Prediction based on big data and neural networks

Development and implementation of predictive maintenance procedures based on big data and neural networks.

A software platform for the analysis and early detection of faults and possible malfunctions of wind turbine components.

 

The aim of this project is to develop a software platform aimed at providing tools and applications for the analysis and early detection of faults and possible malfunctions in the different elements present in wind turbines, with the aim of applying preventive maintenance methodologies to extend the working life of the equipment.

GENERAL INFORMATION

  • Location: Technological Innovation Department of Acciona
  • Duration: January 2019 – July 2020
  • Participants: Acciona
  • Budget: €235,980
  • Note: This company has received aid co-funded 50% by the Government of Navarra and the European Regional Development Fund (ERDF) through the Operational Programme ERDF 2014-2020 for Navarra.

Technical objectives

  • The creation of an application for the handling and treatment of data, generating models and indicators with the aim of detecting malfunctions and calculating RUL (Remaining Useful Life) estimates for the equipment.
  • The development and training of behavioural models by combining different Artificial Intelligence techniques based on neural networks, for the analysis of data in real time and the early detection of faults or malfunctions of equipment to facilitate diagnosis and identify causes.
  • Structure the simultaneous performance of a series of strategies with a dynamic distribution of the total capacity to each one of them.
  • Design of a Big Data architecture that can handle large volumes of data to provide the necessary computational capacity to develop standardized models of equipment performance and search for behaviour patterns.
  • Development of a standardized database from different sources for the purposes of analysis, and the creation of a standard behavioural model for the equipment based on the data obtained.

Participantes en el proyecto

More innovative projects

Discover all of our projects