Contacts

Titles
Canada Research Chair (CRC) Tier II in Multiscale Modelling and Process Systems​
Associate Professor of Department of Chemical Engineering
Member, American Institute of Chemical Engineers (AIChE)
Member, Professional Engineers Ontario (PEO)
Member, Canadian Society for Chemical Engineering (CSChE)​

Email: laricard@uwaterloo.ca
Phone: +(519)-888-4567, x38667​
Location: E6 3014

Education
Ph.D., University of Waterloo, 2008.
MASc, Technological Institute of Celaya, 2000
BASc, Technological Institute of Orizaba, 1997

I received my bachelor's and master's degrees in chemical engineering from Shiraz University and the Sharif University of Technology, respectively. In the Fall of 2022, I had the opportunity to join Professor Luis Ricardez-Sandoval's research team as a Ph.D. student.

In my Ph.D., I am working on developing a dynamic model based on Dynamic Flux Balance (DFB) and its application for monitoring, optimization, and optimal control of mammalian-based bioprocesses. The dynamic metabolic flux model is a modeling paradigm that is gaining popularity in academia and industry since it explicitly considers the metabolic reactions of a given microorganism and can potentially improve process operations. However, the accuracy of DFB is often limited due to a lack of knowledge about the effect of operating conditions such as temperature, pH, and osmolarity on the metabolic response. As the metabolic models cannot typically deal with changes in pH and temperature, these effects will be introduced by empirical correlations between these variables and the parameters in the constraints of the DFB model in this project. This model will calibrate with data provided by Sartorius combined with shake flask data collected at Waterloo. After developing the DFB model, it will be optimized by the Run-to-Run optimization model based on DoE. This project is conducted in collaboration with Sartorius.

I completed my undergraduate degree in Chemical Engineering at the University of Waterloo in 2022. In my co-op experience, I worked in various fields and industries including oil and gas, medical devices, batteries and nanotechnology. However, the areas that have attracted my attention the most are process control, process modelling and projects involving computer simulations in Python and MATLAB. Hence, in my fourth year of undergrad, I started the accelerated master’s program at the University with Prof. Luis as my supervisor. Currently, I am a MASc. student and was awarded the Engineering Excellence Master’s Fellowship from the University of Waterloo for my high academic standing. 

So far, my research has focused on the application of Non-linear Model Predictive Control (NMPC) on large-scale chemical processes. NMPC is the state-of-the-art technology in process control. It involves the formulation of a dynamic mathematical model of a process in an optimization problem, where the optimal control actions are determined based on an objective function, subject to physical and process constraints. We have studied the application of scenario-based robust NMPC, together with state estimators, such as Extended Kalman Filter (EKF) and Moving Horizon Estimator (MHE), on the benchmark Tennessee-Eastman process. The main disadvantages of NMPC are that the optimization problem must be solved at every sampling step, and it only guarantees constraint satisfaction when the model is accurate. Robust and stochastic optimization aim to consider model uncertainty. Explicit NMPC seeks to solve the optimization problem offline to generate control laws that can be evaluated online, under much less computational costs, as direct functions of the state feedback. The research objectives are to develop a controller that can account for process uncertainty by using explicit control laws that can move the computational challenges offline. For this, we are exploring the potential of adjustable robust optimization (ARO), or reinforcement learning to train neural networks that can evaluate the control laws. 

Machine learning techniques have become an attractive alternative to solve optimization problems using historical data and simulations. Supervised and unsupervised learning, which can produce models for forecasting or analysis, have been useful tools for the development of mathematical models. On the cases where data is not available or when the system is subject to continuous changes, there is the alternative to use Deep Reinforcement Learning (DRL) methods. These methods aim to create intelligent agents that can take decisions and learn how to perform tasks, e.g., optimal schedules and control profiles. My research focuses on the implementation of DRL algorithms as optimization methods in chemical engineering scenarios, specifically scheduling problems under uncertainty on job-shop machine environments. The capacity of DRL agents to adapt decisions to changes and to be pre-trained with historical data (using supervised learning methods) makes them an attractive tool to handle this type of dynamic processes. On and off-policy algorithms like Deep Deterministic Policy Gradient (DDPG) and Proximal Policy Optimization (PPO) are state-of-art models that I use to perform these implementations. The objective is to create efficient DRL agents that can handle the effects of uncertainty in on-line scheduling processes.

I received my Ph.D. degree in 2019 from the Systems Engineering and Computer Science Program at Federal University of Rio de Janeiro.

I came to the University of Waterloo to work on a project involving an industrial partner from the analytical services sector. Since June 2021, I have been a postdoctoral researcher at the Faculty of Mathematics, working under the supervision of Prof. Fukasawa from the department of Combinatorics and Optimization, and Prof. Sandoval from the department of Chemical Engineering. The main goal of our project is to help our partner solve the scheduling problem for large-scale industrial production.

My research aims to propose a new model to integrate the personnel allocation and machine scheduling problem for large-scale industrial production. These two optimization problems have been studied independently in the literature and, there are limited studies integrating both problems and presenting results for large-scale instances. I have been working with different optimization techniques applied to solve these real-world problems. My challenge is understanding the customer's problems and presenting a mathematical formulation to solve them. 

I earned a BASc and an MASc in chemical engineering from Universidad de Los Andes (Bogotá, Colombia) in 2020. During my MASc I studied unconventional strategies to handle discrete decisions in Mixed-Integer Nonlinear Programming (MINLP). The optimal design of a Reactive Distillation (RD) column was used as the main case study. The key outcome of the research during my MASc was the development of a reformulation and an algorithm that allowed to optimize the discrete decisions of the RD system, e.g., number of stages, more efficiently than traditional MINLP solvers. I also have applied research experience with an industrial partner in the food industry, developing simulation and optimization solutions for the conceptual design and operation of a production process of baked products. 

Professor Luis Ricardez-Sandoval co-supervised my MASc research, and I joined his group in Fall 2020 to pursue a PhD. Currently, I am developing and testing deterministic optimization algorithms that efficiently handle ordered discrete decisions in MINLP and Generalized Disjunctive Programming (GDP) problems. The chemical engineering applications of these algorithms include optimal design and control, and optimal scheduling and control. 

Catalytic flow reactors play an important role in chemical processes, and as a result there has been substantial interest toward reactor development and optimization. The behaviour of catalytic reactors depends on the macroscopic reactor design as well as the microscopic events taking place on the catalyst surface. As a result, both scales must be accounted for while modeling and designing new and optimal reactors. However, this is significantly impeded by uncertainties present in the reactor’s kinetic mechanisms and system parameters. In catalytic flow reactors, uneven catalyst fouling produces spatially-varying uncertainties in the reaction rates along the reactor length.  My research focuses on the design of robust optimization and control studies for the catalytic flow reactor to identify the values of key design and operation parameters that maximize the reactor performance and minimize the effects of uncertainty. This is accomplished through the use of reduced-order models such as Power Series Expansions (PSEs), Polynomial Chaos Expansions (PCEs), and data-driven models (DDMs) to efficiently propagate spatially-varying uncertainty through the catalytic reactor model. The objective is to address the effects of uncertainty while designing control and robust optimization processes for transient and steady-state multiscale catalytic flow reactors.

Thesis: Uncertainty Analysis and Robust Optimization of a Single Pore in a Heterogeneous Catalytic Flow Reactor System​​

I received an BASc in chemical engineering at the University of Toronto in 2017 and an MSc in process systems engineering at Imperial College London in 2018. My research aims to design algorithms for the real-time economic optimization of process systems under uncertainty.

Real-time optimization (RTO) and modifier adaptation (MA) are the approaches used in the cases of parametric and structural model uncertainty, respectively, for online steady-state economic optimization. The two-step RTO approach estimates and updates uncertain parameters prior to solving an economic optimization problem. In contrast, the MA approach uses bias and gradient-correction terms to modify structurally uncertain process models. For the two-step approach, we have  developed a scheme for accurate estimation in noisy environments. For the MA approach, we have developed a partial adaptation algorithm that has improved computational efficiency as to provide economic benefits to frequently-disturbed systems.

I am also broadly interested in sustainability and the use of economic incentives to operate sustainable processes efficiently. Through the use of online economic optimization, we have developed such schemes for post-combustion carbon capture and recirculating aquaculture systems.

Google Scholar Profile

I received my BASc in Chemical Engineering at the University of Waterloo in 2021. As an undergraduate student, I was interested in modelling and optimization, and joined the group in 2018 through the URA program. My research focuses on the modelling and process control of an emerging carbon capture process known as Chemical Looping Combustion (CLC). In CLC, the air and fuel are kept separate through the addition of a metal-based oxygen carrier (OC), inherently separating the CO₂ into a concentrated stream with no oxygen or nitrogen. I am applying nonlinear model predictive control (NMPC) in order to ensure the system is efficient, producing a concentrated CO₂ stream while generating adequate energy. I have also been investigating the use of biomass as a fuel in this process, which would allow for combustion to be performed with net-negative CO₂ emissions.

Publications :

Toffolo, K., Ricardez-Sandoval, L. (2021, June 13-16). Optimal Design and Control of a Multiscale Model for a Packed Bed Chemical-Looping Combustion Reactor. 11th IFAC Symposium on Advanced Control of Chemical Processes, Fully Virtual.

Valipour, M., Toffolo, K. M., and Ricardez-Sandoval, L. A. (2021). State estimation and sensor location for Entrained-Flow Gasification Systems using Kalman Filter. Control Engineering Practice, 108(104702).

I received my BASc in Chemical Engineering from the Benemerita Universidad Autonoma de Puebla (Mexico) in 2010, a MSc degree in Chemical Engineering from Universidad Iberoamericana (Mexico), and currently pursuing a PhD degree at the University of Waterloo. My research is focused on the integration of design and model-based control under uncertainty. I implement modern control strategies such as nonlinear model predictive control (NMPC). Multiple challenges arise for the solution of the corresponding optimization formulation for design and NMPC-based control:  a) a multilevel formulation (embedded optimization problems), b) introduction of uncertainties and disturbances, c) implementation of structural decision using integer variables. In particular, I have implemented two different strategies for the solution of the multilevel formulation: (i) a decomposition strategy based on the determination of the back-off that allows to determine a dynamically feasible process design [1], and (ii) a direct strategy based on the transformation of the multilevel formulation into a single-level problem using a classical KKT transformation [2]. On the other hand, the introduction of structural decisions increases the complexity of the problem. To circumvent this challenge, I have implemented a discrete-steepest descent algorithm to determine the optimal structure, process design, and control.

[1] Palma-Flores, O., & Ricardez-Sandoval, L. A. (2022). Simultaneous design and nonlinear model predictive control under uncertainty: A back-off approach. Journal of Process Control, 110, 45-58.

[2] Palma-Flores, O., & Ricardez-Sandoval, L. A. (2022). Integration of design and NMPC-based control for chemical processes under uncertainty: An MPCC-based framework. Computers & Chemical Engineering, 162, 107815.

 I started at McMaster University in 2018 and graduated from the Chemical Engineering program in the June of 2022 when I received my B.Eng. I have since continued my studies at Waterloo University pursuing my MASc in sustainable energy systems optimization. 

    I am currently in the process of reviewing Canada’s current position on carbon capture and developing an optimal solution for how we can reach our 2030 goal of 40% emissions reduction proposed by the Federal Government earlier this year (2022). With this review and optimization we will be able to provide a comprehensive basis for the progress that has taken place in carbon capture research and projects. 

    The main focus of my work is on post-combustion carbon capture systems. For my thesis I plan to development, simulate, and optimize an alternative carbon capture process for post-combustion implementation. Although there are many different avenues and methods in which carbon capture can be implemented, post-combustion is still the most common due to its flexibility and retrofitting capabilities. With this process it is my goal to supply the industry with a more feasibly and efficient way to capture CO₂ so that more commercial scale projects can occur in order to continue to making our industries more sustainable. 

I got my BSc degree in Molecular Biology & Biotechnology from University of Liverpool and Xi’an Jiao Tong Liverpool University. I had my MRes degree in System & Synthetic Biology from Imperial College London. After graduation, I was working in the RD department of the Biopharmaceutic industry in China. Then I came to PhD project in Process System Engineering at the University of Waterloo with the supervision of Professor Hector Budman and Professor Luis Ricardez-Sandoval.

My research focuses on model-based optimization of scale-up/down operations in the biopharmaceutic industry. We are using CFD (computational fluid dynamics) techniques to simulate the flow pattern, local gradient, shear-sensitive zone, and biokinetics in small- and large-scale bioreactors. Meanwhile, we are developing a predictive model based on actual experimental data. AI algorithm, PINN (Physical-Informed Neuro Network), is our later objective for a surrogate model in parameter optimization and Advanced Process Control of bioreactor.

I received my Bachelor in science in chemical engineering at the Universidad Iberoamericana in 2022 and got the amazing opportunity to continue my graduate studies at the University of Waterloo, doing a Master's in science in process system engineering (PSE). My research aims to implement novel Machine Learning techniques such as Reinforcement Learning in dynamic optimization processes. 

I received my MASc degree in Chemical Engineering department in University of Waterloo in 2019 then continued my research as a PhD student under the supervision by Prof. Luis Ricardez-Sandoval and Prof. David Simakov. During my MASc, I focused on carbon dioxide methanation reaction over multi-metallic transition metal-based thermo-catalysts (Co/Mo oxide and carbide). During my PhD, I developed K-promoted Fe-based catalysts by the newly developed modified reverse microemulsion method for direct conversion of carbon dioxide into hydrocarbons, which showed decent conversion of carbon dioxide and over 50% selectivity to light hydrocarbons (C2-C4 paraffins and olefins), providing a potential for direct carbon dioxide conversion to valuable hydrocarbons. Currently, my research is on investigation of a potential catalyst(s) for Reverse Water Gas Shift, metal doped ceria-based catalysts, by the combination of Density Functional Theory (DFT) method and Machine Learning algorithm, which aim for optimal materials for carbon dioxide hydrogenation.​