Comparison of functional aspects in two automatic milking systems and auto-tandem milking parlors.

Milk yield, milking frequency, intermilking interval, teat-cup attachment success rate, and length of the milking procedure are important functional aspects of automatic milking systems (AMS). In this study, these variables were compared for 2 different models of AMS (AMS-1, with free cow traffic, and AMS-2, with selectively guided cow traffic) and auto-tandem milking parlors (ATM) on 4 farms each. Data on milking-stall visits and milkings of 20 cows were recorded on 3 successive days by means of video observations. Data were evaluated with mixed-effects models. Milk yield did not differ among the 3 milking systems. Milking frequency in the AMS was 2.47/d [95% confidence interval (CI) = (2.38, 2.56)], and was significantly higher than the 2 milkings/d in ATM. Milking frequency was lower for cows with a higher number of days in milk (DIM) in AMS-1 [change of -0.057/10 DIM, CI = (-0.070, -0.044)], but remained constant for cows with varying DIM in AMS-2 [change of -0.003/10 DIM, CI = (-0.034, 0.027)]. As a consequence, milking frequency was higher in early lactation [by 0.603, CI = (0.102, 1.103)] and lower in late lactation in AMS-1 than in AMS-2 [by -0.397, CI = (-0.785, -0.008)]. The intermilking interval showed the opposite pattern. Teat-cup attachment was more successful in AMS-1 than in AMS-2 (98.4 vs. 94.3% of the milkings), with some variation among farms (range: AMS-1 96.2 to 99.5%; AMS-2 91.5 to 96.1%). The length of the entire milking process did not differ among the milking systems [454 s, CI = (430, 478)], although the preparation phase was longer [changes in comparison with ATM: in AMS-1 by a factor of 2.90, CI = (2.30, 3.65), and in AMS-2 by 5.15, CI = (4.09, 6.48)] and the actual milking phase was shorter in both AMS-1 and AMS-2 than in ATM [changes in comparison with ATM: in AMS-1 by a factor of 0.76, CI = (0.62, 0.94), and in AMS-2 by 0.75, CI = (0.60, 0.93)]. The admission [changes in comparison with ATM: in AMS-1 by a factor of 2.56, CI = (1.55, 4.22), and in AMS-2 by 3.07, CI = (1.86, 5.08)] and preparation phases lasted longer in AMS-2 than in AMS-1, whereas the time required by the cows to leave the milking stall did not differ among the systems [changes in comparison with ATM: in AMS-1 by a factor of 0.89, CI = (0.55, 1.44), and in AMS-2 by 1.02, CI = (0.63, 1.66)]. In conclusion, different technical approaches to automatic milking led to differences in teat-cup attachment success rates, in the duration of several phases of the milking process, and in milking frequency. The capacity of an AMS could be further improved by shortening the preparation phase and reducing the proportion of failed milkings.