A dynamic model of ammonia production within grow-finish swine barns
Ammonia is a nuisance gas in many swine barns. The overall objective of this research project was to model ammonia formation and transmission processes in a grower-finisher swine barn, by first modelling the ammonia production and emission from urine puddles on the floor surface and the ammonia emission from the slurry pit, and then incorporating these emission rates in a dynamic model that separates the room and slurry pit headspace as two separate, but linked, control volumes. A series of studies were conducted to gather more information about the processes affecting the ammonia emission rate from the floor surface and the slurry that were later included in the overall room model developed. The model was then used to investigate ammonia reducing techniques and technologies based on the understanding of ammonia production and transmission incorporated in the model. The first step in modelling the ammonia emission rate from the floor surface was to determine the frequency of urinations by grower-finisher pigs. Male and female pigs were observed three times during their finishing phase to determine their urination frequency over the course of a day. The average measured urination frequency was 0.62 ± 0.11 urinations pig-1 h-1. A sinusoidal dromedary model was developed to describe the daily variation in urination frequency for male and female pigs between 51 and 78 kg.In order for the deposited urinations on the floor surface to emit ammonia, the urea in the urine must first be converted to ammonia and the urease enzyme catalyzes this reaction. Two methods, a fixed-time-point method using the indophenol assay for ammonium-nitrogen analysis and a continuous method using the coupled enzyme assay, were used to measure enzyme activity at the floor surface of a swine barn and were compared to reported urease activity levels in the literature. Using both methods, there appeared to be an ammonia-producing site on the floor surface or within the collected samples that made accurate measurements of urease activity impossible. A review of urease activity levels in the literature from dairy-cow houses suggest that urease activity will be lowest following floor-cleaning and increase quickly following fouling of the floor surface. Based on the literature review, a urease activity value of 5 g NH¬3 m-2 h-1 was suggested for use in ammonia emission modelling of fouled floor surfaces in swine barns until better measurements become available. The ammonia emissions from 36 simulated urine puddles under a variety of temperature, air velocity and initial urea concentration conditions were measured in a bench-scale experimental set-up. The measurements were used to calibrate and validate a dynamic, mechanistic, urine puddle emission model that considered the processes of evaporation, urea conversion, change in liquid concentration and puddle pH in order to simulate the amount of ammonia emitted from a puddle. Based on the correlation coefficients (R) between measured and simulated values for water volume (R=0.99), total ammoniacal nitrogen concentration (R=0.90), and total emission (R=1.00), along with five other statistical tests for each simulated variable, the model was deemed accurate. The measurements and simulations in this experiment showed the impact of puddle pH, urease activity and changing environmental conditions on the average puddle emission rate. Puddle emission continued to occur as long as there was still water. The impact of different slurry compositions on the ammonia emission rate from slurry pits was tested in another bench-scale experimental set-up with emission chambers. The emission chamber concentration data collected was used to calibrate and validate a developed slurry emission model. The collected slurry samples were concentrated mixtures of urine and feces from individually-housed animals fed different diets. An empirical equation was developed to express the amount of total ammoniacal nitrogen in the slurry that was in the form of ammonia (f) and thus volatile to the surroundings. Based on the empirical equation, the simulated value of f was between 0.03 and 0.08 and did not show the sensitivity to slurry pH that has been reported by other authors. The slurry emission model with the empirical equation for f was validated with ammonia emission measurements from eight different slurry samples and simulated hourly concentration measurements within 17% and five-day average concentration measurements within 3%. Further testing was recommended to ensure the model developed for concentrated manure in this study was applicable to the more dilute slurry found in swine barns. Using the information gained in the previous experiments, a mechanistic model describing the dynamic ammonia concentration in the room and in the slurry channel headspace of grower-finisher swine barns, as well as the ammonia emitted to the surrounding environment was developed. Data was collected from two grower-finisher rooms to use as input data to the model and for calibration and validation purposes. The model calibration procedure determined that the amount of emissions originating from the slurry for the simulated room conditions was generally less than 5% of the total room emissions, the air exchange rate through the slatted floor was approximately 4% of the room ventilation rate, and that in the first two weeks of animal activity in a room the urease activity at the floor surface will increase. The model was validated using separate data from that used in the calibration process. The model simulated hourly room concentration levels within 2.2 ppm and 3-day average concentration levels within 1.6 ppm. The model simulations were more accurate for one room that was fed a typical grower-finisher diet compared to another room fed an experimental diet with lower protein content and sugar-beet pulp inclusion. The dynamic model was tested for its sensitivity to various input factors in terms of the floor emission rate, slurry emission rate and total emission rate. An interesting aspect of the simulations was that increases in either floor or surface emission rate were compensated to a small extent by decreases in the other emission rate as a result of a reduced concentration gradient for mass transfer. The ammonia emission rate from the floor was most sensitive to changes in urease activity, fouled floor area and puddle area. The ammonia emission rate from slurry was most sensitive to changes in slurry pH. The impact of input variables on the total emission rate was dependant on the simulated proportion of the total ammonia emission coming from either the floor surface or slurry channel. Three ammonia reduction techniques were tested and evaluated on their impact to the total ammonia emission rate from a room compared to a given set of control conditions.The work in this thesis highlighted the importance of ammonia emission from the floor surface. The proportion of ammonia originating from the slurry and from the floor surface respectively will vary on the specific conditions within the barn, and will impact the effect of any ammonia mitigation technique that is investigated or used.
model validation, convective mass transfer, slurry, urine puddles, modelling, ammonia, urination frequency, urease activity
Doctor of Philosophy (Ph.D.)
Agricultural and Bioresource Engineering
Agricultural and Bioresource Engineering