CFD-modellering van pesticidedepositie en -drift in fruitboomgaarden

The Environmental and health risks of pesticide application have been the subject of research for more than fifty years and still continue to be a great concern. Ground/surface water contamination and bystander exposure from pesticide drift occurring during and after application and health risks of excessive pesticide residues detected on agricultural products are continuously in the spotlight. Orchard spray application received special emphasis due to the relatively large amount of pesticides applied and the complexity of the process involved. Both experimental and modelling approaches have been used in the past to determine pesticide drift for regulatory risk assessment studies. However, the complexity and variability of the drift process and the efforts needed for a proper analysis remain considerable hurdles in development of approaches to mitigate negative impacts of spraying. The main objective of this thesis was, therefore, to contribute to the understanding of spray deposition and drift in orchards by developing a mechanistic drift model of the air assisted spray process using computational fluid dynamics (CFD). An integrated research approach combining experiments and simulations was used throughout the thesis. Spray nozzles and three different air-assisted orchard sprayer designs (a cross-flow sprayer, an axial sprayer and a sprayer with individual spouts) were first characterized under controlled laboratory conditions in terms of their spray liquid distribution and air flow pattern. The results obtained from this indoor trials were used to validate a CFD orchard spray model. The results obtained from both the experiment and the CFD model showed important differences in air flow pattern and liquid distribution generated by the different sprayer designs. It was also seen that the liquid distribution follows the air flow pattern which is strongly influenced by the architecture of the air discharge unit on the sprayers. The CFD model was further used to optimize the outlet design of the sprayer with individual spouts. The measurement and modelling of the spray distribution was done at this stage using a vertical patternator wall rather than the actual trees. An extensive field trial was then performed using the three air assisted orchard sprayers designs and four different training systems of apple and pear to investigate the effect of tree architecture on spray distribution and to check whether the differences in air flow profiles among the three sprayer designs affect the on-target spray distribution. The results obtained showed that total leaf cover, leaf wall porosity and tree volume strongly affected the total on-target deposition. There was a strong relationship between the vertical leaf deposition profile and the air flow pattern from the sprayers. Predicted vertical deposition profiles from a canopy CFD model of spray application in orchards that incorporates the actual tree architecture, were validated against experimental profiles. This CFD model was then used to assess the effect of wind and sprayer design on spray distribution within the four training systems. The analysis performed on the axial sprayer using the CFD model showed that wind blowing in the direction of spraying actually gave the largest deposition on the tree, constraining the spray in the canopy region. A wind direction opposite to the spraying direction, however, resulted in an increase of the ground deposition and the amount of spray remaining in air. Further analysis on the effect of increasing the magnitude of a cross flow wind showed an increase in the off-target spray distribution. The CFD model was further elaborated to predict spray drift from orchard sprayers away from the orchard. This model was validated against drift measurements from an apple orchard with different nozzle arrangements. The model predicted the transient spray profile, calculates the fraction captured by the trees, the airborne spray and the ground deposition up to 40 m from the last tree row in agreement with the measured profiles. This model was then used to evaluate the effect of drift reducing nozzles and fan speed on drift. Drift reducing nozzles reduced the drift distance (the maximum downstream distance droplets travel behind the target canopy), but increased near-tree ground deposition. Further analysis done using different nozzle arrangements showed that the increase in ground deposition near the tree can be avoided (keeping the reduction in the drift distance) by combining the drift reducing nozzles with the standard ones. A reduced sprayer airflow resulted in further reduction of the percentage drift. The analysis showed that the model can be used to analyse different drift mitigation strategies and perform sprayer optimization. Future investigations could focus on developing a database of drift curves that are obtained from in silico experiments performed under different orchard spraying scenarios using the developed model. These drift curves could then be used to develop online drift monitoring tools or combined with recommended drift mitigation scenarios to be used by applicators.status: publishe