Automated Analysis of Renewable Energy Datasets
The goal of this project is to dramatically improve the understanding of renewable energy [RE] and energy efficiency [EE] datasets and the depth and efficiency of their analysis through the application of statistical learning methods (“data mining”) in the intelligent processing of these often large and messy information sources.
The project focused on developing general purpose algorithms for the “automated analysis”:
- automated anomaly detection (e.g., identifying data quality problems)
- data mining to detect patterns (e.g., finding opportunities for selecting representative subsets of large databases)
- sophisticated, comparative data-quality assessments (e.g., determining which portions of competing data sets are superior for particular purposes)
- reasoning “robots” that autonomously explore data sets (e.g., inferring favorable economic niches for renewable energy deployment).
Selected Publications
B. Bush and E. Kalendra, “Multivariate Time Series Analysis Applied to an Irradiance Dataset.”
D. Getman and B. Bush, “Use of Autocorrelation in Detection of Anomalies in Spatiotemporal Datasets,” National Renewable Energy Laboratory, Golden, Colorado.
D. Inman, R. Elmore, and B. Bush, “Imputation and clustering of irregular time series data for improved electricity power and demand modeling in commercial buildings.”
E. Kalendra and B. Bush, “Gaussian Process Modeling of Multivariate Time Series.”