Pasar al contenido principal

DESARROLLO DE METODOLOGIAS MULTIESCALA EFICIENTES MEDIANTE TECNICAS DE INTELIGENCIA ARTIFICIAL PARA OPTIMIZACION TOPOLOGICA ESTRUCTURAL DE ALTA DEFINICION

PID2022-141512NB-I00

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal para Impulsar la Investigación Científico-Técnica y su Transferencia
Subprograma Subprograma Estatal de Generación de Conocimiento
Convocatoria Proyectos de I+D+i (Generación de Conocimiento y Retos Investigación)
Año convocatoria 2022
Plan Estatal 2021-2023 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSITAT POLITÈCNICA DE VALÈNCIA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 5
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Two-Level High-Resolution Structural Topology Optimization with Equilibrated Cells

  • MERLI GISBERT, RAFAEL|||0000-0003-1987-4683
  • Ródenas, Juan José|||0000-0003-2195-7920
  • Bosch-Galera, Marc|||0000-0002-4286-6721
  • Nadal, Enrique|||0000-0002-2808-298X
  • Martínez-Martínez, Antolín
[EN] In today's industry, the rapid evolution in the design and development of optimized mechanical components to meet customer requirements represents a significant challenge for companies. These companies seek efficient solutions to enhance their products in terms of stiffness and strength. One powerful approach is Topology Optimization, which aims to determine the optimal material distribution within a predefined domain to maximize the overall component's stiffness. Achieving high-resolution solutions is also crucial for accurately defining the final material distribution. While standard Topology Optimization tools can propose optimal solutions for entire components, they struggle with small-scale details (such as trabecular structures) due to prohibitive computational costs. To address this issue, our proposed approach introduces a two-level topology optimization methodology considering density-based techniques. The proposed methodology includes three steps: The first one subdivides the whole component in cells and generates a coarse optimized low-definition material distribution, assigning a different density to each cell. Since the output stresses from the coarse problem are not equilibrated into each cell, they must not be directly used in the fine level. Thus, the second step uses the equilibrating traction recovery approach to convert the cell nodal forces into equilibrated lateral tractions over the cell boundary. Finally, taking as input data the densities from the coarse optimization and imposing these lateral tractions as Neumann boundary conditions, each cell is optimized at fine level. The main goal of this work is to efficiently solve high-resolution topology optimization problems using a two-level mechanically-continuous method, which would be unaffordable with standard computing facilities and the current techniques.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Machine Learning Model in Obesity to Predict Weight Loss One Year after Bariatric Surgery: A Pilot Study

  • Nadal, Enrique|||0000-0002-2808-298X
  • Ródenas, Juan José|||0000-0003-2195-7920
  • Benito, Esther
  • Ródenas-Navarro, Ana María
  • Palanca, Ana
  • Martinez-Hervas, Sergio
  • Civera, Miguel
  • Ortega, Joaquín
  • Alabadi, Blanca
  • Piqueras, Laura
  • Real, Jose T.
[EN] Roux-en-Y gastric bypass (RYGB) is a treatment for severe obesity. However, many patients have insufficient total weight loss (TWL) after RYGB. Although multiple factors have been involved, their influence is incompletely known. The aim of this exploratory study was to evaluate the feasibility and reliability of the use of machine learning (ML) techniques to estimate the success in weight loss after RYGP, based on clinical, anthropometric and biochemical data, in order to identify morbidly obese patients with poor weight responses. We retrospectively analyzed 118 patients, who underwent RYGB at the Hospital Clínico Universitario of Valencia (Spain) between 2013 and 2017. We applied a ML approach using local linear embedding (LLE) as a tool for the evaluation and classification of the main parameters in conjunction with evolutionary algorithms for the optimization and adjustment of the parameter model. The variables associated with one-year postoperative %TWL were obstructive sleep apnea, osteoarthritis, insulin treatment, preoperative weight, insulin resistance index, apolipoprotein A, uric acid, complement component 3, and vitamin B12. The model correctly classified 71.4% of subjects with TWL < 30% although 36.4% with TWL ¿ 30% were incorrectly classified as ¿unsuccessful procedures¿. The ML-model processed moderate discriminatory precision in the validation set. Thus, in severe obesity, ML-models can be useful to assist in the selection of patients before bariatric surgery., This research was funded by Instituto de Salud Carlos III, grant numbers PI15/00082, PIE15/00013, PI21/00220, and Conselleria de Educacion, Cultura y Deporte, Generalitat Valenciana, grant number AICO 2019/250 to L.P. and Prometeo/2021/046 to J.J.R. This work was also supported by Ministerio de Economia, Industria y Competitividad, grant number SAF2014-57845-R and by Ministerio de Ciencia e Innovacion, Agencia Estatal de Investigacion (MCIN/AEI/10.13039/501100011033) and Fondo Europeo de Desarrollo Regional (FEDER) (Project PID2022-141512NB-I00) to J.J.R. The authors also received financial support from CIBER de Diabetes y Enfermedades Metabolicas Asociadas (CIBERDEM), a research network of the Ministry of Health, Madrid, to J.T.R.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Accelerating cell topology optimisation by leveraging similarity in the parametric input space

  • Martínez-Martínez, Antolín
  • Muñoz-Pellicer, David
  • Allix, Olivier
  • Chinesta Soria, Francisco José
  • Navarro-Jiménez, José-Manuel|||0000-0002-7333-9377
  • Ródenas, Juan José|||0000-0003-2195-7920
  • Nadal, Enrique|||0000-0002-2808-298X
[EN] The design of high-resolution topology-optimised (TO) structures is important for many industrial and medical applications because of their better mechanical performance under different load conditions. Traditional density-based TO methods, like the Solid Isotropic Material with Penalisation (SIMP) method, can produce detailed designs but are very computationally expensive, especially for fine meshes. While surrogate models using neural networks can speed up the process, they often lack generality and can create discontinuities, making them less effective for solving new problems. This study addresses these issues by introducing a method to speed up cell-level TO within a 2-Level framework, where large structures are built by combining optimised square cells. A data-driven instance-based model provides a better starting point for the standard SIMPbased optimiser, placing it closer to a local minimum and reducing computation time. To avoid the generality problems of other methods, the instance-based model uses a dataset expanded through two strategies: context-based data creation, which generates specific samples for the problem, and data augmentation, which increases dataset size without extra computation. Two similarity metrics, vector-based and energy-based, are used to measure how close the input parameters are. Both metrics are effective, but the energy-based metric is expected to work better in 3D cases, where higher-dimensional input spaces make other approaches less reliable. This methodology addresses important challenges associated with existing instance-based models, enhancing the speed and applicability of high-resolution TO., The authors gratefully acknowledge the financial support of:center dot Grant FPU 19/02103 funded by "Ministerio de Educacion y Formacion Profesional".center dot Grant MCIN/AEI/10.13039/501100011033 funded by "Ministerio de Ciencia e Innovacion".center dot Grant PID2022-141512NB-I00 funded by "Fondo Europeo de Desarrollo Regional-FEDER-".center dot Grants Prometeo/2021/046 and CIAICO/2021/226 funded by "Generalitat Valenciana".O. Allix would like to thank the French National University Council, ENS Paris-Saclay and National Center for Scientific Research (CNRS) for supporting his sabbatical at UPV, which made it possible to closely interact with the colleagues from I2MB/UPV.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Identification of the homeostatic bone-remodelling stimulus from medical images

  • Gutiérrez-Gil, J.
  • Ródenas, J.J.
  • Nadal, Enrique|||0000-0002-2808-298X
  • ATIENZA VICENTE, CARLOS MANUEL|||0000-0002-2522-2131
  • Navarro-Jiménez, José-Manuel|||0000-0002-7333-9377
[EN] Background & objectives: Computational methods enable the effective processing of medical data and the modelling of complex biological systems, thereby supporting the development of Personalized Medicine over recent decades. In this context, bone remodelling simulation is a promising tool, as the prediction of patient-specific bone adaptation may have multiple clinical and experimental applications in bone disease assessment and management. Numerical models combining the Finite Element Method (FEM) with bone remodelling algorithms can be generated from medical data such as CT scans. However, their accuracy strongly depends on how realistically bone properties and boundary conditions are represented. Because case-specific data are often difficult to obtain, generalized assumptions are commonly introduced, which can compromise simulation accuracy. This work aims to improve the personalization of bone remodelling simulations through the identification of the attractor stimulus (PA), a key parameter in bone remodelling analysis. To this end, we propose what is, to our knowledge, the first fully numerical and patient-specific methodology to identify PA directly from medical images. Methods: Three approaches for identifying PA were explored. Method I was based on the assumption of homeostasis observable in medical imaging. Method II proposed a simplified methodology grounded in Mechanostat theory. Method III was designed as an in silico analogy of an in vivo experiment in order to illustrate the limitations associated with that type of experimental approach. All proposed methods were designed to rely only on a patient CT scan and body weight, thus avoiding the need for additional patient-specific measurements and supporting a clinically feasible workflow. Results: Applied to a human femur case study, Methods I and II yielded subject-specific PA values below the generic literature reference (of the order of 30-40 vs 50 MPa/day), whereas the in silico replication of the single-gauge experiment (Method III) produced markedly higher and probe-position-sensitive values, indicating that it is the least reliable of the three. These results suggest that CT-based information combined with patient weight may be a good starting point to estimate this key parameter without additional invasive measurements, supporting a clinically feasible route to personalize bone remodelling simulations. Conclusions: The identification of the attractor stimulus PA from routinely available patient data may enhance the personalization of bone remodelling simulations while preserving clinical feasibility. This approach may contribute to the integration of subject-specific bone remodelling models into real medical contexts, with potential applications in both clinical decision-making and experimental research., The authors gratefully acknowledge the financial support of Generalitat Valenciana (FEDEGENT/2018/025). Grant PID2022-141512NBI00 funded by MICIU/AEI/ 10.13039/501100011033 and, as appropriate, by ERDF A way of making Europe , by ERDF/EU , by the European Union or by the European Union NextGenerationEU/PRTR .
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Tesis de maestría (MasterThesis)

Diseño de estructuras de alta resolución topológicamente optimizadas mediante un algoritmo multinivel con continuidad entre celdas, Design of Topologically Optimized High-Resolution Structures Using a Multilevel Algorithm with Continuity between Cells, Disseny d’estructures d’alta resolució topològicament optimitzades mitjançant un algorisme multinivell amb continuïtat entre cel·les

  • MERLI GISBERT, RAFAEL|||0000-0003-1987-4683
[ES] En el presente trabajo de fin de máster (TFM) se plantea la mejora de una metodología de diseño de estructuras de alta resolución topológicamente optimizadas, útiles para patrones de relleno en procesos de fabricación aditiva. Esta metodología consta de dos niveles: en el nivel superior de discretización se utiliza el método SIMP de optimización topológica basada en densidades, y posteriormente, se obtienen unas tracciones equilibradas en el contorno de cada celda con un método de equilibrado nodal de fuerzas; con estas tracciones se ejecuta una nueva optimización para cada celda de la discretización inicial (segundo nivel), nuevamente mediante el SIMP, hasta obtener la distribución final del material. La definición de las tracciones equilibradas en el contorno (interpolación lineal de cargas nodales) conlleva la aparición de material en todo el perímetro de las celdas optimizadas, obteniendo regiones con distribución de material difusa, que no es fabricable. En este trabajo se plantea una estrategia iterativa para distribuir las cargas en el contorno de cada celda de manera selectiva, obteniendo así distribuciones trabeculares de material, con una correcta definición que además es contínua entre bordes de celdas adyacentes, permitiendo su fabricabilidad desde el punto de vista geométrico. Dicha estrategia de selección se trata en detalle mediante filtros de densidad en el contorno de cada celda, que permiten controlar en la fase de diseño el tamaño de los miembros que atraviesan el contorno entre celdas. Finalmente, se acepta que el método ha convergido cuando la nitidez en el contorno de cada celda es suficiente para los fines de fabricación deseados. Además del método SIMP, se utiliza el Cartesian Grid Finite Element Method (cgFEM), un software propio desarrollado en Matlab® que permite validar la metodología mediante varios ejemplos numéricos de aplicación. La metodología propuesta se ha aplicado a casos 2D, con resultados prometedores para su futura aplicación en estructuras tridimensionales.