Milk yield has an impact of success of milking events in automatic milking systems (AMS)

MASÍA, F.M.; LYONS, N.A.; PICCARDI, M.; BALZARINI, M.; RUSSELL, H.; GARCIA, S.

Palmerston North

Conferencia; Australasian Dairy Science Symposium 2018; 2018

Australasian Dairy Science Symposium (ADSS)

Automatic milking systems (AMS) provide the benefit of unassisted milking events, although a proportion might also be deemed incomplete. Minimising their occurrence is key to overall system performance. A database containing records for 773,483 individual milking events from four AMS farms was used to describe the risk of an incomplete milking event. Each record contained information about farm, cow, lactation, days in milk (DIM), milking interval (MI), milk yield (MY, in kg/milking) and if the milking event was complete or not. An incomplete gamma function with a random cow effect on the intercept was used to model variability amongst cows within their lactation profile, with adjustment for lactation number (1, 2 and 3 or more) and calving period (warm or cool). The best linear unbiased (BLUP) prediction of cow effect was used to categorise lactations into percentile 33 (P33) and 66 (P66), as either high, medium or low MY level respectively. A proportional hazards model was fitted to describe the risk of a complete milking event happening as a function of MY and calving period. First lactation cows with high MY level were 2.05 times more likely (P<0.0001) to have a complete milking event than those with low MY level. The likelihood was only 1.36 (P<0.0001) for cows with three or more lactations. Furthermore, lactations that had commenced in the cool period were more likely to have shorter MI than those that commenced in the warm period. Know the milking interval in which complete milking happens for different productive groups would improve the performance of the robot.