BCAM
Academic
Experts in:
Modeling and simulation
- Manufacturing processes: injection, stamping, forging, additive manufacturing, etc.
- Fluids: gases, liquids.
- Fluid-structure interaction.
- Mechanical or structural.
- Combustion.
- Environmental.
- Multiphysical phenomena.
Statistics and Big Data
- Statistical advice and data analysis.
- Time series forecasting.
- Production of maps from spatial data.
- Environmental and energy statistics.
- Biostatistics, epidemiology.
- Applications in the field of health.
- Statistical applications for industry or public administrations.
Optimization
- Resource optimisation.
- Resource location optimisation.
- Planning of transport routes.
- Optimal decision support.
- Optimisation of industrial and business processes.
Artificial intelligence
- Automatic learning.
- Neural networks.
- Bayesian networks.
- Deep learning.
- Machine learning.
- Hybrid models based on data and physics.
Computing
- Programming in scientific languages: Fortran, C, C++, Pyton, Matlab, R.
- Development of software packages.
- Parallelisation of algorithms.
- Use of free software packages.
- Programming on GPUs.
- Distributed computing.
- High performance computing.
- Data/database modelling
- Quantum computing.
Areas of experience:
Challenge 1: Health
- Visualisation, processing and analysis of medical images. Characterisation of false positives. Learning models using large databases (machine/deep learning).
- Evolution of demographic rates. Population projections and generation of scenario trees over a time horizon.
- Characterisation of population habits. Simulation of user flow by services.
- Optimization of the planning of the location of primary health centres in uncertain environments of the evolution of demographic rates over a time horizon.
- Optimization of ambulance fleet types: Sizing and location in an uncertain environment of service demand.
- Optimization of the distribution of health resources, both human and material, in emergency situations.
Challenge 3: Energy
- Forecasting and planning of energy production for domestic or industrial use.
- Optimization of energy distribution networks.
- Optimization of the location of renewable sources (wind, solar, photovoltaic, etc.).
- Optimization of maintenance planning for electricity generators.
- Numerical simulation of heat transfer and combustion processes.
- Decision-making assistance in energy processes.
Challenge 5: Environment
- Simulation, prediction and control of pollutant emissions.
- Modelling and simulation of forest fires.
- Optimization of sustainable forest exploitation planning.
- Optimization of irrigation planning for agricultural and livestock purposes.
- Optimization of the sustainable planning of the exploitation of hydrographic basins for industrial purposes.
Success stories:
Multiscale Simulations to Develop Advanced Battery Materials
In partnership with CIC energiGUNE, BCAM has developed novel computational strategies to understand and optimize the inner-workings of complex battery materials. Our strategy incorporates quantum mechanical information into accurate atomistic and mesoscale models that run efficiently on parallel computing architectures. In addition, BCAM has introduced AI-based screening tools to efficiently pin-point the “best material for the task” given a set of goals and constraints.
Neiker: Mapping High-Resolution Soil Properties With Geoadditive Models
Neiker required high-resolution maps of carbon stocks and soil texture properties in different land uses at 0-30cm depth in the Basque Country. To take care of this need, BCAM developed and implemented a unified modelling framework for spatial prediction that incorporates the effects of climate variables such as average temperature, min/max temperature and precipitation. These maps contribute to agricultural planning of crops, forest management and environmental protection. Stock carbon and soil texture maps are publicly available at GeoEuskadi.
Machine Learning to improve customer satisfaction in insurance
Seguros Lagun Aro wanted to increase customer satisfaction and prevent them from leaving the company. To fullfil this need, BCAM developed and implemented a machine learning model to predict customer loss and retention probabilities. The software allows re-training the model in a streaming fashion and in a reasonable period of time, considering the great volume of data they work with.
More success stories:
http://www.bcamath.org/documentos_public/archivos/BCAM_Industry.pdf