"multiple weld pass" is not in itself a defect.
Diaphragm connection plate welds crossing the web-to-flange weld is not inherently a problem; the problem areas on the Hoan bridge had much more complex geometry and stresses than that. Intersecting welds are discouraged but are not always automatically a problem (for example, every web-to-flange weld intersects the web and flange butt splices). This particular form of intersecting weld is common in Japan (
http://international.fhwa.dot.gov/Pdfs/SteelBridge1.pdf) and there is research in the U.S. suggesting it would not be a problem (
http://www.utexas.edu/research/ctr/pdf_reports/0_4178_1.pdf). You'll have to have a design engineer evaluate the state of stress at that intersection. If it's not a problem, you could make a worse mess by trying to remedy it.
Spatter is not a weld defect. It can cause a problem with paint and should be ground off before you repaint the bridge, but it's not a weld defect.
Porosity, if at a rejectable level (check AWS D1.5 for acceptance criteria), would be repaired in exactly the same way it would have been in the shop, by grinding and re-welding.
I'm not sure what you're saying about the Charpy results. Is it the base metal or the weld metal that's failing? What retrofit are you proposing there?