Difference between the Simulated result and the Measured result

All about WUFI 2D
Post Reply
zhanqsh
WUFI User
WUFI User
Posts: 11
Joined: Wed Nov 13, 2019 4:23 am -1100

Difference between the Simulated result and the Measured result

Post by zhanqsh » Thu Sep 17, 2020 9:13 am -1100

Hello!
I meet a problem with the calibration between the simulation and the measurement.
I have a wall assembly monitored for 10 months. And then I had all the monitored environmental data (RH, T, rain, rind and radiation...) and the wall assembly simulated. The results generally show a good consistency with the Relative Humidity and Temperature of the measured position (interface of XPS board and Rock Wool). However, in the period when the indoor were air-conditioned, significant differences (5 to 10% RH) occur between the measured (md) RH and the simulated (sm) RH. I used the normal materials for the construction and the material data from WUFI's database for simulation. I have tried to change the properties of the coating, gypsum board and rock wool, but couldn't get a better result. I have no idea to adjust the setting in the simulation now.
Could anyone please help me with that? Thank you!

Best regards,
Qiaosheng
Attachments
幻灯片2.JPG
Wall assembly construction detail
幻灯片2.JPG (90.07 KiB) Viewed 213 times
幻灯片1.JPG
Comparision of the results
幻灯片1.JPG (128.08 KiB) Viewed 213 times

Thomas
WUFI Administrator
WUFI Administrator
Posts: 302
Joined: Sun Jun 19, 2005 10:33 pm -1100

Re: Difference between the Simulated result and the Measured result

Post by Thomas » Thu Sep 17, 2020 11:02 pm -1100

Hi Qiaosheng,

I notice that the deviations which you point out always occur during periods where the indoor temperature is held constant, and where the indoor temperature is quite different from the temperature in the component, while it follows the component temperature more closely during the rest of the time.

Your construction appears to be relatively vapor-permeable on the indoor side, so the vapor pressure at the measurement point will be close to the vapor pressure in the indoor air (you can use your measurements to check whether this is indeed true). Because most of the time the temperature at the measurement point is close to the indoor temperature (as shown by the measurements), the relative humidity at the measurement point is also quite close to the indoor relative humidity (this is also shown by the measurements).

However, there are periods where the indoor temperature is held constant and is quite different from the temperature at the measurement point. During these times, the vapor pressures of the indoor air and at the measurement point will be similar (due to the vapor-permeability of the construction), but because of the different temperatures, the same same vapor pressures will correspond to different relative humidities. This can also be seen in your data:

In the first two circled periods, the temperature in the component is noticeably lower than the indoor air temperature, and consequently the relative humidity is noticeably higher there than the indoor humidity. In the last two circled periods, the temperature in the component (red and orange) is higher than the air temperature (pink), and consequently during these periods the relative humidity in the component is much lower than the indoor humidity (with the vapor pressures expected to be very similar).

Both the measurement and the simulation show this expected behavior, but the reaction of the relative humidity is much stronger in the simulation than in the measurement. It may be possible to find the reason if you look a the vapor pressures. We expect the vapor pressure at the measurement point and in the indoor air to be quite similar. If the temperatures at these points are similar, the relative humidities will be similar, too. If the temperatures are different, the relative humidities will be different in just the right way to correspond to similar vapor pressures.

In your circled cases, we have different temperatures, but conflicting values for the relative humidity in the component. Only one of these conflicting humidity values can correspond to the vapor pressure being the same in the component and in the indoor air. By looking at the vapor pressures you may be able to find out which one of the conflicting values is the more plausible one. (This assumes that there are no other major humidity sources and the main transport process occurring is the vapor diffusion exchange between the measurement point and the indoor air.)

For example, looking at the last two circled cases: Here the simulated humidity is lower than the measured humidity. If you find that the simulated humidity corresponds to similar vapor pressures at the measurement point and in the indoor air, this may be the more plausible result, and the higher measured humidity corresponds to a measured increase of the vapor pressure above the vapor pressure in the indoor air. In this case, additional moisture may have penetrated somehow into the component, or some parts of the component (e.g. the timber) may have released additional moisture. These additional moisture sources have then not been included fully in the simulation.

On the other hand, if the measured humidity corresponds to similar vapor pressures, this may be the more plausible result, and the simulation will correspond to a temporary decrease of vapor pressure below the indoor level. One would then have to find out why this happens in the simulation.

Kind regards,
Thomas

zhanqsh
WUFI User
WUFI User
Posts: 11
Joined: Wed Nov 13, 2019 4:23 am -1100

Re: Difference between the Simulated result and the Measured result

Post by zhanqsh » Fri Sep 18, 2020 4:45 am -1100

Hi, Thomas. Thank you for your quick reply.
Unfortunately, I didn't measure the vapour pressure, but similar vapour pressure between the measured position and the interior surface during the periods were seen in the simulation result by WUFI2DMotion. However, I have measured the temperature and relative humidity of the different interfaces. And it's found that maybe the condensation played as the additional moisture in somehow. Because during the first two circled periods, the Rh of the measured position (as discussed before) was significantly higher than the opposite side of the Rockwool (contact with the gypsum board) (85% to 65% Rh in average). And during the last two circled periods, it's much lower (60% to 75% Rh in average), while the simulated Rh is also lower than the measured Rh at the interface between Rockwool and gypsum board (65% to 75% in average ).
How do you think about the possible influence of condensation, which might be underestimated in my simulation by any reasons?

Thank you!
Best,
Qiaosheng

Thomas
WUFI Administrator
WUFI Administrator
Posts: 302
Joined: Sun Jun 19, 2005 10:33 pm -1100

Re: Difference between the Simulated result and the Measured result

Post by Thomas » Fri Sep 18, 2020 6:52 am -1100

zhanqsh wrote:
Fri Sep 18, 2020 4:45 am -1100
Unfortunately, I didn't measure the vapour pressure [...] However, I have measured the temperature and relative humidity of the different interfaces.
So you can calculate the vapor pressure from relative humidity and temperature. For the interfaces and for the indoor air.

If, for example, the RH is 80 % (= 0.8 ) and the temperature is 20°C, then the vapor pressure is 80 % of the saturation vapor pressure at 20°C:

p = 0.8 * psat(20°C) = 0.8 * 2342 Pa = 1874 Pa.
And it's found that maybe the condensation played as the additional moisture in somehow. Because during the first two circled periods, the Rh of the measured position (as discussed before) was significantly higher than the opposite side of the Rockwool (contact with the gypsum board) (85% to 65% Rh in average).
Condensation and later release of the condensed moisture may be a possibility. In your diagram, however, no condensation is obvious, at least not at the measured interface, because the relative humidity never reaches 100 % there. The temperatures and the humidities at that position look quite harmless. But maybe the timber stud visible in the cross-section diagram is a thermal bridge where condensation might occur. It does not even need to be condensation, maybe some parts of the wall (like the timber) can simply take up some normal sorption moisture during more humid periods and release it during drier periods. The simulation should automatically include this effect, but maybe it does not do this well enough for some reason. And of course the problem might be something completely different after all...

Regards,
Thomas

Post Reply