Knowledge management for engineers is a notoriously difficult thing. One of the problems lies in how engineers actually use information: they don't use formal sources, they rely on personal contacts, and if they can't find information they need they'll just make do.
Perhaps we're studying the wrong information seeking dimension of engineers. Perhaps we should look at how they interact with other epistemic communities. Engineers are notoriously paranoid about litigation and go to great pains to document their interactions with other actors such as suppliers, architects, consultants, and vendors. Maybe KM insights lie in these documents.
In large civil engineering projects, a notorious artifact that spans various epistemic cultures--a boundary object according to Susan Leigh Star--is the change order or engineering change notice. Perhaps we can find some insights in studying these documents as boundary objects.
Susan Leigh Star
Sample spec on change orders
Academic Paper: Representation of Activity Knowledge for Project Managemen