samples with 4 mol.% (YST04) and 8 mo...

Hybrid Real-Time Optimization (HRTO) approaches consist of the economical steady-state optimization performed after the dynamic adaptation of the process model, usually carried out by an Extended Kalman Filter. Despite the increasing number of works concerning this technique, the literature still lacks a larger number of case studies supporting the...

One of the main drawbacks of the so-called two-step approach in Real-time Optimization (RTO) is the long waits for stationary operation. To overcome this issue, a hybrid RTO (HRTO) approach has been proposed in the literature in which a dynamic estimation is carried out, followed by economic optimization. Despite generally presenting a good perform...

An optimal control framework was employed to obtain optimal supersaturation/temperature policies for controlling the crystal mass, size, and shape that meet target product specifications. It uses a bivariate population balance model that includes crystal nucleation, growth, dissolution, and disappearance. The optimal control scheme, solving a dynam...

The growing market economics demands the companies for even more efforts in the direction of an increase of operational efficiency, enabling the processes to produce more income with less cost in order to provide products with competitive prices. In this context, real-time optimization (RTO) is one of the tools of the new concept of industry 4.0 th...

There is a need to develop models that can accurately predict the individual flow rate of the producing wells, due to the lack of multiphase flow measuring equipment for each well in a platform. Thus, this paper shows a methodology to predict the liquid and gas flow rates of each well as a function of real-time process data, choke valve specificati...

The VariCol process is a variant of the conventional simulated moving bed (SMB) process, distinguished by the asynchronous shifting of the inlet and outlet ports of the chromatographic column train. This feature allows for a more flexible operation in column utilization and can also achieve higher separation performances. However, to take full bene...

Na produção de petróleo em ambientes offshore, o gás natural separado do óleo é comprimido, condicionado e pode apresentar três destinos: Exportação para unidades de processamento de gás natural (UPGNs), injeção em reservatório e uso em gas-lift. Para todos os casos listados, o gás natural deve ser comprimido para alcançar pressões de escoamento es...

The short-term oil production optimization of offshore platforms which use continuous gas lift and consider several operational constraints is a challenging task, specially when decision variables include both continuous and integer ones. The intrinsic non-linearity of the Gas Lift Performance Curves (GLPC) naturally yields to Mixer-Integer Nonline...

The divided wall column (DWC) can achieve sharp separations of three or more components in a single shell, substituting conventional sequences of two or more binary distillation columns, with lower expenses. Despite these advantages, DWC models are not available in commercial chemical process simulators. To simulate DWC, users must employ instances...

Electric submersible pumps (ESPs) are one of the most widespread oil artificial lifting technologies. In the operation of an ESP there are a large number of parameters that must be monitored and held within operational constraints in order to guarantee stable and optimal operation. Manual control is subject to sub-optimal production and constant vi...

Solid-state fermentation (SSF) with Yarrowia lipolytica requires appropriate medium supplementation and bioreactor configuration for efficient lipase production, which are related to the temperature profile due to the heat release by the metabolic activity and the heat removal mechanisms of the bioreactor. Thus, we evaluated supplementation with co...

According to the Brazilian federal program called Renovabio, biofuels producers will earn a number of carbon credits, known as CBios, proportionally to the reduction of greenhouse gases emissions. These CBios will be sold by biorefineries to fuel distributors in Brazil. Thus, the environmental performance will have a direct impact on the economic p...

This paper presents a comprehensive interface and electrochemical phenomena review of different anodes for SOFC. In this way, it was possible to select some mixed electronic and ionic conductor materials alternatives to the conventional anode, Ni-YSZ conductor. New materials must present a satisfactory electronic and ionic conductivity at intermedi...

Multiphase flow in petroleum pipelines became more challenging due to the emergence of deeper wells at extreme environmental conditions. Flow assurance strategies ensure the production of hydrocarbons uninterruptedly. In this work we propose a simple and straightforward framework for design strategies for flow assurance constraints, implementing th...

Beer-Lambert-Bouguer law is for a limiting case and, therefore, it is not useful to describe the relationship between absorption signal and enantiomer concentration in a stream when there are nonlinear phenomena present. In this work, the Chiral Detector (CD-2095 JASCO) equipment was used to measure simultaneously the UV–Vis and circular dichroism...

Front Cover: Cellulose amorphous‐paracrystalline structures of 36‐chains with 4‐10 glucose units per chain are obtained, and their thermophysical properties (glass‐transition, expansion, compressibility, and heat capacity) at different temperatures are investigated by Molecular Dynamics using the fluctuation method with CHARMM36 force field. The fi...

Although there are many studies in the literature concerning viscoelastic fluids, it remains a challenging subject in the field of polymer rheology to determine which is the most appropriate model and parameter set to describe the rheological behavior of a given viscoelastic fluid under real flow. The aim of this work is to present an optimization...

An experimental methodology for inferring brine dissolution rate in monoehtylene glycol (MEG) solutions at different temperatures using a webcam combined with a mathematical model is presented. The measurement system is designed to track the RGB (red, green, and blue) color variations during the dissolution process. A dynamic model augmented with t...

For the engineering and process design of chemical and pharmaceutical plants, the knowledge of thermophysical properties is essential. Here, glass transition temperature (Tg), curves of heat capacity (Cp), isotropic thermal expansion (∝p), and isothermal compressibility (βT) are computed for amorphous/paracrystalline (Am‐Par) structures of cellulos...

Taking advantage of both shape and chemical anisotropy on the same nanoparticle offers rich self-assembly possibilities for nanotechnology. Through Dissipative Particle Dynamics (DPD) calculations, in the present work, the directed assembly of Janus nanorod aggregates and their capability to assemble into metastable novel structures at an interfaci...