I am currently modeling a coal combustion boiler for electricity production (350 MWe). I have read several papers looking for new ideas to validate the models. I see that it is usual to compare measured values with calculated values for: GAS TEMPERATURE at different planes, %O2 at different planes and NOx (if modeled). I think that these validations are not good enough, because I am not very confident on the goodness of the plant measurements (particularly those regarding gas temperature, where is it really measured?). I would be grateful if any people could give me ideas and opinions about this, thanks. ELI

I will have similar issues when I get to validating my model for underground coal gasification.

It seems to me that you need to validate these complex models at different scales. This is because you can't possibly validate many of the details using plant data, but on the other hand lab data, which might provide high quality high detail data for a specific problem, won't encompass all the details of the full-size plant, which you ultimately want to use the model to predict.

Thus perhaps validate the reaction mechanisms and rates against high quality lab data where the environment is more easily controlled (local validation). And then use data from full-scale to validate the global predictions of the model.

I suppose it would be best then if you also actually did the measurements at each scale yourself so as to avoid any uncertainty in what was measured.

Unfortunately, this is rearely possible.