Control charts to monitor growth in dairy heifers

VISSIO, CLAUDINA; MANCILLA, BRENDA; TURIELLO, PAULA; LARRIESTRA, ALEJANDRO

Congreso; Virtual Annual Meeting; 2020

There is a need to monitor processes in dairy systems, and there are methodsof statistical quality control, like control charts, to effectively do it.Although widely applied in manufacturing industry, control charts havenot been much used in animal growth processes. The objective of thisstudy was to build a control chart application as a management tool in acommercial dairy for monitoring and surveillance of the growing processof replacement heifers. Control charts were built from BW data of 2959-mo old heifers born during the years 2016 to 2018 in a commercialdairy farm in Córdoba province. During the first phase we built the controlcharts for X-bar and R from 30 samples of 10 heifers each, following theinstructions of Montgomery (2005). The center line of the X-bar chartwas the average of the sample means (260 kg), and the upper and lowercontrol limits (UCL and LCL, respectively) were calculated with theaverage of the sample means, the mean range and a constant A2 (UCL =X-bar-bar + A2 x R-bar; LCL = X-bar-bar - A2 x R-bar). For the R chart,the center line was the average of the sample ranges (70 kg), and the UCLand LCL were the average of the sample ranges multiplied by a differentconstant for each (UCL = D4 x R-bar; LCL = D3 x R-bar). The constantsA2, D4 and D3 are tabulated values for different sample sizes. Controlcharts were built and 5 samples out of control were removed after findingassignable causes. From the final graphs the dairy manager can addfuture sample results and determine that the process is under control if itsaverage is between 238 and 281 kg and the range is between 16 and 124kg. If the contrary happens the manager can suspect that an assignablecause has occurred, and some investigation should be made and correctiveaction should be taken to remove the unusual source of variability.These control charts could be used to monitor the growing process of theheifers in the evaluated farm, to detect when some changes are requiredto take the process back to an in-control state thus reducing the processvariability, and to estimate heifer weight among other process variables.