In situ measurement of heat transfer coefficient on a building wall surface: h-measurement device based on a harmonic excitation

Heat transfers from the surface of a building represent a significant part of the overall building energy consumption. It has been shown that the uncertainties on the internal and external heat exchange coefficients have a large impact on building energy simulations (BES) outputs. In addition, accurate in situ measurements of building heat losses require an estimation of the heat transfer coefficients. In both situations, these coefficients are seldomly measured in situ because of the complexity of the process. Standard values and empirical correlations are rather used. While the former are often not adaptable to the situation experienced (specific geometry, wind speed, temperature gradient, etc.), the later are based on laboratory experiments which are usually not representative of conditions encountered in buildings. This makes the estimation of the total heat transfer coefficient, noted h, possibly highly inaccurate. Moreover, the uncertainties are rarely quantified.

Existing heat transfer coefficient measurement methods are usually only applicable in steady-state and are not able to capture temporal evolutions of the coefficient. In addition, they rely on the measurement of the exchange temperature (also called the “operative” temperature in the building physics community: a weighted average of the air and mean radiant temperatures). This temperature is complex to measure. It is usually supposed equal to the air temperature, which is not the case in many configurations, especially outdoor. Finally, the common approach consists in estimating the convective and radiative coefficients separately, so that the measurement uncertainties add up.

This paper proposes a novel in situ measurement method of the total heat exchange coefficient h. It is based on a periodic excitation. First, it has the advantages of being easy to implement (only one sensor to fix on the wall) and able to perform measurements within only a few minutes. Second, it allows the continuous monitoring of the h value, provided the variations of the environment temperature are not too fast. Third, it does not require any knowledge of the operative temperature. This is interesting when the air temperature significantly differs from the mean radiant temperature. This happens in non-insulated buildings for example, or in dynamic regimes when the building is being heated up or cooled down rapidly.

The sensor is mainly made out of an array of thermoelectric coolers, a heat flux meter and a thermocouple. The optimal sensor design and operating conditions (oscillation period, amplitude, number of periods to consider, etc) were determined thanks to a comprehensive analysis. The method was validated in laboratory on test cases both in steady-state and fast varying conditions. For the latter, four 500 W electrical heaters were used to heat up the indoor air of the room which generated highly off-equilibrium conditions.

The technique proved a good reproducibility. Finally the measurement uncertainties were quantified: in standard conditions, the h-value could be measured with a 0.5  W.m − 2.K − 1 uncertainty.