
This might be related to the (technically correct) behavior in following "stationary process", in which fluid is seemingly flowing, but the quantities do not change anywhere after the initial transient phase:

(I personally would optimize such closed loops away, à la belt optimizations, to save computations, but that's beside the point)