Automatic milking

The general conditions for milk production are defined by the Commission Directive 89/362/EEC (1989). Fulfilment of this directive is presently a problem with automatic milking systems because visual inspection of foremilk is not performed. Technical solutions may replace visual inspection for detection of abnormal milk either before or during milking, and consecutive separation. However, unequivocally and generally accepted definitions of acceptable milk quality and abnormal milk are not available. In order for AMS companies to develop sensors to detect abnormal milk, a precise definition of abnormal or unacceptable milk is needed. Such a definition has consequences not only for abnormal milk but also for the amount of consumable milk. Therefore, the objectives of this workpackage are:

• Definition of acceptable milk quality of raw milk at time of milking 
• Testing the state-of-the-art of automatic milking systems for their ability to automatically detect and separate abnormal milk.

The activities are divided into six parts: 
Part 1-3 (subpackage 1): Data analysis as preparation for a workshop on acceptable milk quality. 
Part 4 (subpackage 2): Organisation of the workshop to set definitions for acceptable milk quality. 
Part 5-6 (subpackage 3): Analysis and report writing of the consequences for milk quality given the definition of abnormal milk.

1. Milk quality on the cow level. 

A dataset with 6000 visual inspections and CMT-score of foremilk of 1500 cow milkings is available from three herds with more than 100 cows each and a prevalence of about 5 % clinically and 20 % subclinically infected cows. Besides visual inspection of foremilk, the dataset includes fat, protein, lactose, urea, freezing point, and cell count of composite milk. Working title for a paper: Visual inspection and CMT-scoring of foremilk of individual quarters as tool to predict cell count of composite milk.

2. Milk quality on quarter basis 

The automatic milking system on our experimental farm will be equipped with semi-automatic sampling devices for quarter milk. Data will be collected of at least 50 cases of clinical mastitis and 300 cases of subclinical mastitis. This requires sampling of at least 2000 quarters. Visual inspection and CMT-score will be performed on foremilk. Quarter milk samples will be analysed for cell count. Conductivity and milk flow, milk yield, and temperature will be collected by the automatic system. Working title for a paper: The relationship between visual inspection and CMT-scoring of foremilk and milk composition of individual quarters as tool to predict cell count of composite milk.

3. Time of decision of acceptable milk quality during the milking process 

The experimental farm will report clinical mastitis cases and individual cows will be monitored and sampled for about 10 milkings hereafter. The sampling device for part 2 will be modified and used to take individual time related samples during milking. Samples will be taken of foremilk, at least at 3 different time periods during milking, and of strip yields, and analysed for fat, protein, lactose, and cell count. Foremilk will be scored visually and by CMT and data from the automatic milking system will be collected as in part 2. Historical data from individual quarters will be available as well. The herd will be monitored for about one year in order to get data from about fifty cows with clinical mastitis. Working title for a paper: Changes in milk composition during milking in relation to health status determined by visual inspection and CMT-scoring of foremilk.

Statistical analysis 

Visual appearance and CMT-score of foremilk are each individually and complementary regarded as "golden standards" for the test of different separation criteria. Statistical analysis will be performed of single samples and of consecutive samples. The test results will be presented as sensitivity, specificity, and positive and negative predictive values.

4. Workshop 

Results will be presented at a workshop for invited scientists and EU legislators (DG-Agri) with the goal of setting definitions of acceptable milk quality. The workshop will be held at the end of the second year of the project. Working title of a paper: Consequences for milk quality and amount of saleable milk by different definitions of abnormal milk.

5. Evaluation and demonstration phase 

Consequences of our definitions will be demonstrated and evaluated within a data set available from the official test of automatic milking systems on the Danish market and by statistical analysis of data collected in WP 4. Working title of a paper: The ability of automatic milking systems to detect and separate abnormal milk.

6. Final report to EU legislators 

A final report will be forwarded to EU legislators with the proposal for definition of acceptable milk quality.

• D6: Report describing consequences of definitions of acceptable milk quality for the practical use of automatic milking systems. 
• D7: Report on the possibilities of automatic milking systems to separate milk based on quality.

© Animal Sciences Group - Wageningen UR. Last update: 20-02-2008 10:03. Mail any questions and remarks to: webmaster.asg@wur.nl