Unexpected behavior in CHP system model

[amyeallen1]

I noticed unexpected behavior in the CHP system model generated by the heat_with_CHP test on the MBL_v9 branch. (This may well exist in versions of the model generated without v9, as well.)

When I ran the simulation for a 2-month period, I noticed quite a few periods of negative mass flow at the supply temperature sensor (also corresponding to negative mass flow at the return mass flow rate sensor). The negative flow values are contrary to expectation per the assumed direction of flow through the sensor. (There was virtually no flow through the bypass.) This behavior occurred frequently, so it could be observed with a shorter-duration run as well.

This plot shows a snapshot of the unexpected behavior. The supply and return mass flow rates are almost identical during this period (no bypass flow), and thus appear coincident.