Can parallel computing work for smart grid data?
December 20, 2011
Having too much information is a commonly heard complaint today. Yet, in the energy and utility industry, millions of smart meters are being deployed with the promise of providing data at unprecedented levels. Many agree that the ability to digest a multitude of smart grid data will be the key to success for the industry. How can we process and analyze the imminent flood of information in a timely and efficient manner?
One solution seems simple: reduce the data to fit our existing methods. For instance, with statistical methods, we can design representative samples or aggregate datasets to levels that are more manageable, i.e., smaller summaries of data that are still useful. But, these methods also have limitations. In some cases, utilities may be required to run analyses on an entire population, or companies may be interested in mining live streams of data to provide quick feedback to customers. In these situations, they can’t apply or do not want to use statistical methods. So, they face the challenge of working with larger datasets. As datasets increase via an influx of information, the idea of what constitutes a manageable dataset will evolve, and analysts will need to adapt to new sizes.
For the full article see: http://www.kemautilityfuture.com/can-parallel-computing-work-for-smart-grid-data
