Browsing by Person "Junge, Melanie"

Now showing 1 - 1 of 1

Verhaltens- und Gesundheitsmonitoring für die Gruppenhaltung tragender Sauen
(2015) Junge, Melanie; Jungbluth, Thomas
The number of pig farmers in Germany with breeding sow herds has continually re-duced over the past 15 years. Simultaneously, herd size has increased. This trend is intensified through the introduction of mandatory group housing for gestating sows. While larger herds represent high demands on management and monitoring of indi-vidual animal health, the situation also offers potential for automation of work proce-dures or in data recording as part of indicator-based systems. The primary objective of this work was conception, implementation and evaluation of a monitoring system for determining health and behaviour deviations of gestating sows in group housing. Hereby, sensor and data technology infrastructure was to be designed for recording animal-individual indicators as feeding or drinking events as well as minimum distances travelled within a sow gestation stable for a large dynamic group. Materials used were the available RFID technology of two electronic feeding stations (EFS) and a boar recognition system. These were supplemented by the mounting of additional RFID antennae besides drinkers and on the two doors between exercise and indoor areas. For determining volumes of water metered out, flowmeters were fitted in the supply pipelines for all eight drinkers. The EFS data protocols were used for assessment of feeding events and calculation of relative eating ranks. Over the combined time-related sequencing of registration of the 13 RFID antennae within the sow gestation stable, the animal-individual minimum distances travelled were calculated on a daily basis. Animal-individual assessment of health status and changes in behaviour of the sows was conducted as part of an observation study. Examined were relationships between the automatically recorded indicators feeding events, drinking events and distances travelled, in addition to changes in health and behaviour were examined. The potential for implementation of a monitoring or prediction model was then assessed. From 13.04.2012 to 31.05.2013, recorded and analysed were 29,552 day data sets from 199 gestating sows (parity 2 to 11). In this period, only a few effects on sow health through cases of disease were documented. However, during the twice-weekly gait assessment of the sows via locomotion scoring, some cases of medium to serious lameness were identified. During 372 days, 69,577 drinking events with water volumes of from 2 ml to 11.45 l were recorded. On average, the sows drank 0.53 l water 2.4 times per day although, for 25% of the daily data sets, no drinking events were determined. In addition, a clear 24-hour cycle of water consumption was established with maxima during morning and afternoon. Where sows did not take feed from the EFS, they then tended to drink less water from the drinkers. Parity and temperature differences appeared to have only a limited influence on water consumption. Contrary to this, the sows identified as lame showed a highly significant reduced count of animal-individual drinking events, metered amounts of water at the drinkers, lower calculated distances travelled and time spent at the boar recognition in comparison with sows showing no symptoms of lameness. Detection of sows returning to heat via observation of animal-individual periods of attendance at the boar recognition and then applying a threshold value model proved very effective. Comparing standard-behaviour sows and those returning to heat also led to identifying a tendency towards reduction in metered drinking water consumption and number of drinking events for the latter. The assessments of feeding sequences at the EFS and calculation of relative eating ranks gave no significant differences between non-standard behaviour and standard behaviour animals. A tendency for influences of age, lameness and return to heat on EFS feeding sequence, or on relative eating ranks, was observed. For locomotion behaviour of gestating sows kept in large groups, reference values for possible distances travelled could be collected. Up to now, only very little literature was available on this subject. Management-associated measures such as the integration of new animals in the group appeared to have very little influence on the investigated indicators. The feasibility of animal-individual monitoring through real time processing of sensor data recorded from a group of gestating sows and of integration with a management software program, could be demonstrated. In total, however, the very large animal intra and inter variability for the parameters drinking events, feeding events and minimum distances travelled complicated the definition of the individual-animal monitoring models for indicator-based early identification of health problems. Conceivable here are further follow-up investigations regarding indicators, sensors and assessment algorithms.