How about an ELISA? You can buy kits for this. Here are two:

http://www.abcam.com/human-albumin-elisa-kit-ab108788.html

https://www.thermofisher.com/order/catalog/product/EHALB

Because of budgetary limitations I was looking for something more DIY. 

OK. My next idea is to use SDS-PAGE. Albumin has a molecular weight of about 66 kDa. It is a fairly abundant protein in BALF (see page 39 of this publication: http://www.ers-education.org/lrmedia/1999/pdf/44063.pdf). Run samples of BALF on a gel along with standards containing a range of known amounts of purified albumin. Stain the gel with Coomassie Blue and photograph the gel. Then use an image analysis program to compare the amount of stain in the albumin bands from BALF to the standards. (If you have access to a densitometer, you can use that instead of photography.) If there are too many other proteins of around the same molecular weight, you may have to resort to 2-D gels. If albumin is not a major protein in BALF, then use Western blotting with an anti-albumin antibody instead of Coomassie staining. Do not use BSA as a blocking agent, of course.