0.25
0.5
0.75
1.25
1.5
1.75
2
Novel Geo-spatial Interpolation Analytics for General Meteorological Measurements
Published on 2014-10-071945 Views
This paper addresses geospatial interpolation for meteorological measurements in which we estimate the values of climatic metrics at unsampled sites with existing observations. Providing climatologica
Presentation
Novel Geo-spatial Interpolation Analytics for General Meteorological Measurements00:00
IBM Smarter Energy Research Institute: Advance the Future of Utility 12:52:55
Electrical (Power) Engineering101 43:31:02
Electric Utility is a Capital Intensive Industry 83:33:03
Much more Distribution Transformers than Power Transformers 97:54:50
Optimal Asset Maintenance and Capital investment is of Importance 106:58:54
But How to Obtain Ambient Temperature of Transformers? 139:41:08
Challenge: there are only Limited Number of Weather Stations 153:40:36
Paucity of Weather Station Network is a Common Issue 170:40:25
Physical-based Model is typically Computationally Intensive 184:23:14
Learning-based Model: Trade-off between Accuracy and Speed 195:18:24
Major Contributions of this Work 201:13:52
Estimate DCT Coefficients to Minimize Reconstruction Error with Student-t prior227:48:11
Student-t is a better Prior than Laplace for Meteorological Metrics 238:24:05
BCST as a General Meteorological Interpolation Service 244:19:04
Experimental Setup 251:12:10
Temperature Interpolation Results on KNMI (metric: percentage) 255:47:41
Runtime Comparison (in ms) 262:54:26
Example: Barometric Pressure Map in NYC 265:59:45
Conclusions & Future Work 268:15:24