[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Vertical U-Tube Ground Heat Exchanger and Ground source

At 09:40 AM 3/16/2004 -0500, Michel Bernier wrote:

The incorporation of manufacturer data in the file
"heatpump.dat" is not an easy task.

I have been using TYPE127 for a number of years
(which I beleive is the predecessor of TYPE504).
I think that TYPE127 was created based on available
coefficients from one manufacturer (ClimateMaster) .
It is sometimes difficult to use other manufacturer's data
because it is not presented the same way.

There was a fundamental change in moving from Type 127 to Type 504 - namely in the way the model reads the performance data. In the older version that Michel is using you had to create a 'normalized' data file such that the model could scale the results quickly to any size heat pump. The normalization process, while not difficult, wasn't a simple process either. For this reason the Type 504 model relies on catalog data files that pertain to one size heat pump. It is now MUCH easier to enter the required information and it takes about 5 minutes to enter a brand new heat pump file into the program.

The other problem you will run into with TYPE127 (or TYPE504)
is the time of operation of the heatpump during a time step.
The way it is setup right now, the heat pump is either on or off
during a time step which makes it difficult to control. For example,
say that you have a cooling load of 1kw, a heat pump capacity of
10 kW and a one-hour time step, if your controller send a "on" signal to
your heat pump then you will have a 9 kW excess capacity given
to your building during one hour.

This is the fundamental approach of "temperature-level" control in TRNSYS; that equipment is either on or off - just like in real-life. The consequence of having dynamic interaction is that it requires timesteps of less than 1 hour. I personally use 5-minute timesteps whenever possible and am considering changing my personal default to 1-minute - the speed of todays PCs makes this feasible.

The solution to this problem is to use
smaller time steps but you run into problems if you go down below
15 minutes time step with TYPE56.

The only "problem" with running sub-15 minute timesteps is that the transfer functions for some walls in Type 56 cannot be calculated with a transfer function time-base of less than 15-minutes. Simulations then run at less than 15 minute timesteps then use the same wall functions for 15 minutes instead of recalculating the effects every timestep - a very small consequence in an annual simulation.

Jeff Thornton
TESS