Development and field test of an on-line computerized instrumentation system for air velocity , temperature and differential pressure measurements in poultry houses

Experimental research on airflow patterns and thermal comfort in livestock buildings require the development of an on-line multi-point computerized system for environmental measurements. The characteristics, development and tests of such equipment are described in this article. The system was based on a laptop, a data acquisition card with 16 input channels and a set of air temperature, air velocity and differential pressure sensors. Sensors were resistance temperature detectors for temperature and air velocity measurements. The accuracy of the computerized sensing system after calibration was, for air velocity measurements, 0.05 m s in the range of 0.1 to 1.5 m s, and from 0.03 to 0.14 m s in the range of 1.5 to 7 m s; ± 0.2°C for air temperature measurements; and ± 5 Pa for pressure measurements. The system has been used in several research studies in commercial poultry buildings and the developed system has been proven to be robust and stable. Additional key words: air velocity, measurement, poultry, temperature, ventilation.

However, when the outdoor temperature and humidity is high (e.g.T = 30°C, RH = 70%), the evaporative cooling systems cannot cool the incoming air adequately, the micro-environment around the birds becomes too hot and humid, and animals are heat stressed.In these conditions, a uniform and high air velocity over the animal can be very effective for increasing the convective heat losses from the animals, reducing the effective temperature and improving their thermal comfort (Loot et al., 1998;Furlan et al., 2000;May et al., 2000;Yahav et al., 2001;Yanagi et al., 2002;Simmons et al., 2003;Tao and Xin, 2003;Yahav et al., 2004).
Poultry farms with inadequate ventilation systems suffer higher mortality rates when indoor air is hot, humid, and still in the zone occupied by animals (Tao and Xin, 2003).Air velocity and temperature uniformity in the animal occupied zone is also important to prevent animal migration into better ventilated but already crowded areas, which also contributes to increase animal mortality (Tabler et al., 2002).Determining the uniformity in the vertical direction is also important to ensure that environmental control sensors (usually located about 50-70 cm above floor) reflect the conditions at bird's level (Wheeler et al., 2003).However, the ability of standard environmental control systems and building designs, to create suitable air velocities in the animal occupied zone in poultry farms, is not well documented.
Air velocity distribution can be evaluated from direct aerodynamic analysis by means of air velocity measurements in commercial farms (Boon and Battams, 1988;Lee et al., 2003;Wheeler et al., 2003).However, several issues have to be considered when measuring air velocities in ventilated animal houses.
First, broiler buildings in Europe and USA tend to large structures housing many animals.The evaluation of these large structures must be carried out by measuring environmental variables (including air temperature and velocity) at enough points inside the indoor space.Airflow in ventilated rooms is in general turbulent and therefore it requires continuous measurement of air velocity during a certain time for subsequent calculation of averages.Therefore, simultaneous measurements at different locations in the animal building are necessary for determining the spatial distribution of these variables inside the building in an efficient way and within a reasonable time.Earlier works on airflow patterns and temperature in livestock buildings under Mediterranean conditions (Blanes-Vidal et al., 2001) have also revealed the importance of having a computerized system to simultaneously measure environmental variables at many points inside livestock buildings.
Second, direct measurement with portable hand-held anemometers requires an operator, which unavoidably interferes with the air velocity and therefore distorts the measurement output.In this sense, Wheeler et al. (2003) studied air temperature and velocity uniformity in tunnel and conventional ventilation broiler houses, and stated that air velocities measured by portable hot wire anemometers were highly influenced by the operator who used the anemometer.In that study, they concluded that an automatic velocity measurement system would be an important improvement.
The differential pressure between indoor and outdoor air is the variable most commonly used in Mediterranean poultry houses, to adjust the opening of inlets, in order to get a suitable air velocity at the animals' level.Therefore, the computerized system also included differential pressure sensors.This article describes the characteristics of the developed on-line (16 input channels) measurement system, which can be used as a basis for the development of a measurement system with a larger number of sensors, which would be just a repetition of those described here.
The specific objectives of this work were: (1) to develop an on-line multi-point computerized system to measure and monitor, air velocity, temperature and differential pressure inside poultry buildings; (2) to measure air velocity, temperature and differential pressure in a commercial poultry house, in order to test the measurement system under practical conditions, and to carry out a field evaluation of the climate comfort of the animals.

Description of system components
The measurement system was based on a portable computer, a data acquisition card (DAQ), and a configurable set of sensors and associated electronic circuits (Fig. 1).Sensor instrumentation and some additional circuits were developed specifically for this application.Signals from sensors were buffered, amplified, and then, transmitted in current mode to the acquisition point, where they were converted to voltage according to the voltage range indicated in the DAQ datasheet (from 0 to 10 V).
The system was designed to measure air temperature and air velocity, although it admitted other auxiliaries' inputs, such as data from differential pressure sensors.Table 1 summarizes the type and quantity of the key components used in the instrumentation system.They are briefly described below.

Portable computer and data acquisition card
A laptop and a data acquisition card were used for continuously recording the measured data.The basic specifications of the data acquisition card (DAQCARD-6024E) were: 16 single-ended analog input channels, 12 bits A/D converter, and maximum sampling frequency of 200 kHz.

Software development
Specific software for monitoring and acquiring the data was developed based on LabView 6.0.(National Instruments, 2004).The developed software works as follows: a certain number of data (n) is acquired through each channel of the DAQCARD, with a sampling frequency of ω.After a period of time t, the software calculates the mean and standard deviation of the n measurements.Means and standard deviations of successive periods of t seconds, are collected, and can be saved at the end of the experiment.All three parameters (n, ω and t) can be defined by the user in the front panel of the software (Fig. 2).

Temperature sensors
Temperature sensors were resistance temperature detectors (RTDs).Each sensor consisted of a thin layer of platinum deposited on a small ceramic substrate, of 2 × 10 mm.This element complies with DIN 43760 and BS 1904 standards.Voltage difference measured at a low constant current is related with ambient temperature.Self-heating was below 100 µW with a current of 1 mA.
Temperature sensors were calibrated by means of a set of three standard resistors (precision of ± 1 mΩ) equivalent to three temperatures (0°C, 27°C and 40°C), according to DIN 43760.Due to the narrow range of ambient temperature, a linear response for the sensor was assumed.The calibration line had a correlation coefficient equal to 1.The accuracy for the computerized sensing system after calibration was ± 0.2°C for air temperature measurements.

Air velocity sensors
Air velocity was estimated by means of constant temperature hot wire anemometry.An RTD sensor (the same as for temperature measurements) was maintained at a constant temperature of 60°C.The voltage (U) applied to the sensor needed to maintain this temperature under different conditions of flow is related to air velocity (ν) over the sensor, according to King's law (Eq.[1]): where T w is the temperature of the sensor (60°C), T is the air temperature, a and b are constants related to convective effect and K is a constant dependent on the electronic instrumentation.T w is set equal to 60°C so that the temperature difference between T w and T under normal conditions is in the 30°C to 100°C range (Webster, 1999).Equation [1] can be generalized using ν m , with  an exponent m slightly inferior to 0.5 (Zhang et al., 1996).The value of the exponent and the other coefficients were experimentally determined for each individual sensor and its associated circuitry.The anemometers were calibrated in an air steady flow rig unit.The test unit included a flux generator (airflow range from 5 to 300 kg h -1 ), regulating valves, electronic pressure gauges (Digima FP autozero, Special Instruments, maximum error 0.5%), and calibrated nozzles (ISO 5167-1).To ensure the proper formation of the velocity profile, a straight duct section was provided upstream of the anemometer location.The calibration was carried out by placing the RTD sensor in the centerline of a calibrated nozzle, measuring the voltage (U) applied to the sensor and calculating the air velocity at the centerline from the pressure drop across the nozzle (Goldstein, 1996).To validate velocity measurements, initial data were transformed according to the Eq.[2]: [2] where T is the air temperature measured by the temperature sensors, and T w and K are constants for each sensor.Then, a linear regression was performed between X and Y variables.The obtained regression coefficient R 2 for each air velocity sensor was higher than 0.996 (Fig. 3).Air velocities measured by the developed system compared to reference air velocities are given in Figure 4 for two velocity ranges, 0.1-1.5 m s -1 and 1.5-7 m s -1 .The accuracy for the computerized sensing system after calibration was, for air velocity measurements, 0.05 m s -1 in the range of 0.1 to 1.5 m s -1 and from 0.03 to 0.14 m s -1 in the range of 1.5 to 7 m s -1 .
On-line instrumentation system for environmental measurements in poultry houses 573

Differential pressure sensors
Differential pressure sensors consisted of an integrated silicon gauge sensor.External instrumentation was exclusively for amplifying the voltage to the level required by the data acquisition card.The accuracy for the computerized sensing system after calibration was ± 5 Pa for differential pressure measurements.

Experimental building and animals
Air velocity, temperature and differential pressure in a commercial poultry house located in the Valencian Community (Spain) were measured by the computerized sensing system in a field experiment carried out in summer season (June).The experimental building was a typical commercial Mediterranean poultry farm (Fig. 5), equipped with conventional cross-ventilation by negative pressure.Building dimensions were: length, 69.8 m; width, 15 m; side-wall height, 2.36 m, and maximum distance from floor to ceiling, 3.94 m.The animal house had 56 sidewall prefabricated inlets (Gasnet S.L., Villarreal, Spain) located 1.8 m above floor.Nine exhaust fans were located in the opposite wall: six were type a (Model Euroemme EM50-1.5 CV, Munters Europe, Sollentuna, Sweden) with a diameter of 0.63 m, and an airflow rate of 36,000 m 3 h -1 ; and three were type b (Model FC 063-0.7 kW, Ziehl Abegg AG, Künzelsau, Germany), with a diameter of 0.32 m and ventilation rate of 10,500 m 3 h -1 .The poultry house was occupied by broilers aged 50 days.The stocking density was of 12 animals m -2 and the average weight of the animals was about 2.8 kg.

System setup and measurements
The experiment was carried out under high ventilation rate conditions, which were: Fans 1, 3, 4, 5, 6, 7 and 9 running, and inlets opened 75% of the maximum (inlet slot 17.2 cm).Measurements were taken in the test zone shown in Figure 5.
The animals' comfort in commercial occupied broiler buildings has to be evaluated from measurements taken at birds' level.However, measurements are usually taken at higher height above floor, out of birds reach, due to the difficulty of measuring close to animals, especially when sensitive sensors, such as hot-wire anemometers, are used.In our field test, three air velocity sensors and three air temperature sensors were placed The measurement post was sequentially located at nine coordinates of the plane xy shown in Figure 7, taking measurements at one xy location at every time, during 6 min per location.Measurements were taken at the centreline of the building (A2, C2, D2) and aligned with the drinker lines closer to the side walls (B1, C1, D1, B3, C3 and D3) (Wheeler et al., 2003).The measurement time was similar to that used by Demmers et al. (2000).The parameters n, ω and t were set equal to 10, 0.2 Hz and 360 s, respectively.Therefore, at each location, the measurement post took 720 measurements per sensor.The remaining temperature sensors were located at two of the inlets (points B0 and D0) and two of the exhaust fans (points B4 and D4), and the air velocity sensors at two inlets (points B0 and D0).The differential pressure sensors were placed at points B0, D0, B4 and D4.

Statistical analysis
The effect of the location in the building (x, y and z coordinates) on the indoor air velocity and temperature was evaluated by stepwise multiple regression analyses.Linear and quadratic effects were considered, and the p-value was set at 0.05.The analyses were carried out in MATLAB.

Results and discussion
Table 2 shows mean and standard deviation of measured air velocity at inlets, and differential pressure   The statistical analysis on air velocity showed a variation of air velocity with the distance above the floor (z 2 ) and the distance to the inlets (y and y 2 ) (Table 3).The distance along the longitudinal axis of the building (x and x 2 ) did not have a significant effect on the air velocity.The reason for this is the high number and uniform distribution of air inlets along the longitudinal axis of the building, as the uniformity of air distribution throughout the building, depends primarily upon the location and size of the air inlets (ASAE, 2001).
Regarding the vertical profiles, air velocity generally was highest at 2 m.Air velocity at 0.6 m above the floor was very similar to air velocity at bird's level when measuring at the centreline of the building (A2, C2, and D2) and close to the outlets (B3, C3 and D3) (Fig. 8).However, close to the inlets wall (B1, C1 and D1) air velocity at 0.2 m s -1 was higher than at 0.6 m (maximum difference of 0.11 m s -1 ), probably due to a vortex motion (Van Brecht et al., 2004).In a similar study, Wheeler et al. (2003) detected that in most cases, the bird level velocities were higher than their corresponding human level velocity.
Regarding the horizontal variability, air velocity did not vary along the longitudinal direction (x), but it did along the cross direction (y) (Table 3).At birds' level, air velocities were generally highest close to the sidewall where the inlets were mounted (average of 0.26 m s -1 from points B1, C1 and D1) (Fig. 9).Wheeler et al. (2003)  Table 3. Regression analysis relating air velocity (m s -1 ) and air temperature (°C) with location in the building (coordinates x, y, z)   Air velocity at point B3 (0.38 m s -1 ) was higher than at C3 (0.08 m s -1 ) and D3 (0.06 m s -1 ), as it was highly affected by the large fan located at B4 (Fig. 8).Therefore, although both fans contributed to indoor air renewal by exhausting contaminated air from the building, only the largest fan increased the air velocity around the animals (from 0.06 m s -1 to 0.38 m s -1 ).
Measured air velocities at bird's level were in general terms low in comparison to recommended values.Air velocity in the centre line of the building (points A2, C2 and D2) and close to the side wall of the fans (C3 and D3), under high ventilation rate, did not exceed the minimum recommended air velocity for broilers (0.1 m s -1 , according to Cedó, 2002).Furthermore, according to Dozier et al. (2005), broilers from 36 to 42 days of age subjected to an air temperature between 25 and 30°C, dew point temperature 23°C, and an air velocity of 2 m s -1 had improved BW gain, feed consumption, feed conversion, and incidence of mortality compared with broilers grown in still air at the same temperature.This velocity (2 m s -1 ) was not reached at any point at birds' level.These results are in agreement with previous studies (Bottcher et al., 1992;Wheeler et al., 2003), which have revealed that in general terms, air speeds measured at bird's level in conventional (not tunnelventilated) broiler houses are low (from 0 to 0.5 m s -1 ).
One of the objectives of the ventilation during summer is to mix incoming air with indoor air and remove the heat gain of the building (caused by the addition of heat transfer, thermal radiation and heat produced by the animals).As a consequence, air temperature in the building is expected to be lower at the inlets than at the outlets.The results of the statistical analysis showed that air temperature significantly varied along the y direction (Table 3), being higher at the centreline and at the outlets in comparison to the temperature at the inlets.The minimum temperature in the animal house was 23.5°C and the maximum temperature 26.3°C (Fig. 8).Air temperature at birds' level, close to the outlets (B3 and D3) was about 2°C higher than close to the inlets (B1, C1 and D1).Both temperatures were higher than the recommended air temperature for 7 weeks age broilers.In a previous experiment in a conventional ventilated house, Xin et al. (1994) measured a birdlevel temperature gradient across the house between 1.3 to 2.1°C.Temperature uniformity is important to avoid animal migration, as birds tend to congregate in the areas where the temperature is closest to their re-quirements (Xin et al., 1994).Overcrowded areas should be avoided, because inadequate space from that recommended means reduced rate of growth, thermal discomfort and poor welfare (McLean et al., 2002).Air temperatures did not vary along the vertical direction (Table 3).In connection to this, previous authors (Xin et al., 1994;Wheeler et al., 2003) also found that vertical temperature gradients were negligible in crossventilated poultry houses under hot summer conditions.The maximum difference between temperatures at 0.2 and 0.6 m high, was 0.31 °C (at point B1).

Further implementations of the measurement system
The system has been used for measuring air velocity, temperature and differential pressure in several studies carried out with different objectives, as for example, in the evaluation of the influence of the differential pressure on indoor air velocity measured at the transverse section of a typical broiler house (Blanes-Vidal et al., 2007).Furthermore, airflow computational fluid dynamics (CFD) simulations have been carried out and validated in mechanically cross-ventilated, commercial poultry buildings with multiple wall inlets by using the described system (Blanes-Vidal et al., 2001, 2004a,b, 2005, 2008;Blanes-Vidal and Torres, 2005).

Conclusions
An on-line computerized system for environmental to measure environmental parameters in poultry houses was developed.The system was based on a laptop, a data acquisition card and a set of air temperature, air velocity and differential pressure sensors.Sensors were RTD's (resistance temperature detectors) for temperature and air velocity measurements.The system can acquire up to a maximum of 16 independent signals.The accuracy of the computerized sensing system after calibration was, for air velocity measurements, 0.05 m s -1 in the range of 0.1 to 1.5 m s -1 , and from 0.03 to 0.14 m s -1 in the range of 1.5 to 7 m s -1 ; ± 0.2°C for air temperature measurements; and ± 5 Pa for pressure measurements.
A field test of the developed system was carried out in which air velocity, temperature and differential pressure were measured in a commercial poultry house under practical conditions.The measurements were used to analyse the air velocity and temperature variability inside the building.The results showed that air temperature at 0.6 m above floor can be used as an estimation of temperature at bird's level.However, air velocity measured at 0.6 m above the floor did not fairly represent the air velocity conditions at birds' level.Air velocity and temperature at birds' level ranged from 0.06 to 0.38 m s -1 , and from 23.6 to 26.3°C.Therefore, sensors for the environmental control system have to be located at suitable coordinates of the horizontal plane, in order to provide useful measurements under the experimental conditions described in this study.Air velocity under high ventilation rate conditions did not exceed the minimum air velocity recommended for 7 week poultry, and air temperature was about 5°C higher than recommended.
The field test described in this article, together with several research studies carried out in commercial poultry buildings have confirmed that the developed system is useful, robust and stable.

Figure 2 .
Figure 2. Front panel of the developed software.

Figure 3 .
Figure 3. Calibration line for air velocity sensors.p β1 : probability value for the slope.p β0 : probability value for the intercept.RSD: residual standard deviation.

Figure 4 .Figure 5 .ᮡ
Figure 4. Measured and true air velocity from the calibration of one of the sensors.a) Range from 0.1 to 1.5 m s -1 ; b) range from 1.5 to 7 m s -1 ; p β1 : probability value for the slope; RSD: residual standard deviation.

Figure 6 .
Figure 6.Measurement post showing sensors at three levels (a) and photo showing the sensors (b).

Figure 7 .
Figure 7. Measurement points in the test zone (distances in meters) and mobile post with sensors at three heights: 0.2 m, 0.6 m and 2 m.
Measured air velocities (v), temperatures (T) and differential pressure (∆P) at inlets and outlets.25.3 ± 0.4 ∆P, Pa 26.5 ± 3.1 and temperatures at the inlets and outlets.Air velocities and temperatures measured at the 27 points inside the test zone (nine locations of the plane XY where the post was placed, at three heights: 0.2 m, 0.6 m and 2 m) are shown in Figure 8.

Figure 8 .
Figure 8. Air velocity (a) and air temperature (b) measured at nine locations of the test zone, at three heights: 0.2 m, 0.6 m and 2 m.Error bars show standard deviations.
inlets' wall in comparison to the centerline or points located close to the opposite (outlets) wall.

Table 1 .
Instrumentation system components