Energy storage is an important topic as many countries are seeking to increase the amount of electricity generation from renewable sources. An open and accessible online database on energy storage technologies was created, incorporating a total of 18 energy storage technologies and 134 technology pages with a total of over 1,800 properties. In this database information on technical maturity, technology readiness level and forecasting is included for a number of technologies. However, since the data depends on various sources, it is far from complete and fairly unstructured. The chief challenge in managing unstructured data is understanding similarities between technologies. This in turn requires techniques for analyzing local structures in high dimensional data. This paper approaches the problem through the use and extension of t-stochastic neighborhood embedding (t-SNE). t-SNE embeds data that originally lies in a high dimensional space in a lower dimensional space, while preserving characteristic properties. In this paper, the authors extend the t-SNE technique with an expectation-maximization method to manage incompleteness in the data. Furthermore, the authors identify some technology frontiers and demonstrate and discuss design trade-offs and design voids in the progress of energy storage technologies.
|Title of host publication||International Society for Scientometrics and Informetrics|
|Subtitle of host publication||ISSI|
|Number of pages||10|
|Publication status||Published - 2015|
- energy storage technologies
- renewable technologies