Incorrect cloud dimensions/concentration for BritterMcQuaidPuff

[aefarrell]

The BritterMcQuaidPuff model should return a cylinder of constant concentration at a given time t. However, it does not. I think it is calculating the domain based on the supplied x and not the cloud center x_c, which is the source of the error.

[aefarrell]

The typical process in the Britter-McQuaid workbook is to calculate the downwind distance x given a concentration, e.g. calculate the distance to the LFL or something like that, whereas I am trying to do the opposite: calculate the concentration at a given point. I think that is the source of the confusion.

But I also built my unit tests from a worked example, so I will have to look at it more closely to see how I went wrong without catching with the tests.

[aefarrell]

In fixing this one bug I found another problem, which was masked by how I created the unit tests. For the unit test I by-passed the normal constructors and gave values for the release and substance parameters to match what was in the Britter-McQuaid workbook example 6.3. The problem with this is that there is an error in the Britter-McQuaid example.

They give a release of 38t of ammonia expanding into an air/ammonia cloud with ~6% (v/v) ammonia as the starting point for the correlations. Unfortunately, the numbers they give correspond to 3.8t of ammonia.

The workbook example does not make the non-isothermal correction, so perhaps I should also make that a user-selectable option? If you are trying to replicate what is in the workbook, it won't work out of the box, even though the non-isothermal correction is part of the model specification. The workbook examples are inconsistent about whether or not they apply it.