arXiv subjects

Paulina Quintanilla

Publications and source records attributed to Paulina Quintanilla.

7 recordsLinked to original sources

arXiv2026

Hierarchical Control via MPC-RL for Multi-Timescale Battery Systems

Multi-timescale systems present a fundamental challenge, where fast operational decisions must coexist with long-horizon sustainability targets. In this work, we propose a new hierarchical control framework via Model Predictive Control (MPC) and Reinforcement Learning (RL) to separate decision-making on two distinct timescales. The high-level MPC optimizes long-horizon setpoints at the slow dynamic and on a fast timescale, a low-level pretrained RL agent tracks these setpoints in real time to maximize short-term objectives. RL is introduced to learn nonlinear control policies, without relying on model linearizations or requiring the heavy online computation from solving repeated optimal control problems. The framework is applied to a Battery Energy Storage System (BESS) operating in frequency regulation markets to balance fast profit opportunities (seconds) and slow battery degradation (weeks to months). The design employs a degradation-aware RL agent trained offline to generate safe long-horizon setpoints, and a degradation-unaware agent fine-tuned from it for fast runtime setpoint tracking. Compared to MPC baselines, the proposed approach successfully extends battery lifetime by 84% and increases operational profit by 34%.

eess.SY↗

arXiv2026

DASyR-LLM: Domain-Aware Symbolic Regression with LLMs for Kinetic Model Discovery

Kinetic model discovery is a central challenge in chemical engineering, as accurate rate expressions are essential for understanding and controlling chemical and biological processes. Symbolic regression (SR) has emerged as a powerful data-driven approach for identifying interpretable kinetic models, but usually operates without domain knowledge, often exploring physicochemically implausible models. Large language models (LLMs) offer a promising avenue for injecting domain expertise into this search. Here, we introduce an LLM-guided SR framework, embedding an LLM module within an iterative SR algorithm for automated kinetic model discovery. The LLM performs two roles at each iteration: (1) a qualitative physicochemical critique of the best SR candidates, and (2) the proposal of new candidate rate expressions guided by the SR-generated models and embedded chemical knowledge. Our framework is evaluated on four in silico case studies of increasing complexity, spanning heterogeneous catalysis and bioprocess systems. Results show the LLM-guided framework reduces iterations to identify the ground-truth model by $41.7-79.3\%$ versus a state-of-the-art SR framework, with the LLM directly proposing the correct model structure in over half of the guided runs. In practical settings, where each iteration typically requires a new wet-lab experiment, this translates into a substantial reduction in experimental effort. Predictive performance on an independent validation set is equivalent between both approaches, with $R^2>0.98$ in all case studies. Ablation studies indicate that both the SR component and the LLM scale contribute to this performance, with a reduced-size LLM largely retaining discovery efficiency. These findings demonstrate that LLMs can effectively inject domain knowledge into scientific model discovery, paving the way toward fully automated, domain-aware kinetic modelling pipelines.

cs.LG↗

arXiv2026

From data to decisions: Bayesian modelling and global sensitivity analysis for flotation control

This work presents a data-driven framework for interpretable modelling and decision support in flotation systems, integrating Gaussian Process (GP) regression with Global Sensitivity Analysis (GSA) via Sobol indices and local interpretability using SHapley Additive exPlanations (SHAP). Based on laboratory-scale experimental data, a static GP surrogate model is developed to capture how superficial air velocity, overflowing froth velocity, froth height over the lip, pulp height, bubble size, and tailings flowrate influence the measured air recovery. The trained GP enables the computation of Sobol indices to quantify the contribution of each variable and their interactions to the overall variance in air recovery. The combination of Bayesian inference and Sobol-based sensitivity metrics provides a systematic approach to identify the dominant and interacting variables governing air recovery. This study links Bayesian learning, sensitivity quantification, and explainability to provide a foundation for data-driven control and optimisation of flotation processes.

eess.SY↗

arXiv2025

Digital twin with automatic disturbance detection for an expert-controlled SAG mill

This study presents the development and validation of a digital twin for a semi-autogenous grinding (SAG) mill controlled by an expert system. The digital twin integrates three key components of the closed-loop operation: (1) fuzzy logic for expert control, (2) a state-space model for regulatory control, and (3) a recurrent neural network to simulate the SAG mill process. The digital twin is combined with a statistical framework for automatically detecting process disturbances (or critical operations), which triggers model retraining only when deviations from expected behaviour are identified, ensuring continuous updates with new data to enhance the SAG supervision. The model was trained with 68 hours of operational industrial data and validated with an additional 8 hours, allowing it to predict mill behaviour within a 2.5-minute horizon at 30-second intervals with errors smaller than 5%.

eess.SY↗

arXiv2024

Experimental implementation of an economic model predictive control for froth flotation

We present the implementation of a novel economic model predictive control (E-MPC) strategy for froth flotation, the largest tonnage mineral separation process. A previously calibrated and validated dynamic model incorporating froth physics was used, which overcomes the limitations of previous simplified models reported in the literature. The E-MPC's optimal control problem was solved using full discretization with orthogonal collocation over finite elements, employing automatic differentiation via CasADi. This approach was applied in a 30-litre laboratory-scale flotation cell, significantly improving mineral recovery from 9% to 29% under feed flowrate disturbances while maintaining a minimum concentrate grade of 20%.

eess.SY↗

arXiv2024

Digital twin with automatic disturbance detection for real-time optimization of a semi-autogenous grinding (SAG) mill

This work describes the development and validation of a digital twin for a semi-autogenous grinding (SAG) mill controlled by an expert system. The digital twin consists of three modules emulating a closed-loop system: fuzzy logic for the expert control, a state-space model for regulatory control, and a recurrent neural network for the SAG mill process. The model was trained with 68 hours of data and validated with 8 hours of test data. It predicts the mill's behavior within a 2.5-minute horizon with a 30-second sampling time. The disturbance detection evaluates the need for retraining, and the digital twin shows promise for supervising the SAG mill with the expert control system. Future work will focus on integrating this digital twin into real-time optimization strategies with industrial validation.

eess.SY↗

arXiv2024

Grey-box Recursive Parameter Identification of a Nonlinear Dynamic Model for Mineral Flotation

This study presents a grey-box recursive identification technique to estimate key parameters in a mineral flotation process across two scenarios. The method is applied to a nonlinear physics-based dynamic model validated at a laboratory scale, allowing real-time updates of two model parameters, n and C, in response to changing conditions. The proposed approach effectively adapts to process variability and allows for continuous adjustments based on operational fluctuations, resulting in a significantly improved estimation of concentrate grade - one key performance indicator. In Scenario 1, parameters n and C achieved fit metrics of 97.99 and 96.86, respectively, with concentrate grade estimations improving from 75.1 to 98.69 using recursive identification. In Scenario 2, the fit metrics for n and C were 96.27 and 95.48, respectively, with the concentrate grade estimations increasing from 96.27 to 99.45 with recursive identification. The results demonstrate the effectiveness of the proposed grey-box recursive identification method in accurately estimating parameters and predicting concentrate grade in a mineral flotation process.

eess.SY↗