In the power test code ( PTC 4.1 or newer) there is a curve that shows expected % loss in efficiency vs total heat input to the burners. This is a rough estimate and might not be accurate if the boiler is a criculating fluid bed boiler ( or some types of MSW units) with a lot of uncooled refractory components.
A more accurate method ( and proposed for the new PTC) is to directly measure the heat loss using an infrared pyrometer. An audit would be conducted during full load operation , and a sample reading of each type of casing used at the site( ie, windbox, waterwall, heat rcovery area, hoppers, flues, ducts, etc) and weighted according to the surface area and emmissivity of the surface.
Please note the emmissivity can be deliberately jiggled in the design phase in order to minimize the amount of insualtion needed to hold a 140 F max surface temperature. Some tricks used are heat emmitting paint(e=0.9) and anodized aluminum ( e=0.85), which seem to be much cooler than standard aluminum lagging ( e=0.10).