INTERNET OF THINGS (IOT) AND DAIRY FARM AUTOMATION

The objectives of the thesis were: (i) to evaluate the use of automatic systems and the related sensor-based technologies (Precision Dairy Farming – PDF - systems) in three important areas of a dairy farming; (ii) to assess different methods of estimating liner compression (LC) by using a new test device and a novel artificial teat sensor, both specifically designed and built. Four studies were carried out to achieve these goals. In the first study “Use of a proactive herd management system in a dairy farm of northern Italy: technical and economic results” the reproductive and economical performances of an AMS farm that adopted a proactive herd management system (Herd Navigator™) were analyzed. Reproductive and economic data were recorded before and one year after the installation of Herd Navigator™. Number of days open reduced from 166 to 103 days, number of days between the first and second insemination decreased from 45 to 28 days, and days for identifying an abortion were 80 % less, from 31 to 6 days. The preliminary results highlighted the usefulness of the proactive herd management system implemented for the reproduction management. A basic economic model was proposed to evaluate the potential economic benefits coming from the introduction of this technology. The model considered the benefits deriving from the reduction of reproduction problems and, consequently, of days open. Considering the effects related to the above mentioned aspects in a case study involving 60 dairy cows, a return on investment over 5 years was calculated. In the second study “Evaluation of an electronic system for automatic calving detection on a dairy farm”, a GSM-based remote alarm system for automatic calving detection was evaluated - in terms of sensitivity and PPV- as useful and reliable tool to detect the exact moment of calving in the field. Up to date, various monitoring technologies and protocols have been proposed to predict the exact moment of the calving but none of them have been adopted widely by producers due to high costs, difficulties of execution or lack of quality staff. Visual observation of the cow’s behavior is still the most frequent. The system object of the study, showed very high sensitivity and PPV, respectively 100% and 95 %, allowing the farm staff to be present at the moment of calving in 100 % of cases when cow were monitored using this system. Cows not monitored by this system, were assisted only in 17% of cases (P<0.001). The farm staff, if present during this crucial and important moment, could assist the animal preventing possible problems for the cow and the calf. This possibility would be of great interest particularly with heifers and with problematic cows. In the third study “Evaluation of the performance of the first automatic milking system for buffaloes”, the response of buffaloes to automatic milking and the related performance of the system were investigated. Automatic milking systems (AMS) are a revolutionary innovation in dairy cow farming and can now be considered a well-established technology. In 2008, automatic milking of dairy buffaloes was introduced for the first time in a commercial farm in southern Italy. The aim of this study was to evaluate the response of buffaloes to automatic milking, examining the relationships between milking interval, milk production, and milking time for this species. A total of 7,550 milking records from an average of 40 buffaloes milked by an AMS were analyzed during a 3-mo experimental period at a commercial farm with Italian Mediterranean buffaloes in southern Italy. Date and time of animal identification, milk yield, milking duration, milking interval, and average milk flow rate were determined for each milking. The results were also used to predict the maximum number of milkings per day and the optimal number of buffaloes per AMS for different levels of milk production. The average interval period between 2 consecutive milkings was 10.3 h [standard deviation (SD) 3.3]. Overall, 3.4 and 25.7% of the milkings had an interval of ≤6 h or >12 h, respectively. Milking duration averaged 8.3 min per buffalo per milking (SD 2.7). The average milk flow rate was 1.3 kg/min (SD 0.5) at a milk yield of 2.8 kg per milking (SD 1.4). Assuming that the milking station is occupied 80% of the time, the number of milkings ranged from 136 to 152 per day and the optimal number of buffaloes per AMS ranged from 59 to 66 when the production level increased from 2 to 5 kg of milk per milking. Automatic milking systems seems suitable for buffalo, opening new options for the management of dairy buffalo farms. In the last study “Methods of estimating Liner Compression” the aim was to compare different methods of estimating liner compression (LC) by using a new test device and a novel artificial teat sensor, both specifically designed and built. Liner compression (LC) is the pressure applied to the teat end when liner collapses during the d-phase of pulsation. Liners with higher LC are thought to increase the occurrence of teat-end hyperkeratosis. Overpressure (OP)has been proposed as a relative indicator of LC. By using the new test device developed, two methods of measuring overpressure were compared: liner overpressure (OP) was measured with no pulsation (OPnp) and with limited pulsation (OPlp) repeatedly on the same cow during a single milking. Each of the six liners (three round liners and three triangular liners) used in this study were tested on the same six experimental cows. OPnp and OPlp were measured on all four teats of each experimental cow twice for each liner. The order of OPnp and OPlp alternated sequentially for each cow test. The OP results for the six liners were also compared to LC estimated on the same liners with a novel artificial teat sensor (ATS). The OPlp method showed small but significantly higher values than the OPnp method (13.9 kPa vs. 13.4 kPa). The OPlp method would be recommended as the preferred method as it more closely approximates normal milking condition. OP values decreased significantly between the first and the following measurements, (from 15.0 kPa to 12.4 kPa). Thus, performing the OP test at a consistent time, one minute after attaching the teatcup to a well-stimulated teat, to reduce the variability produced by OP changing during the peak flow period would be recommend. The new test device showed several advantages over previously published methods of measuring OP. A high correlation between OP and LC estimated by the ATS was found, however, difficulties were noted when using the ATS with triangular liners.