[esp-r] Convergence of explicit plant model

[Lomanowski, Bartosz]

Hello,
 
A question about explicit plant modeling: 
 
I am developing a model with a plant heating system that utilizes a hydronic heat pump. The source side of the heat pump is a closed loop between a heat exchanger and the evaporator node of the heat pump. The load side charges a storage tank. The hydronic heat pump is set up similar to a 2-node fluid-fluid heat exchanger. The closed loop on the source side is proving to be highly unstable. I am using a 1 min time step on the plant side and running the plant solver fully implicit. The solution converges if the flow rates are constant but blows up when the flow rates are toggled by an on/off control.
 
When I combine the heat exchanger with the evaporator on the source node of the heat pump, elminating the closed loop, the solution is very stable.  I can then run it with 5 min ts and the default solver implicitness which is a function of the time constant of the components. In this scenario, I have to make a crude assumption as to the heat exchanger efficiency so I would prefer to get the closed loop working. 
 
Does anyone have any hints on modeling closed loops in the plant domain? 
 
Thanks,
Bart
 
NRCan, Canada