This paper presents initial results of methodologies developed for extracting useful information from online textual data in the field of power networks. Since there are no tools developed specifically for power engineering, an attempt was made to take advantage of the existing tools for specialized web browsing and data extraction. Two methodologies are explored: one for extracting on-line documents (journal papers, technical reports, etc.) that are highly related to a specific topic, and the second one for extracting highly related sentences, providing a literature summary on the topic. The first results are promising, showing great potential for text-mining applications as a part of power networks data analytics.