Development, validation, and demonstration of the rotary parlor model

Rotary milking systems have potential to reduce the labor requirement of the milking process on dairy farms. However, there is a need to identify the most effective strategies that lead to consistently efficient milkings. We developed a mechanistic, process-oriented model that accurately simulates the milking efficiency of rotary parlors operating under a diverse range of conditions. The rotary parlor model (RPM) was developed using milking efficiency data collected from a sample of commercial Irish dairy farms (n = 10) and the Teagasc Moorepark Dairy Research Farm (Teagasc, Ireland). To validate the performance of RPM, simulated milking process times (MPT, s) and efficiency metrics (cows milked per hour [cows/h], liters of milk harvest per hour [L/h], MPT per cow [s/cow]) were compared with empirical data recorded on the Moorepark Dairy Research Farm across 2 recording periods. Model validation produced mean absolute percentage error values of 3.3%, 3.5%, and 2.9% for MPT, cows/h, and L/h metrics, respectively. We defined rotation time as the time taken for the rotary platform to pass 1 bail position (s/bail); this value decreases as the platform rotates faster. The functionality of RPM was demonstrated by simulating the milking efficiency (cows/h) of parlor sizes ranging from 40 to 120 clusters, operating at rotation times of 5 to 25 s/bail. We found that the effect of decreasing rotation time (≤15 s/bail) on milking efficiency was greater for larger parlors (50, 60 clusters) than lower-sized parlors (40 clusters). For example, decreasing rotation time from 15 to 10 s/bail increased milking efficiency by 3% for a 40-cluster parlor, compared with 26% for a 60-cluster parlor. As rotation time decreased for all parlor sizes, there was an increased occurrence of go-around cows at milking (i.e., cows with a milking time longer than the platform time, thereby requiring additional rotations). A sensitivity analysis investigated the effect of automatic cluster remover (ACR) threshold change from 0.2 to 0.8 kg/min on the milking efficiency of 40-, 50-, and 60-cluster parlors operating at rotation times of 6, 8, 10, 12, 16, and 20 s/bail for a 300- and 600-cow herd. Using increased ACR thresholds reduced the milking time duration of individual cows. This lowered the number of go-around cows, and thereby rotations required at milking, as a result, milking efficiency increased. The ACR threshold increase effect was largest among lower-sized parlors with lower rotation times. For example, increasing the ACR threshold from 0.2 to 0.8 kg/min for a 40-cluster parlor with a rotation time of 12 s/bail lowered MPT by 17%. In contrast, for a 60-cluster parlor operating in the same conditions, the increased ACR threshold reduced MPT by only 5%. Optimal go-around cow occurrence ranged between 2% to 20%, depending on herd size, parlor size, rotation time and ACR threshold. Through the development and application of RPM, this study provides greater understanding into the effects of parlor size, rotation time, rotation numbers, herd milking characteristics, ACR thresholds, and go-around cows on rotary milking efficiency.