impact
- Potential Improvement in energy efficiency between 7% and 10% depending on equipment.
- Potential improvement in energy consumption by 15% approx. in peripheral systems such as cooling.
Developed AI Technologies
Pattern analysis, prediction and intelligent classification.
Challenges
- Energy efficiencys diagnosis of a radio and television communication infraestructure based in Barcelona area.
- Deployment of the computing node at the edge for high-speed data sampling and online analytics.
results
- Dataset size: 128.06 GB.
- Number of variables: 15.
- Sampling Time: 125 μ s.
- Different radio frequency equipment has been analyzed to study their energy consumption patterns with respect to the transmission power of 20kW.
- Probabilistic clustering techniques have been used to find different groups based on the equipment consumption data.
Consumption patterns found by Gaussian mixture models clustering techniques for the same type of equipment, but from different radio stations.
- The diagnosis approach has helped to identify improvement opportunities in terms of the configuration of equipment parameters and also according to the type of equipment.