Least Squares Support Vector Machine Classifiers
Neural Processing Letters
Learning and Soft Computing: Support Vector Machines, Neural Networks, and Fuzzy Logic Models
Learning and Soft Computing: Support Vector Machines, Neural Networks, and Fuzzy Logic Models
A computer system for predictive maintenance of wind generators
ICCOMP'08 Proceedings of the 12th WSEAS international conference on Computers
| Hi-index | 0.00 |
The use of open-source software in many institutions and organizations is increasing. However, a balance should be considered between the software cost and the cost of its technical support and reliability. In this article, a maintenance system for wind farms will be presented. It is connected to an information system for maintenance, called SMIT (Terology Integrated Modular System) as a general base to manage the assets and as a support strategic line to the evolution of this system, which incorporates on-condition maintenance modules, and the support to the research and development done around this theme. The SMIT system is based on a TCP/IP network, using a Linux server running a PostgreSQL database and Apache web server with PHP, and Octave and R software for numerical analysis. Maintenance technicians, chiefs, economic and production management personnel can access SMIT database through SMIT clients for Windows. In addition, this maintenance system for wind systems uses also special low cost hardware for data acquisition on floor level. The hardware uses a distributed TCP/IP network to synchronize SMIT server master clock through Precision Time Protocol. Usually, the manufactures construct, deploy and give the means for the suppliers to perform the wind system's maintenance. This is a very competitive area, where companies tend to hide the development details and implementations. Within this scenario, the development of maintenance management models for multiple wind equipments is important, and will allow countries to be more competitive in a growing market. For on-condition monitoring, the algorithms are based on Support Vector Machines and time series analysis running under Octave and R open-source software's.