Convert Flow and Temperature Difference to kWh

[desperate]

Nice one Klaus,, look forward to your results, I will be away for the week, see you next weekend.

Take care

Desp

[desperate]

Hallo Klaus, any temperature readings with your new sensor positions yet? I have been scouting around for some reasonable flow meters, it looks as if about 100 quid needs to be spent to measure lowish flow rates to an accuracy of 1% or so, wow.

I am wondering weather it is worth measuring the inlet and outlet temps of the solar coil at the cylinder, and forgetting about s1 s2 etc. Those 2 readings with an accurate flowrate would enable one to calculate heat delivered to the cylinder, without all the heatlosses in the pipework etc.

Seeya

Desperate

[ericw]

Desperate
Measuring the temperature at the cylinder must give a more accurate answer.

But have you considered the effects of the errors in temperature measurement on the overall accuracy?
As you are subtracting two values any errors have a large effect.  So if you were measuring to 0.1 degree, then if you have a temperature difference of 10 degrees the error can be 2%.
Despite what the display may show, actually measuring temperatures to 0.1 degree is not a trivial task,  for instance the spec for DS18B20  is only +/- 0.5 degrees (0.2 typ) , before taking into account any errors caused by the insertion into the flowing water.

So unless you spend a lot of time (or money) doing calibration, you are unlikely to get great overall system accuracy.

[hiccup]

Hi

Yes Eric is right. That's the problem with the TDC4 which comes with a very good Grundfos flow sensor (VFS) but relies on one of the temperatures being measured to only 1C resolution, so you get very lumpy and inaccurate measurements of energy production.

(The temperature sensor built into the flow sensor is measured to 0.1C, but the Pt1000 is measured to 1C.)

Grundfos do a temperature sensor only version (RPS) that matches the flow sensor and they even do an energy meter (HMD II) although I could only find this version from Caleffi that uses the same sensors.
http://www.caleffi.us/en_US/caleffi/Details/News/files/Caleffi_WMZ-G1_EnergyMeter.pdf

Note the temperature and flow measurement accuracies quoted!

Also note the difference between accuracy and resolution. What you need in this application is accurate temperature difference, so as long at the temperature sensors are matched, and read to a reasonable resolution, the absolute temperature measurements do not have to be very accurate.


Hope this helps

Hic!

[EccentricAnomaly]

Also note the difference between accuracy and resolution. What you need in this application is accurate temperature difference, so as long at the temperature sensors are matched, and read to a reasonable resolution, the absolute temperature measurements do not have to be very accurate.

Also, finding the offset between the two sensors should be pretty easy.  E.g., put them both on the same bit of pipe for a while and plot a graph of difference in the readings against the indicated temperature.  The attributes needed for sensors in this application are resolution and long-term stability (so you don't have to recalibrate too often) rather than absolute accuracy.

[KLD]

Hallo Klaus, any temperature readings with your new sensor positions yet?

Hi Desperate

Sorry for the late reply, I've been away a lot recently.
The new sensors-in-pockets do quite well. I'll attach a few  graphs in order to highlight the main remaining issue: the correlation between temperature measurement and pump operation.
IIUC it is your goal to measure the heat transferred into the tank. First off, a transfer only happens while there is flow. Sounds trivial, and is so when the pump runs for long periods of time. The more difficult area is for intermittent pump runs. The temperatures around the HE coil and the panel in- / outlets relax during the pump's off state, and show quite drastic swings. In order to calculate reliable values for the heat transfer, your data logging will have to be faster then the typical time constant of the system (sorry for the physics speak).

To get an impression of what variations you can get in the heat transfer measurements, just compare Dhaslam's figures with mine. Two very different approaches lead to 11kWh per day per 20tubes in Ireland, versus 3.5kWh / day / 20tubes in Oxfordshire. If only we could agree on a measurement procedure that Eric's 2% errors were the limiting factor 


Klaus

[desperate]

Hey Klaus that,s brilliant, I will study these graphs more tommorow, it is getting a bit late for my old brane now I have amassed stacks of figures over the summer, I will try to plot and post them soon.

Desperate

[KLD]

What I really wanted to learn from those sensors is about how much energy is lost in the pipework. And, whether bundling the flow and return pipes together before adding insulation has any detrimental effect on the system performance.
Here are two more plots, showing the temperature differences (a) across the heat source and sink, and (b) along the pipework.

While the dT across the panel and across the coil in the store seem to mirror each other nicely, the pipework plot shows that there is a small measurement error in the flow pipe: early in the morning the water arriving at the coil appears warmer than the water leaving the panel (by 0.5 degrees or so.) Most likely this is due to insufficient insulation around the panel pipework, but could be due to sensor differences (see Eric's earlier comment about the DS18B20 one-wire sensors).
Had there been a significant heat transfer from flow pipe to return pipe, I'd have expected to see negative dT for the return pipe (ie. panel entry warmer than coil exit). This is not the case. I've got about 10m of parallel flow and return pipe wrapped inside the same insulation (rockwool).

Sanity check: from the 7.5°C across the coil and the slope of the temperature increase in the tank (0.395°C per 5min interval, equals 1.15kW heating power), I can calculate a flow rate of 2.2L/min. That sounds plausible to me.

Klaus

[desperate]

Hallo Klaus, how are you?
Good work you have done here, well worth some karma

I keep loosing posts for some reason, maybe I am tabbing around between threads too much, I will compose my thoughts first and return soon

Desperate