Physical Trough Model with High Freeze Protection

When you use the physical trough model in SAM 2014.1.14 for a system that requires a lot of freeze protection energy, simulations may fail. Here is a description of why with a possible workaround solution.

The CSP models use iterative algorithms to calculate temperature, mass flow rate, and other values for each time step of the simulation. The algorithms involve large numbers of calculations, and a relatively small error can cause an algorithm not to converge in a given time step. SAM allows a number of convergence failures before stopping the simulation and generating an error report. Often when simulations run slowly but do not stop, it is because of these convergence failures. You can see a full report of the convergence failures by pressing the Shift-F9 keys, and navigating to the [case_name]/base/trnsys.dtr.log file.

The problem for systems with significant freeze protection is that, especially for high temperature applications, the freeze-protection strategy does not always converge. The freeze protection controller attempts to increase the loop inlet temperature until the loop outlet temperature is above the freeze protection temperature. In these high temperature cases, the additional thermal losses to the environment caused by increasing the operating temperature offset the effect of the higher inlet temperature, and the outlet temperature cannot reach its target. Eventually, the freeze protection controller sets the inlet temperature to a really large value and convergence fails.

This happens because when the model detects that there is no DNI (when freeze protection energy is typically needed), it sets the mass flow rate to the minimum mass flow rate. Our original justification for this behavior was that it minimizes the pumping energy requirement. However, one negative consequence of this approach is that a larger temperature difference is required from the HTF to absorb (or reject) a given amount of thermal energy.

A workaround to this problem is to increase the mass flow rate during times when freeze protection is needed. You can do that by increasing the “Min single loop flow rate” on the Solar Field page. For example, in one case, we found that changing the value from the default 1 kg/s to 4 kg/s for a system that was generating errors resulted in a successful simulation.

One potential problem with this approach is that it may limit the outlet temperature of the HTF during time periods with average-low DNI, so there's a trade-off. In an actual system, there is no physical reason that the HTF mass flow couldn’t be limited to 1 kg/s during DNI operation and set to 4 kg/s, for example, during freeze protection. We’ll work on a fix to address this for the next version of SAM.

Best regards,
Paul.