44 https://uvadoc.uva.es/handle/10324/1168 2026-04-10T05:26:30Z 2026-04-10T05:26:30Z Pulido, Belarmino Zamarreño Cosme, Jesús María Merino, Alejandro Bregón Bregón, Aníbal https://uvadoc.uva.es/handle/10324/83818 2026-04-08T06:36:43Z 2019-01-01T00:00:00Z

Reliable and timely fault detection and isolation are necessary tasks to guarantee continuous performance in complex industrial systems, avoiding failure propagation in the system and helping to minimize downtime. Model-based diagnosis fulfils those requirements, and has the additional advantage of using reusable models. However, reusing existing complex non-linear models for diagnosis in large industrial systems is not straightforward. Most of the times, the models have been created for other purposes different from diagnosis, and many times the required analytical redundancy is small. The approach proposed in this work combines techniques from two different research communities within Artificial Intelligence: Model-based Reasoning and Neural Networks. In particular, in this work we propose to use Possible Conflicts, which is a model decomposition technique from the Artificial Intelligence community to provide the structure (equations, inputs, outputs, and state variables) of minimal models able to perform fault detection and isolation. Such structural information is then used to design a grey box model by means of state space neural networks. In this work we prove that the structure of the Minimal Evaluable Model for a Possible Conflict can be used in real-world industrial systems to guide the design of the state space model of the neural network, reducing its complexity and avoiding the process of multiple unknown parameter estimation in the first principles models. We demonstrate the feasibility of the approach in an evaporator for a beet sugar factory using real data. Keywords: Intelligent fault diagnosis; Model-based diagnosis; State space neural networks; Data-driven modelling; Grey-box models.

2019-01-01T00:00:00Z Beneyto Rodríguez, Alicia Sáinz Palmero, Gregorio Ismael Galende Hernández, Marta Fuente Aparicio, María Jesús de la Cuenca de la Cruz, José María https://uvadoc.uva.es/handle/10324/83810 2026-03-25T20:01:31Z 2026-01-01T00:00:00Z

Water reuse is a key point when fresh water is a commodity in ever greater demand, yet also becoming more accessible. Furthermore, the return of clean water to its natural environment is also mandatory. Therefore, wastewater treatment plants (WWTPs) are essential in any policy focused on these serious challenges. WWTPs are complex facilities which need to operate at their best to achieve their goals. Nowadays, they are largely monitored, generating large databases of historical data concerning their functioning over time. All this implies a large amount of embedded information which is not usually easy for plant managers to assimilate, correlate and understand; in other words, for them to know the global operation of the plant at any given time. At this point, the intelligent and Machine Learning (ML) approaches can give support for that need, managing all the data and translating them into manageable, interpretable and explainable knowledge about how the WWTP plant is operating at a glance. Here, an eXplainable Artificial Intelligence (XAI) based methodology is proposed and tested for a real WWTP, in order to extract explainable service knowledge concerning the operation modes of the WWTP managed by AQUAVALL, which is the public service in charge of the integral water cycle in the City Council of Valladolid (Castilla y León, Spain). By applying well-known approaches of XAI and ML focused on the challenge of WWTP, it has been possible to summarize a large number of historical databases through a few explained operation modes of the plant in a low-dimensional data space, showing the variables and facility units involved in each case

2026-01-01T00:00:00Z Rivero Contreras, Rogelio Emilio Zamarreño Cosme, Jesús María Tadeo, Fernando https://uvadoc.uva.es/handle/10324/81239 2026-04-07T11:21:38Z 2025-01-01T00:00:00Z

Este artículo resume diversas iniciativas para implementar el control predictivo (MPC) en controladores lógicos programables (PLCs), a partir de la experiencia acumulada, algoritmos de MPC utilizados, métodos de optimización, sistemas de proceso considerados y los estándares de programación y marcas comerciales de dispositivos PLCs empleadas. Los estudios demuestran la viabilidad de implementar algoritmos de MPC clásicos junto con métodos de optimización en forma embebida. Destacan, en particular, los algoritmos de optimización como el método de Hildreth y qp OASES simplificado, que han demostrado ser eficientes y codificables según el estándar IEC 61131-3, ampliamente utilizado en estos dispositivos. Además, los estudios resaltan la necesidad de reducir los requerimientos de memoria y cálculos computacionales, y que esto permita escalar estos algoritmos desde simulaciones hardware-in-the-loop (HiL) y procesos a escala de laborarorio hacia plantas industriales. Las tendencias actuales se orientan hacia la simplificación del uso de recursos computacionales en los PLCs, la mejora de los algoritmos de MPC y optimización, y la integración de estos algoritmos en dispositivos modernos basados en internet de las cosas (PLCs-IoT).

2025-01-01T00:00:00Z López Muñoz, Paola Capellán Pérez, Iñigo Carpintero Redondo, Óscar https://uvadoc.uva.es/handle/10324/80495 2025-12-16T20:01:17Z 2025-01-01T00:00:00Z

Well-being impacts of climate change, particularly on human integrity (Health) and living conditions (Place), are severe but often underrepresented in Integrated Assessment Models (IAMs). When included, these impacts are typically modelled using simplistic top-down approaches, while bottom-up representations linking hazards to impacts, which offer high transparency and process detail, are largely overlooked. Recent trends connecting IAMs with the Impact, Adaptation, and Vulnerability (IAV) community offer an opportunity to improve the representation of well-being damages. Here, we conduct a scoping review resulting in a mapping of 37 modelling studies, revealing a diverse range of approaches, with variation in hazards, impacts, and modelling choices. Key gaps include weak representation of inequality, a lack of multi-channel assessments, and an overrepresentation of northern regions. We propose a roadmap to enhance climate impacts representation on Health and Place in IAMs, using improved data and large-scale multiregional models to generate results that better support decision-making.

2025-01-01T00:00:00Z Zamarreño Cosme, Jesús María Ríos, Juan C. Alonso, Gabriel https://uvadoc.uva.es/handle/10324/79621 2025-12-18T08:14:15Z 2025-01-01T00:00:00Z

Virtual and remote laboratories are excellent tools to complement real on-site laboratories. This paper presents the development of a virtual and a remote laboratory to support and enhance the students’ experience when dealing with a real pilot plant for control engineering education. Both labs, virtual and remote, are developed in Easy Java/JavaScript Simulations (EjsS) with similar interfaces and equivalent tasks as the real lab. The evaluation of the complementary labs through a student survey is presented at the end of the paper to validate the usefulness of this approach.

2025-01-01T00:00:00Z Pulido Junquera, José Belarmino Zamarreño Cosme, Jesús María Merino, Alejandro Bregón Bregón, Aníbal https://uvadoc.uva.es/handle/10324/79620 2026-04-08T06:38:19Z 2019-01-01T00:00:00Z

Reliable and timely fault detection and isolation are necessary tasks to guarantee continuous performance in complex industrial systems, avoiding failure propagation in the system and helping to minimize downtime. Model-based diagnosis fulfils those requirements, and has the additional advantage of using reusable models. However, reusing existing complex non-linear models for diagnosis in large industrial systems is not straightforward. Most of the times, the models have been created for other purposes different from diagnosis, and many times the required analytical redundancy is small. The approach proposed in this work combines techniques from two different research communities within Artificial Intelligence: Model-based Reasoning and Neural Networks. In particular, in this work we propose to use Possible Conflicts, which is a model decomposition technique from the Artificial Intelligence community to provide the structure (equations, inputs, outputs, and state variables) of minimal models able to perform fault detection and isolation. Such structural information is then used to design a grey box model by means of state space neural networks. In this work we prove that the structure of the Minimal Evaluable Model for a Possible Conflict can be used in real-world industrial systems to guide the design of the state space model of the neural network, reducing its complexity and avoiding the process of multiple unknown parameter estimation in the first principles models. We demonstrate the feasibility of the approach in an evaporator for a beet sugar factory using real data.

2019-01-01T00:00:00Z Duque Domingo, Jaime Caccavale, Riccardo Finzi, Alberto Zalama Casanova, Eduardo Gómez García-Bermejo, Jaime https://uvadoc.uva.es/handle/10324/79353 2025-12-16T20:01:16Z 2025-01-01T00:00:00Z

We present a framework that enables a collaborative robot to rapidly replicate structured manipulation tasks demonstrated by a human operator through a single 3D video recording. The system combines object segmentation with hand and gaze tracking to analyze and interpret the video demonstrations. The manipulation task is decomposed into primitive actions that leverage 3D features, including the proximity of the hand trajectory to objects, the speed of the trajectory, and the user’s gaze. In line with the One-Shot Learning paradigm, we introduce a novel object segmentation method called SAM+CP-CVV, ensuring that objects appearing in the demonstration require labeling only once. Segmented manipulation primitives are also associated with object-related data, facilitating the implementation of the corresponding robotic actions. Once these action primitives are extracted and recorded, they can be recombined to generate a structured robotic task ready for execution. This framework is particularly well-suited for flexible manufacturing environments, where operators can rapidly and incrementally instruct collaborative robots through video-demonstrated tasks. We discuss the approach applied to heterogeneous manipulation tasks and show that the proposed method can be transferred to different types of robots and manipulation scenarios.

2025-01-01T00:00:00Z Galende Hernández, Marta Sáinz Palmero, Gregorio Ismael Tarrero Fernández, Ana Isabel Duque Domingo, Jaime Giménez Olavarría, Blanca https://uvadoc.uva.es/handle/10324/78416 2025-10-07T19:00:59Z 2025-01-01T00:00:00Z

Integrating emerging Artificial Intelligence (AI) tools into the teaching of Industrial Engineering courses has become a crucial aspect of the university environment. This requires both the continuous training of faculty members and the integration of new AI-based tools into teaching, as well as careful consideration of the significant ethical and social impact of these technologies. To address these objectives, the University of Valladolid (Spain) has funded an Innovative Educational Project at the Industrial Engineering School, with the participation of five different technological departments involved in its engineering degrees. This paper presents the main results and conclusions obtained during the initial phases of this project, which have primarily focused on training lecturers in the use of AI tools to effectively integrate them into their teaching methodologies and to develop new educational materials. Specifically, AI tools are being used to generate questionnaires for student self-assessment, based on the content covered in each session. A key finding is the necessity of encouraging students to develop a strong critical mindset when using AI tools, particularly in analysing the reasoning behind AI-generated solutions. It has been observed that AI-driven self-assessment is particularly beneficial for theoretical knowledge, although students are guided to critically evaluate the AI's output, especially in problem-solving, where errors in intermediate steps can occur. The paper also presents specific examples related to systems and automation engineering using different AI tools and analysing their responses collaboratively with students to identify their strengths and weaknesses, highlighting both the potential and the current limitations of AI tools in these practical domains. To summarize the main conclusions derived from this study, it is essential to encourage students to develop a strong critical mindset when evaluating responses provided by AI-based tools. To achieve this, it is important to allow and encourage the use of such tools in the classroom, guiding students in identifying inconsistencies in AI-generated texts or results, and comparing them with other sources of knowledge. Moreover, it is a priority for lecturers to stay continuously updated on advancements in AI-based tools. Anticipating the impact of these tools on teaching is crucial, as their development and applications are constantly evolving and improving. Furthermore, with the gradual integration of AI into teaching activities, future engineers will be well-positioned to adapt to and lead technological changes in their industrial careers. Initiatives and experiences like this study will help both lecturers and students in their daily activities, allowing them to adapt to constant technological changes and helping them become better professionals in the not-so-distant future.

2025-01-01T00:00:00Z Rivero Contreras, Rogelio Emilio Merino, Alejandro Zamarreño Cosme, Jesús María Vilas, Carlos Tadeo Rico, Fernando Juan https://uvadoc.uva.es/handle/10324/75219 2025-03-04T20:00:56Z 2024-01-01T00:00:00Z

En este artículo se describe una propuesta de entorno virtual para aplicación en la evaluación de estrategias de control en controladores lógicos programables (PLC). Para ello, se ha considerado el uso de un gestor de la simulación del proceso, el cual comanda el modelo digital, y un gestor de comunicación de componentes que integra los elementos del entorno virtual (PLC virtual y modelo digital) a través del protocolo de comunicación industrial OPC UA. El entorno virtual se ha validado mediante un caso de estudio que considera como sistema de proceso una unidad de esterilización de alimentos envasados, el cual requiere de lógicas de control discretas y continuas para llevar a cabo su operación. Esto permite verificar las lógicas de control en el entorno virtual, en un ambiente de simulación, para su posterior su escalabilidad al proceso real utilizando protocolos adecuados para dispositivos de campo.

2024-01-01T00:00:00Z El Fezazi, Nabil Fahim, Mohamed Idrissi, Said Farkous, Rashid Álvarez Álvarez, María Teresa Tissir, El Houssaine https://uvadoc.uva.es/handle/10324/75158 2025-02-27T20:01:16Z 2024-01-01T00:00:00Z

In this paper, the finite-time stability and stabilization of nonlinear systems with delays is studied, via a Takagi–Sugeno approach. By using a novel Lyapunov–Krasovskii functional and introducing some fuzzy free-weighting matrices, sufficient conditions are derived, for bounded and differentiable time-varying delays in terms of an upper bound of the delay derivatives. Then, we achieve closed-loop stabilization in finite time through an efficient parallel distributed compensation design. The sufficient conditions are formulated as linear matrix inequalities to achieve the desired performance. Finally, the proposed methodology is applied to various case studies, highlighting its significance.

2024-01-01T00:00:00Z Rey Diez, María Isabel Galende Hernández, Marta Sáinz Palmero, Gregorio Ismael Fuente Aparicio, María Jesús de la https://uvadoc.uva.es/handle/10324/75071 2025-03-06T13:55:48Z 2013-01-01T00:00:00Z

Fuzzy modeling is one of the best known techniques to model systems and processes. In most cases, as in data-driven fuzzy modeling, these fuzzy models reach a high accuracy, but show poor performance in complexity or interpretability, which are key aspects of Fuzzy Logic. There are several approaches in the literature to deal with the complexity and interpretability challenges for fuzzy rule based systems (FRBSs). In this paper, a post-processing approach is proposed via a genetic rule selection based on the relevance of each rule (using Orthogonal Transformations (OTs), in this case P-QR) and the well-known accuracy-interpretability trade-off. The main objective is to check the true significance, drawbacks and advantages of the rule selection based on OTs to manage the accuracy-interpretability trade-off. In order to achieve this aim, a neuro-fuzzy system (FasArtFuzzy Adaptive System ART based) and several case studies from the KEEL Project Repository are used to tune and check this selection of rules based on rule relevance by OTs, genetic selection and accuracy-interpretability trade-off. This neuro-fuzzy system generates Mamdani FRBSs, in an approximate way. SPEA2 is the multi-objective evolutionary algorithm (MOEA) tool used to tune the proposed rule selection, and different interpretability measures have been considered.

2013-01-01T00:00:00Z Fuente Aparicio, María Jesús de la Sáinz Palmero, Gregorio Ismael Galende Hernández, Marta https://uvadoc.uva.es/handle/10324/75068 2025-02-17T20:00:58Z 2023-01-01T00:00:00Z

Decentralized monitoring methods, which divide the process variables into several blocks and perform local monitoring for each sub-block, have been gaining increasing attention in large-scale plant-wide monitoring due to the complexity of their processes. In such methods, the dynamic nature of the process data is a relevant issue which is not usually managed. Here, a new data-driven distributed dynamic monitoring scheme is proposed to deal with this issue, integrating regression to automatically divide the blocks, a multivariate and dynamic statistical analysis (Canonical Variate Analysis, CVA) to perform local monitoring, and Bayesian inference to achieve the decision making. By constructing sub-blocks using regression, it is possible to identify the most commonly associated variables for every block. Three regression methods are proposed: LASSO (Least Absolute Shrinkage and Selection Operator), which forces the coefficients of the less relevant variables towards zero; Elastic-net, a robust method that is a compromise between Ridge and Lasso regression; and, finally, a non-linear regression method based on the Multilayer Perceptron Network (MLP). Then, the CVA model is implemented for each sub-block to consider the dynamic characteristics of the industrial processes and the Bayesian inference provides a global decision for fault detection. The Tennessee Eastman benchmark validates the efficiency and feasibility of the proposed method regarding some state-of-the-art methods.

2023-01-01T00:00:00Z Reñones Domínguez, Aníbal Galende Hernández, Marta https://uvadoc.uva.es/handle/10324/74895 2025-02-13T20:00:58Z 2020-01-01T00:00:00Z

Practical research in AI often lacks of available and reliable datasets so the practitioners can try different algorithms. The field of predictive maintenance is particularly challenging in this aspect as many researchers don't have access to full-size industrial equipment or there is not available datasets representing a rich information content in different evolutions of faults. In this paper, it is presented a dataset with evolution of typical faults (commutator, winding and brush wear) in inexpensive DC motors under extensive monitoring (vibration, temperature, voltage, current and noise). These motors exhibit a particularly short useful life when operating out of nominal conditions (from 30 minutes to 6 hours) which make them very interesting to test different signal processing algorithms and introduce students and researchers into signal processing, fault detection and predictive maintenance. The paper explains in detail the experimentation and the structure of the real, un-processed, dataset published in the AI4EU platform with the aim of complying with the FAIR principle so the dataset is Findable, Accessible, Interoperable and Reusable.

2020-01-01T00:00:00Z Gacto, María José Galende Hernández, Marta Alcalá, Rafael Herrera, Francisco https://uvadoc.uva.es/handle/10324/74723 2025-01-31T20:01:20Z 2013-01-01T00:00:00Z

This paper addresses the challenging problem of fuzzy modeling in high-dimensional and large scale regression datasets. To this end, we propose a scalable two-stage method for obtaining accurate fuzzy models in high-dimensional regression problems using approximate Takagi-Sugeno-Kang Fuzzy Rule-Based Systems. In the first stage, we propose an effective Multi-Objective Evolutionary Algorithm, based on an embedded genetic Data Base learning (involved variables, granularities and a slight lateral displacement of fuzzy partitions) together with an inductive rule base learning within the same process. The second stage is a post-processing process based on a second MOEA to perform a rule selection and a fine scatter-based tuning of the Membership Functions. Moreover, it incorporates an efficient Kalman filter to estimate the coefficients of the consequent polynomial functions in the Takagi-Sugeno-Kang rules. In both stages, we include mechanisms in order to significantly improve the accuracy of the model and to ensure a fast convergence in high-dimensional regression problems. The proposed method is compared to the classical ANFIS method and to a well-known evolutionary learning algorithm for obtaining accurate TSK systems in 8 datasets with different sizes and dimensions, obtaining better results.

2013-01-01T00:00:00Z Rey Diez, María Isabel Galende Hernández, Marta Sáinz Palmero, Gregorio Ismael Fuente Aparicio, María Jesús de la https://uvadoc.uva.es/handle/10324/74718 2025-03-06T13:55:08Z 2011-01-01T00:00:00Z

Fuzzy modeling is one of the most known and used techniques in different areas to emulate the behavior of systems and processes. In most cases, as in data-driven fuzzy modeling, these fuzzy models reach a high performance from the point of view of accuracy, but from other points of view, such as complexity or interpretability, the models can present a poor performance. Several approaches are found in the specialized literature to reduce the complexity and improve the interpretability of the fuzzy models. Here, a post-processing approach is taken into account via the definition of the rules selection criterion that aims to choose the most relevant rules according to the well-known accuracy-interpretability trade-off. This criterion is based on Orthogonal Transformations, here the QRP transformation is taking into consideration, and its parameters are tuned genetically. The main objective is to check the true significance, drawbacks and advantages the firing matrix of the rules, that is the foundation of the most usual approaches based on orthogonal transformations for the complexity reduction of the fuzzy models. A neuro-fuzzy system, FasArt (Fuzzy Adaptive System ART based), and several case studies, data sets from the KEEL Project Repository, are used to tune and check this approach. This neuro-fuzzy system generates Mamdani fuzzy rule based systems (FRBSs), each with its own particularities and complexities from the point of view of fuzzy sets and rule generation. NSGA-II is the MOEA tool used to tune the criterion parameters based on accuracy-interpretability ideas.

2011-01-01T00:00:00Z Gacto, María José Galende Hernández, Marta Alcalá, Rafael Herrera, Francisco https://uvadoc.uva.es/handle/10324/74373 2025-02-06T20:00:49Z 2014-01-01T00:00:00Z

In this contribution, we propose a two-stage method for Accurate Fuzzy Modeling in High-Dimensional Regression Problems using Approximate Takagi–Sugeno–Kang Fuzzy Rule-Based Systems. In the first stage, an evolutionary data base learning is performed (involving variables, granularities and slight fuzzy partition displacements) together with an inductive rule base learning within the same process. The second stage is a post-processing process to perform a rule selection and a scatter-based tuning of the membership functions for further refinement of the learned solutions. Moreover, the second stage incorporates an efficient Kalman filter to learn the coefficients of the consequent polynomial function in the Takagi–Sugeno–Kang rules. Both stages include mechanisms that significantly improve the accuracy of the model and ensure a fast convergence in high-dimensional and large-scale regression datasets. We tested our approach on 28 real-world datasets with different numbers of variables and instances. Five well-known methods have been executed as references. We compared the different approaches by applying non-parametric statistical tests for pair-wise and multiple comparisons. The results confirm the effectiveness of the proposed method, showing better results in accuracy within a reasonable computing time.

2014-01-01T00:00:00Z