Date of Award

18-6-2025

Document Type

Doctoral Thesis

Degree Name

Doctor of Philosophy

First Advisor

Dr. Michael D. Murphy

Second Advisor

Dr. John Upton

Abstract

The abolishment of European Union milk quotas in 2015 has increased the strain on dairy farm infrastructure and labour availability. For both pasture and confinement-based production  systems, the milking process requires the largest annual labour input. Hence, this area  represents an opportunity for achieving gains in labour efficiency. The aims of this thesis were to 1) document and quantify factors that affect milking process efficiency and 2) develop models that can simulate milking parlour efficiency with respect to these factors. Data were collected from a sample of commercial, pasture-based Irish dairy farms across the Republic of Ireland that used herringbone and rotary parlours. Milking efficiency was evaluated through;  1) cows milked per hour (cows/h), 2) cows milked per operator per hour (cows/h per operator) and 3) litres of milk harvested per hour (L/h). A multi-faceted methodology was used, involving the use of video cameras, infrastructure surveys, and herd milk yield databases. Mechanistic models were developed to simulate milking efficiency through varying parameters of parlour infrastructure, automation specification and management practices. Model results  were then analysed to identify optimisation strategies relative to those parameters.  The first objective of this study involved documenting milking efficiency with respect to the  effects of parlour infrastructure, automation frequency, management practices and seasonality. For parlour infrastructure, on average, rotary parlours achieved higher levels of milking efficiency than herringbones (+81% cows/h, +81% cows/h per operator, and +52% L/h). Conversely, herringbone cluster utilisation values were greater than rotaries (+67% cows/cluster per h, +33% cows/cluster per operator per h, and +65% L/cluster per h). The  author observed a strong positive correlation (0.75, 0.74) between herringbone automation  frequency and milking efficiency (cows/h, L/h). However, the minimal variation in rotary  parlour automation frequency made their effect on milking efficiency indeterminable. For management practices, two-operator milkings for herringbone (H) and rotary (R) parlours xix resulted in increased cows/h (+19% H; +34% R) and L/h (+21% H; +12% R) but decreased cows/h per operator values (-35% H, -14% R). For both parlours, an inverse, seasonal  relationship existed between litres of milk harvested per hour and cow throughput per hour  across lactation. The second objective was to quantify the effects of parlour size, automation  specification, and management practices on the milking and operator efficiency of herringbone and rotary parlours. Increasing parlour size yielded greater gains inmilking process time (MPT, s/cow) for herringbones (-1.3 s/cow) than rotaries (-0.2 s/cow). Automatic cluster removers (ACRs) significantly reduced herringbone total process time (TPT, -13.3 s/cow), cluster time (CT, -18.9 s/cluster), and work routine time (WRT, -32.6 s/cow). Rapid exit (RE) significantly 407 lowered herringbone CT (-18.6 s/cluster). The use of a backing gate (BG) significantly increased rotary WRT (+2.9 s/cow). An additional operator significantly reduced herringbone TPT but not MPT or CT. For rotary parlours, it was found that a second operator had no effect on TPT, MPT, CT or WRT values.The third objective was to develop, validate, and demonstrate the rotary parlour model (RPM). RPM was developed using sample farm data. Validation was achieved by comparing simulated  results to empirical data using validation metrics. RPM achieved mean absolute percentage error (MAPE) values of 3.5% and 2.9% for cows/h and L/h, respectively, and 10.7% for platform times (s/rotation), indicating sufficient prediction accuracy. The author found the effect of parlour size is primarily observed at low rotation times (s/bail) (i.e., at 10 s/bail, increasing parlour size from 40 to 60 clusters increased cows/h by 44%). Further reducing rotation time from 15 to 10 s/bail was less effective at increasing efficiency for lower sized parlours (40 clusters, +3% cows/h) than larger parlours (60 clusters, +26% cows/h). Increasing the ACR threshold primarily enhanced the efficiency (+18% cows/h) of lower sized parlours (40 clusters) at low rotation times (12 s/bail). The optimal go-around cow occurrence at milking ranged from 2-20%, depending on herd size, parlour size, rotation and time and ACR threshold xx at milking. The fourth objective was to develop, validate, and demonstrate the herringbone parlour model (HPM). Data from the herringbone farms sample were used to develop HPM. The validation of HPM was achieved by comparing simulated results with empirical data using verification metrics. HPM achieved MAPE values of 9.6% and 8.4% for cows/h and milking  process time (MPT, s/cow), respectively, and 7.9% for row times. Increasing parlour size (16 to 20 clusters, 20 to 24 clusters) increased cows/h (+51%, +24%) but decreased operator idle time (-49%, -39%). Using ACRs and rapid exit together had a greater effect on increasing  milking efficiency for smaller parlours (16 clusters, +14% cows/h) than larger parlours (24 431 clusters, +8% cows/h). Conversely, using ACRs and rapid exit together had a greater effect on increasing operator idle time for larger parlours (24 clusters, +240%) than smaller parlours (16 clusters, +65%). Increasing the ACR threshold was found to have a variable effect on milking efficiency across parlours, increasing cows/h by 9%, 1% and 7% for 16, 20 and 24 cluster parlours respectively. The main outputs of this study were: (1) a quantifiable understanding of the effects of parlour size, automation specification, and management practices on herringbone and rotary parlour milking efficiency; (2) identified strategies towards the optimisation of herringbone and rotary parlour milking efficiency; (3) data-driven models that can be used to explore a variety of different milking system scenarios relative to various configurations; and (4) an accurate and  holistic methodology for comprehensive evaluation of herringbone and rotary parlour efficiency. The outputs of this thesis will inform dairy farmers, operating in either pasture or confinement-based production systems, in the optimisation of operator performance and parlour throughput at milking. This knowledge will help to reduce annual labour requirements associated with the milking process on dairy farms and enhance the vitality of the international  dairy industry.

Access Level

info:eu-repo/semantics/openAccess

Download

Share

COinS