TECHNICAL EVALUATION OF ARTIFICIAL LIGTHING SYSTEMS FOR BROILER HOUSES

The spatial distribution of illuminance and the electric consumption of artificial lighting system is one of the main problems related to broiler production. Therefore, the aim of this study was to evaluate the spatial distribution of luminance level and energy efficiency of different lighting systems for broiler houses. Six types of lamps were tested in two different configurations to find the minimum illuminance of 20 and 5 lux. The tested lamps were incandescent (IL) 100 W, compact fluorescent (CFL) 34 W, mixed (ML) 160 W, sodium vapor (SVL) 70 W, T8 fluorescent tube (T8 FTL) 40 W and T5 fluorescent tube (T5 FTL) 28 W. The first four were evaluated with and without reflective light fixture and the latter two without light fixture. It was observed that the tested system with light fixtures negatively affected the spatial distribution of illuminance inside the house. The systems composed by IL and ML without light fixture led to better results in meeting the minimum illuminance of 20 lux and 5 lux, respectively. T5 FTL presented the lowest energy demand.


INTRODUCTION
Lighting is an important environmental factor that affects the production of animals by interfering with their physiological and behavioral responses.Its influence covers various bodily functions and features of the animals as it controls biological rhythms such as feeding, breeding, hormone concentrations, enzyme activities and metabolic processes among others (BAÊTA & SOUZA, 2010).
Light intensity affects the behavior of broilers (KRISTENSEN et al. 2007;ALVINO et al., 2009).Besides this, other illumination aspects such as photoperiod, light source, wavelength, spectral distribution and spatial distribution of the lamps in the shed affect the quantity and quality of production (BUYSE & SIMONS, 1996;LEWIS & MORRIS, 1998).According to LEWIS et al. (2007), the lack of exposure to ultraviolet rays (UV) for birds kept in a controlled environment can be detrimental to their physiology and behavior.Different from man, birds have sensitivity and ability to see in this spectral range of light radiation.
Therefore, as it is related to the development and health of birds, proper lighting helps to achieve better zootechnical and economic results from this activity (MORAES et al., 2008, KAWACHI et al., 2008), and also has possibility of reducing the electricity consumption.
In this context, evaluation, followed by an appropriate design of lighting systems for broiler houses is essential for the proper performance of birds.Among the methods for analysis, there is geostatistics, which enables the study of spatial variability in the level of luminance in the houses through kriging interpolation.Thus it is possible to obtain contour maps to understand the spatial distribution which aids in the planning and control of lighting (YANAGI JUNIOR et al., 2011, FARIA et al. 2008;MIRAGLIOTTA et al., 2006).
According to YANAGI JUNIOR et al. (2011), the spatial distribution of the variables related to the production environment provides the obtainment of more detailed information about the studied system, and, with geostatistics analysis it is possible to predict values at un-sampled points in addition to the obtainment of semivariogram models and their parameters (GOMES et al., 2007).
In addition, lighting is an important factor when reducing energy consumption.According to data from the MME (2009), Brazil's energy consumption in 2007 was 412,130 GWh, 17% of this total refers to artificial lighting (MME, 2007).The share of electricity consumption for lighting in the agricultural sector was 3.4% (MME, 2007).
In poultry houses, the electricity consumption by feeding equipment, lighting and climatization are of great importance in the composition of production costs (BUENO & ROSSI, 2006).In poultry production chain, the energy is the second most used input, where feed is the first.The lighting of the building, in most cases, is responsible for major loss (JORDAN & TAVARES, 2005).TURCO et al. (2002), by measuring the energy consumption of equipment (feeders, fans and lamps) of a poultry house, during 40 days, during the winter, found out that 55% of the total energy consumed was due to the lighting system.
The objective of this study was to evaluate the spatial distribution of the level of iluminance and energy efficiency of different lighting systems for broiler houses.

MATERIAL E METHODS
To evaluate lighting systems for broiler houses to lower power consumption and appropriate luminance levels, six types of lamps were tested, with and without light fixture, and in two different configurations, needed for different stages of the bird's life.According to BROODING FUNDAMENTALS GUIDE ( 2009), for 0-3 days old Cobb broilers chickens the luminance should be 20-60 lux, for 4-7, 20 lux, for 8-14, 10 lux and for 15-28, 5 lux.In turn, the BROODING MANAGEMENT GUIDE (2009) recommends 25 lux for chicks up to seven days; thereafter the intensity should gradually be reduced from 10 to 5 lux and 48 hours before slaughter it is recommend an increased of 10-20 lux.OLANREWAJU et al. (2006) suggests 20 lux for the first 13 days and 5 lux until the slaughter.Similarly, RUTZ & BERMUDEZ (2004) recommend 20 lux for the first six days, and then to slaughter the lighting must be 5 lux.
The criterion for selection of lamps and luminaires evaluated based on the usability in commercial sheds for breeding of broilers and economic viability.
The first four were evaluated with and without reflective light fixture type plate, and the last two only without the light fixture.The configurations of lighting systems evaluated are specified in Table 1.Data collection was divided into two steps.The first was conducted in an experimental shed like those used to grown broilers.A limited region of this experimental shed had the following dimensions: 10 m width, 36 m length and 3.6 in column height.The structural characteristics are as follows: gable roof with French-type ceramic shingles, a slope of 16.70° with regard to the horizontal plane and 1.3 m overhang.
The lighting systems were installed at a height of 2.35 m, similar to those used in some broiler houses.The criterion adopted to define the distance between the lamps was the supply of desired minimum illuminance (5 or 20 lux) where it was considered as critical points (points B, C, F and G of Figure 1) to obtain the minimum value required.Considering the overlap of the luminous flux, these are the points with lower levels of luminance in the experimental plot (Figure 1).No theoretical method for calculating light levels was adopted as literature reports some inconsistency between the data measured and estimated (DAVID, 2007;JORDAN & TAVARES, 2005).When using IL, CFL and ML lamps, three lines of lamps were required to meet minimum illuminance of 20 lux.To provide a minimum of 5 lux, the central line was turned off.When using the SVL, T8 and T5 TFL only two lines were needed for the first scenario, and one lamp (interleaved) of each line was turned off to meet a minimum of 5 lux.
The collection of these data was performed at night to avoid interference from natural light.The luminance was measured with a single digital light-meter (ICEL, model LD-510, accuracy ± 3% for incandescent lamps and ± 5% for the other) at the height of 0.30 m from the floor in a grid of points at every 1 meter, in three replicates made in sequence.The sampling area was determined to reach the established minimum illuminance (Points E and H in Figure 1).Based on the experimental data of luminance, the total number of lamps in a hypothetical commercial broiler house was determined, and the spatial distribution of luminance levels was simulated through the geostatistics techniques.In the second stage of data collection the parameters referring to energy consumption of the systems were measured.The same lamps used in the first stage were installed on a laboratory workbench.With an energy analyzer (Fluke 435, accuracy ± 0.03% for power factor, ± 0.1% for voltage and ± 0.5% for current) the voltage (V, V), current (I, A ) and power factor (dimensionless) were measured.From these data and the number of lamps previously, the estimated active demands (P, W), reactive (Q, Var) and apparent (S, VA) were calculates for a commercial broiler house.
There were three replications for each set of lamps assessed, with each test lasting 30 minutes.From these data and the number of lamps calculated in step 1 was performed to evaluate the total energy parameters considering a commercial broiler house.

RESULTS E DISCUSSION
Results of descriptive statistical analysis of the data are listed in Table 2.The results indicate large variability in the lighting distribution, reinforcing the need to characterize the spatial distribution of illuminance using geostatistics.The ideal lighting levels vary according to the bird's age.According to COBB-VANTRESS (2009a), 0-3 days should be 20 to 60 lux, 4-7, 20 lux, 8 to 14, 10 lux and 15 to 28, 5 lux.In turn, COBB-VANTRESS (2009b) recommends 25 lux for chicks up to 7 days old, thereafter the intensity should be decreased gradually from 10 to 5 lux, and 48 h prior to slaughter should be increased from 10 to 20 lux.OLANREWAJU et al. (2006) suggest 20 lux for the first 13 days and 5 lux until slaughter.RUTZ & BERMUDEZ (2004) recommend 20 lux for the first 6 days, and then until slaughter the lighting must be 5 lux.
Through Table 3, it is observed that the spatial dependence of luminance and the range parameter indicates the effect radius of each system.The sodium vapor lamp had a wider range, indicating that it is necessary to use fewer units in the shed.This result is consistent with studies conducted by JÁCOME (2009) and JORDAN & TAVARES (2005), which compared different types of artificial lighting for egg laying production.When evaluating lighting technologies for the production of chrysanthemum cuttings, DAVID & ROSSI (2010) also used fewer sodium vapor lamps than other treatments.An important aspect worth noting about the semivariogram parameters concerns the value of the nugget effect.The nugget effect measures the unexplained spatial variability between the shortest distance sampled (DIGGLE & RIBEIRO JR, 2007).In this case, it is clear that the vast majority of lighting systems presented the value for nugget effect equal to zero, showing that all spatial variability could be explained.This fact should be highlighted, because usually this value is obtained only theoretically.
However, in all cases, the nugget effect showed a value lower than 25% of the value obtained for the level ( 0 / CC ).According to CRESSIE & HARTFIELD (1996), this indicates a strong spatial dependence on the studied variable.The evaluated systems with light fixtures had lower range when compared to those without the apparatus.Therefore, despite the benefits offered by such luminaries as the direction of the luminous flux down, and decrease of its depreciation due to a reduced accumulation of dirt on the lamps, these systems have reduced the range of the lighting system.This behavior is illustrated in Figure 2. Another factor observed in this figure is the higher occurrence of high levels of luminance on systems that have light fixtures.
Figure 3 depicts the frequency of occurrences of the luminance levels from the evaluated systems, clearly illustrating the differences between systems with and without light fixtures.Systems equipped with light fixtures showed less uniformity and higher frequency of luminance above the ideal.JORDAN & TAVARES (2005) report that situations like this can be detrimental to broiler production, favoring the development of aggressive behavior, hyperactivity and cannibalism, and also causing fat deposition, higher incidence of leg problems, metabolic and circulatory diseases.JÁCOME (2009) observed a lower egg production when hens were exposed to high illuminance.OWADA et al. (2007), using the fuzzy sets theory, estimated that luminance close to 1 lux would be ideal for the welfare of broilers, between 1 and 5 lux would be viable, and other situations would not be considered great.Note: fa means fixture absence and fp is fixture presence.
In Table 4 the total numbers of lamps, voltages, currents and power factor are listed for each of the systems evaluated, simulated from the data obtained in this study.For this estimate, a commercial for broiler house was considered, with dimensions of 12x125m, and installation of lighting systems at 2.35 m from the floor.For the analysis of power factor, the Agência Nacional de Energia Elétrica (ANEEL) (National Agency of Electrical Energy) considers as the minimum value allowed at 0.92.Only compact fluorescent and sodium vapor had power factor below the limit allowed.Therefore, these systems provide higher values of reactive demand, which increases the cost of generating electricity.For industrial and rural consumers, there are specific rates and charges on the reactive power consumed, justifying the adoption of equipment with power factor close to 1.These results are similar to those observed by DAVID & ROSSI (2010) when evaluating different types of lamps for the production of chrysanthemum cuttings.The values were 1.0 for IL, 0.56 for CFL, 0.98 for T8 TFL, 0.95 for SVL.JÁCOME (2009), testing different types of lamps for housing laying hens found 0.83 for SVL, 0.58 for CFL and 0.99 for IL.
In tables 5 and 6 these demands are listed for illuminance levels of 5 and 20 lux, respectively.IL systems without light fixtures (IL 20lux,fa and IL 5lux,fa ) had a higher demand and therefore higher consumption and greater cost.This result is because this lamp has the characteristic of turning most of its energy into heat, and only the rest into light energy.IL systems with light fixture (IL 20lux,fp , IL 5lux,fp ) showed, however, lower demands than the ML systems (ML 20lux,fa , ML 20lux,fp , ML 5lux,fa , ML 20lux,fp ).Therefore, for IL, the use of light fixture decreases the demand due to the smaller number of lamps used.For the illuminance of 5 lux, SVL and CFL showed no significant difference in active demand, but as CFL has a lower power factor, which brings more reactive demand, SVL had a better result by presenting lower apparent demand.JORDAN & TAVARES (2005), analyzing different types of lamps to produce fertile eggs, verified that the sodium vapor lamp 70W offers major advantages, because it showed a 76% savings compared to incandescent 100W.JÁCOME (2009), comparing incandescent lamp 100W (control) with sodium vapor lamp 70W, found a 54.14% reduction in demand, and comparing the control with compact fluorescent lamp 23W, the reduction was 73.76%.DAVID & ROSSI (2010) found that replacing incandescent lamps 100W with compact fluorescent lamps 23W showed a 75% reduction in electricity consumption for the production of chrysanthemum cuttings, without affecting the final output.
The data in Table 6 indicate that the best results are obtained by T5 20lux,fa and T8 20lux,fa TFL systems, respectively, by presenting lower demands for electricity and higher power factor.It is presented in tables 7 and 8 the values for reduction of active and apparent demand in the evaluated systems regarding treatments IL20 lux,fa and IL 5lux,fa , which showed the highest values for these parameters.For the reactive demand the comparison was made regarding the treatment ML 20lux,fa and ML 5lux,fa .The lamp that showed the best performance was a T5 fluorescent tube, followed by T8.
FIGURE 1. Sketch of transversal (A) and longitudinal (B) elevations with the scheme used to determine the distances between the lamps.

TABLE 2 .
Mean, mediun, minimum, maximum, variance and standard deviation for the illuminance values (lux) from the evaluated lighting systems.
Note: fa means fixture absence and fp is fixture presence.

TABLE 3 .
Estimates of experimental semivariogram parameter for the variable illuminance level in the evaluated lighting systems.
1 C 0 / C = Level of spatial dependence.Note: fa means fixture absence and fp is fixture presence.

TABLE 4 .
Total number of lamps, voltage (V), current (A), and power factor simulated for a commercial broiler house, for the evaluated systems.
Note: fa means fixture absence and fp is fixture presence.

TABLE 5 .
Active (W), reactive (VAr) and Apparent (VA) demand for the evaluated systems that meet the minimum illuminance of 5 lux.Averages followed by the same letter in the column do not differ by the Scott-Knott test at 5% significance.Note: fa means fixture absence and fp is fixture presence.

TABLE 6 .
Active (P), reactive (Q) and apparent (S) demand for the evaluated systems that meet the minimum illuminance of 20 lux.Averages followed by the same letter in the column do not differ by the Scott-Knott test t 5% significance.Note: fa means fixture absence and fp is fixture presence.

TABLE 7 .
Reduction (percentage) of the maximum electric demand, compared to the IL 5lux,fa , which meet the minimum illuminance of 5 lux.
Note: fa means fixture absence and fp is fixture presence.

TABLE 8 .
Reduction (percentage) of the maximum electric demand, compared to the IL 20lux,fa which meet the minimum illuminance of 20 lux.
Note: fa means fixture absence and fp is fixture presence.