Milking efficiency of swingover herringbone parlours in pasture-based dairy systems

Identifying strategies to enhance the milking and operator efficiency of herringbone and rotary parlor systems in Ireland

International trends of increasing dairy herd sizes coupled with scarcity of labor has necessitated the enhancement of labor efficiency for dairy production systems. This study quantified the effects of infrastructure, automation, and management practices on the milking and operator efficiency of herringbone and rotary parlors used on pasture-based farms in Ireland. Data from 592 milkings across 26 farms (16 herringbones and 10 rotaries) was used. The metrics of cows milked per hour (cows/h), cows milked per operator per hour (cows/h per operator) and liters of milk harvested per hour (L/h) described milking efficiency. The metrics of total process time per cow (TPT, s/cow), milk process time per cow (MPT, s/cow), work routine time (WRT, s/cow), cluster time (CT, s/cluster), and attachment time per cow (ATC, s/cow) described operator efficiency. Automations investigated were backing gates, cluster flush, plant wash, cluster removers (ACRs), feeders, entry gates, rapid-exit, and teat spray. Additional operator presence at milking was also investigated. Herringbone and rotary parlors were assigned to quartiles from their cows/h per operator values to examine infrastructure, automations, and management practices variations. Fourth quartile herringbones based on cows/h per operator values (Q4) averaged 93 cows/h per operator using average system sizes of 24 clusters with 5 parlor automations. Q4 rotaries averaged 164 cows/h per operator using average system sizes of 47 clusters and an average CT of 13 s/cluster. Cows/h per operator values for Q4 herringbone and rotary parlors were 82% and 54% higher, respectively, than values observed on Q1 parlors, indicating the considerable potential to improve efficiency. To determine if infrastructure, automations, or additional operators at milking significantly affected operator efficiencies, general linear mixed models were developed. For parlor infrastructure, additional clusters had greater significance on operator efficiencies (MPT) for herringbones (-1.3 s/cow) as opposed to rotaries (-0.2 s/cow). Hence, increases in system size was likely to result in improved efficiencies for herringbones but less so for rotaries. For automations, ACRs significantly reduced herringbone TPT, CT, and WRT values by 13.3 s/cow, 18.9 s/cluster, and 32.6 s/cow, respectively, whereas rapid-exit significantly lowered CT by 18.6 s/cluster. We found no significant effect on rotary TPT, MPT, CT, or WRT values from the use of automatic teat sprayers. An additional operator at milking was found to significantly reduce herringbone TPT but not MPT or CT. For rotaries, a second operator had no significant effect on TPT, MPT, CT, or WRT values. We documented strong negative correlations between operator efficiencies (TPT, MPT) and milking efficiency (cows/h) for both herringbone (-0.91, -0.84) and rotaries (-0.98, -0.89). Strong negative correlations between the herringbone automation count and TPT (-0.80), MPT (-0.72), and CT (-0.75) suggested highly automated parlors were likely to achieve greater operator efficiencies than less automated parlors. The strong negative correlation (-0.81) between rotary milking efficiency (cows/h) and CT suggested lower CT values (i.e., rotation speed) resulted in increased milking efficiency. In conclusion, our study quantified the effects of parlor infrastructure, automation, and management practices on the milking and operator efficiency of herringbone and rotary parlors.

The effects of seasonality, management, infrastructure, and automation on the milking efficiency of herringbone and rotary milking parlors in Ireland

The objective of this study was to document the milking efficiency of a sample of Irish dairy farms and to understand the effects of (1) seasonality, (2) management practices, (3) parlor infrastructure, and (4) parlor automations on milking efficiency metrics. A novel methodology based on empirical data from video cameras, infrastructure surveys, and milk yield data allowed for the accurate computation of milking efficiency metrics and quantification of the effects of seasonality, number of operators, and parlor automations on milking efficiency across 2 parlor types. The data for this study were collected over 2 periods: period 1 (July 28, 2020, to October 23, 2020, peak-late production) and period 2 (April 12, 2021, to May 19, 2021, early-peak production) from a sample of 16 herringbone and 10 rotary commercial Irish dairy farms. Milking efficiency was evaluated on each farm using 3 key performance indicators: (1) cows milked per hour (cows/h), (2) cows milked per operator per hour (cows/h per operator), and (3) liters of milk harvested per hour (L/h). Milking efficiency key performance indicators were calculated using "total process time," defined as the time between the first cow entering the holding yard and the end of the cleaning process. Average herd sizes for herringbone and rotary farms were 180 and 425 cows, respectively. Average system sizes for herringbone and rotary farms were 20 and 50 clusters, respectively. For herringbone farms, the average milking efficiency was 94 cows/h, 73 cows/h per operator, and 1,012 L/h, whereas rotary farms achieved an average milking efficiency of 170 cows/h, 132 cows/h per operator, and 1,534 L/h. Parlor size was strongly correlated with milking efficiency (cows/h) for herringbone parlors (0.91) but was only moderately correlated for rotary parlors (0.50). Hence, we documented the effect of parlor size on milking efficiency is relative to parlor type. Cluster utilization values on herringbone farms were 5 cows/cluster per h, 4 cows/cluster per operator per h, and 51 L/cluster per h, which were 67%, 33%, and 65% greater than rotary farms, respectively. We found for both herringbone and rotary farms hourly cow throughput (cows/h, cows/h per operator) were greatest during period 1 and that the volume of milk harvested per hour (L/h) was greatest for period 2. Thus, we documented an inverse seasonal relationship between hourly rates of cows milked and milk harvested. We observed that for herringbone farms, milking efficiency (cows/h, L/h) had a strong positive correlation (0.75, 0.74) with the levels of automation use. However, the minimal variation in automations used among rotary farms made it difficult to evaluate their effect on milking efficiency. Similarly, we found that the effect of automations on milking efficiency was dependent on parlor type. On average, a second operator at milking for both herringbone (H) and rotary (R) farms increased values for cows/h (+19%, H; +34%, R) and L/h (+21%, H; +12%, R) but lowered values for cows/h per operator (-35%, H; -12%, R). The holistic methodology applied in this study allowed us to add novel data to the literature by quantifying the effects of seasonality, the number of operators present at milking, and parlor automation use on milking efficiency across 2 parlor types.

Improving parlor efficiency in block calving pasture-based dairy systems through the application of a fixed milking time determined by daily milk yield and milking frequency

Adjusting end-of-milking criteria, in particular applying a maximum milking time determined by expected milk yield at an individual milking session, is one strategy to optimize parlor efficiency. However, this strategy can be difficult to apply practically on farm due to large differences in session milk yield, driven by milking interval, which affects milking routines and can be limited by in-parlor technology. The objective of this study was to test the hypothesis that a single fixed milking time (duration) could be applied at all milking sessions without compromising milk production or udder health for a range of milking intervals. To test the hypothesis, 4 experimental herds were established: (1) herd milked twice a day (TAD) using a 10- and 14-h interval, (2) herd milked TAD using an 8- and 16-h interval, (3) herd milked 3 times in 2 d using a 10-19-19-h interval, and (4) herd milked once a day (OAD). Herds consisted of 40 cows each, and were established for two 6-wk experimental periods, one in peak lactation and the other in mid-late lactation. Within each herd, half the cows had an end-of-milking criterion of 0.35 kg/min (Flow), and the other half had milking ended after a fixed period of time (FixedT) based on the average milking session yield, the daily milk yield divided by average number of milkings per day, irrespective of milking interval. We found no differences in daily milk yield between end-of-milking criteria due to residual milk from one milking likely increasing the proportion of milk in the udder cistern at the next milking session for the FixedT treatment. However, fat yield was compromised when the percentage of the herd with a truncated milking exceeded an estimated 33% at a milking session, which occurred in the TAD 8-16 herd due to the divergence from the average milking interval (in the case of TAD, 12-12 h). Applying a fixed milking time had no detrimental effects on udder health, except in the OAD herd in mid-late lactation, which had both a higher cell count and new intramammary infection rate. This warrants further investigation, although the majority of cultured bacteria were coagulase-negative staphylococci (CNS). Consequently, we conclude that, in general, with appropriate monitoring (e.g., weekly inspection) to ensure the proportion of the herd with truncated milkings does not exceed 33%, farmers in pasture-based dairy systems can use a fixed milking time to improve parlor efficiency.

Evaluating rates of technology adoption and milking practices on New Zealand dairy farms

This study assessed technology use and evaluated rates of technology adoption and milking practices on New Zealand dairy farms. Industry surveys were conducted in 2008 and 2013, when farmers were asked a series of questions relating to their physical farm details, their role in the business, their attitudes towards technology, the technologies they had on-farm and their levels of satisfaction. In total, 532 and 500 respondents were questioned in the two surveys, respectively, with a similar representation of rotary and herringbone dairies. Questions relating to attitudes towards new technologies were subjected to a cluster analysis using the 2013 dataset. Farmers were classified into two categories, ‘fast’ and ‘slow’ adopters. Fast adopters are more likely to have a rotary, with a larger farm and more cows. The most common technology in herringbone dairies is automatic vat washing and in rotary dairies automatic cluster removers (ACR). Rotary dairies equipped with ACR, automatic drafting and automatic teat spraying achieve greater labour utilisation (cows/labour unit). Around half of farmers with herringbone dairies sometimes or always wait for slow-milking cows to milk out and 85% of farmers do not know the their ACR settings, highlighting significant potential to improve milking efficiency. Overall, technology is associated with greater labour utilisation. However, the benefits of each technology should be scrutinised to ensure appropriate investment decisions are made by farmers.