Lots of it is already freely available (Score 2)

I work at Lawrence Berkeley National Laboratory (LBNL), a large (4000-person) Department of Energy research lab that is, unfortunately but understandably, often confused with Lawrence Livermore National Laboratory. We get lots of funding from DoE, of course, but also from other federal and state agencies.

Almost all of the work we do is published in the form of "LBNL reports", most of which are freely available, although hard to find. Much of the work is later published in scientific journals, and it is sometimes problematic that an LBNL report of the work already exists: some of my colleagues have had papers rejected on the grounds that the work was already 'published' as an LBNL report. (That's bad because LBNL reports are not usually peer reviewed, except for an internal review process). Perhaps because of that, LBNL does not make the LBNL report database searchable by outsiders. However, most (maybe all?) of the reports are supposedly available through the Science.gov portal. But the search facilities there are so primitive that I've never been able to find what I'm looking for there (for instance, even searching on my own name plus a few words from one of my report titles doesn't work reliably, turning up hundreds of hits that may or may not include my report).

Fortunately, the search engine of your choice is probably adequate. If you're looking for work that was done here at LBNL, simply putting "LBNL report" at the start of your search request will probably work. For instance, Google [LBNL report building electric load] and you will find a bunch of reports on analyzing electricity data from buildings, usually from the relevant LBNL department website. A lot of this work was also published in journals that are behind paywalls, but the same content is here for free. Often there is additional material that had to be cut in order to make it into the journal, so sometimes the reports are better than the papers. On the other hand, the papers do benefit from modifications due to reviewer comments, and are often prepared with more care, so sometimes the papers are better. I think several other federal agencies have similar policies.

So if you find a journal paper that you're interested in but can't read because it's behind a paywall, and the authors work for the government, then try searching on the author names and a few of the words from the title and add "report" to your search (or "LBNL report" or "NASA report" or whatever). You have a pretty good chance of finding what you want.