Measuring apparatus

Figure 24: Insulated box for measurements of reference temperatures

Following connections have been made: One regular thermocouple type-T was connected directly to the channel on multiplexer board very close to the point where the thermistor measuring reference temperature was situated. This was done for each of four used multiplexers. Those four thermocouples were inserted into the external box to measure reference temperatures of used multiplexers. Additionally, the temperature in the insulated external box was checked with two pieces of platinum resistance thermometers with accuracy of 0.1 °C (PT100 sensors). Reference thermocouple with the smallest deviation from PT100 sensor’s reading was used as our main reference temperature. The temperature difference between chosen thermocouple and PT100 sensor was 0.1 °C being acceptable for purposes of this project. All the measurements at points of interest within the test room and plenum were then made by connected T-type thermocouples and were related to the main reference temperature. The electric potentials from all thermocouples were measured and the resulting temperatures were calculated according to the Eq.(1), Eq.(2), Eq.(3) which express 3^𝑟𝑑 order power polynomial equation developed by Tomas Lund Madsen [56]. At known reference temperature t_ref and measured electric potential S, the algorithm allows to calculate the required temperature t directly.

𝑇 = 𝑇_𝑟𝑒𝑓+ 𝑇_𝑑 Eq.(1)

𝑇𝑑= 𝑆(25,9 − 0,06 𝑇𝑚+ 0,00027 𝑇𝑚2 − 0,000001 𝑇𝑚3) Eq.(2)

𝑇_𝑚= 𝑇_𝑟𝑒𝑓+ 12,95 𝑆 Eq.(3)

40

Where: 𝑆 is measured electrical potential [mV], 𝑇𝑟𝑒𝑓 is measured reference temperature in insulated box [°C], 𝑇_𝑚 is auxiliary temperature value [°C], 𝑇 is resulting temperature [°C].

Resulting set-up led to better accuracies during our investigations. Heller investigated accuracy of Lund Madsen’s algorithm and proved that absolute temperature deviation was maximum 0.27 K for room temperature range of 10 °C to 30 °C compared to frequently reported 1 K for direct connection of multiplexer [56]. The scheme of used connection of thermocouples is depicted in Figure 25.

Figure 25: Schematic connection of thermocouples

Other measure was taken to reduce errors during our measurement. The thermocouples situated in test room and plenum were not led directly to the multiplexers, but instead those were wired the shortest possible way to bus bars where we connected them to copper multi-conductor cables leading to multiplexers, see Figure 26.

41

Figure 26: Connection of thermocouples to multi-conductor cable

A single constantan cable led from each bus bar alongside to the shielded multi-conductor cable to particular multiplexers creating a circuit for each connected thermocouple. In this way we were able to minimize the undesirable interferences from surrounding environment, such as noise in electromagnetic field. This installation was feasible due to the fact that Agilent works as switch unit, measuring a single channel at the time. This installation saved almost 500 m of rather expensive copper-constantan wire.

Altogether, 75 T-type thermocouples were installed in the room and four 20-channel multiplexers were used to connect all the used thermocouples. Fifteen T-type thermocouples were installed at different heights on three moveable, vertical stands situated within the occupied zone of room. Thermocouples were fixed at height levels 0.1 m; 0.6 m; 1.1 m; 1.7 m and 2.1 m respectively, in order to be able to see vertical temperature profile which is used as a factor of local thermal comfort evaluation in ISO 7730 [29].

Furthermore, 4 thermocouples were placed on floor and 2 thermocouples measured temperature of window’s surface.

There were 36 thermocouples installed in the plenum in total; 3 horizontal planes with 12 pieces at each plane covering the area of plenum in 3 different height levels. Each plane had 4 rows of 3 thermocouples. In the top plane there were thermocouples measuring the surface temperature of the concrete hollow core ceiling deck, in the middle plane thermocouples measuring the air temperature in the plenum, and in the bottom plane thermocouples measuring the surface temperature of the porous suspended ceiling.

Additional 12 thermocouples were placed on bottom side of porous suspended ceiling. Temperatures were measured continuously for the entire period of experiment and recorded once in a minute interval.

The precision of measurements could be influenced by differences of temperature between soldering point and thermocouple wire close to the soldered end. This is why each of thermocouple installed in the room

42

and plenum was coiled about 10 cm from soldering point to eliminate this co-called diffusion error which is otherwise difficult to spot [58]. Thermocouples measuring the air room and plenum temperatures were placed inside the 5 cm long aluminum cylinder working as protection against effect of radiation, see Figure 27. In this way the influence of radiation from surrounding surfaces could be eliminated to very high degree. Influence on results caused by radiation is most often error made during measurements of room air temperature [58].

Figure 27: Aluminum cylinder used for shielding of thermocouples

Wall surface temperatures were measured with 12 thermocouples. The T-type thermocouple works in this solution as a contact thermometer. According to ISO 7726, the use of contact thermometer can lead to the measurement errors since contact thermometer can change the heat exchange between the measured surface and the environment, especially when properly shielded from environment [59]. The surface thermocouples were shielded against effects of thermal radiation by aluminum foil with dimensions of 5 cm x 5 cm. This is however needed in order to eliminate the direct influence of environment on measurement.

In this project however, the measured surface has rather high thermal conductivity (λ=2,6W/mK) which eliminates mentioned error. The rule of thumb is that heat exchange between surface and sensor is much higher than heat exchange between sensor and environment. This was achieved by using special, thermally very conductive paste to attach the thermocouple to the measured surface in order to enhance the heat transfer and by using only small area of covering foil.

Furthermore, the air temperatures in the room were also measured on moveable stands by SensoAnemo5100SF transducers with accuracy of measurement of 0.2 °C [60].

Outside air temperature measurements were made by sets of HOBO Data loggers (model U12) with temperature accuracy ± 0.35 K [61].

43

6.1.2 Air velocity measurements

Air velocities in test room were measured by use of air velocity transducer SensoAnemo5100SF with accuracy of ±0.02 m/s and ±1% of reading for velocities below 5 m/s [60]. Measurements were carried out in the occupied zone on positions of vertical stands at heights 0.1 m; 0.6 m; 1.1 m and 1.7 m respectively.

Sufficient times were allowed for the air in the room to stabilize before any readings were taken.

6.1.3 Tracer-gas measurements

In order to find a pattern of ventilated air distribution in test room, the thorough investigations with use of tracer gas were performed. The units from producer Innova were used for those purposes [62]. The tracer gas was dosed and sampled by unit Brüel & Kjær Innova 1303 (multipoint sampler and doser), and its concentration was measured by use of Innova 1312 (photo acoustic multi gas monitor). Dozing and sampling tubes were firstly connected to special filters in order to prevent dust entering the sampler unit which could cause its malfunction and then were connected to the units. The dosing pipes with diameter of 3 mm were used. Experiments were carried out using Freon R134a (C2H2F4) a tracer gas. Freon has advantage over other gases (such as CO2) that it is usually not present in outside air (zero concentration), which can help to avoid any discrepancy when analyzing the results of the measurements. This feature makes the results of investigations more precise. The accuracy of dosing is ±2%.

6.1.4 Pressure drop

The micromanometer FC0510 from Furness Control with the accuracy of 0.25% was used to measure the pressure drop between plenum and investigated room [63].

6.1.5 Thermo-graphic investigation

Thermo-graphic investigation with use of radiometric thermal camera Hot-Find D was carried out to investigate the proper functioning of the radiant systems [64]. Rather large temperature differences between walls with activated radiant systems and surrounding internal surfaces allowed us to uncover any problematic areas. Used thermo-graphic camera operates at temperature range from -20 °C to 250 °C with measuring accuracy ± 2 K and ± 2 % of reading.