Validation of Individual Computerized Sow Feeding Systems in Lactation

Two experiments evaluated the accuracy of individual computerized feed delivery systems for lactating sows (GESTAL Solo, JYGA Technologies, Inc., Quebec City, Canada). The feeders volumetrically dispense feed based on rotations of a screw auger. In Exp. 1, 29 prototype feeders were used across 3 farrowing groups. On d 0, four feeders were selected to calibrate the computer system to the bulk density of the lactation diet. Feeders were programmed for 5 feeding periods per day with feeding period allowing 2 to 4 feed drops (depending on time of day) triggered by the sow at a minimum of 15 min intervals. Sows activate a trigger within the feed bowl to receive a targeted amount of feed (680 g), and the computerized feeder records the delivery amount based on calibration values. Total lactation feed intake was recorded by weighing the quantity of feed provided to the feeding system for each sow throughout lactation. Feed delivered by a single trigger activation on d 0, d 10 and d of weaning was collected and weighed with a scale and compared to the computer record. Also, total feed delivered over the lactation period was compared between the recorded computer measurement and scale weight. Average percentage difference between the two measurements ranged from 0.01 to 36.6% (P<0.001, SEM 3.0) for a single trigger event. Computer recorded total lactation feed was marginally less (P<0.089) than actual weight of feed delivered (102.8 vs. 107.1 kg; SEM 1.8). Individual feeders had recorded total feed delivery ranging from 77% to 122% of actual weight delivered. Based on the variation observed, a new feeder design was evaluated (plastic hopper manufacturing was injection molded vs. rotational molded). In Exp. 2, 29 feeders were used in a single farrowing group to evaluate the variation of the new feeders. Feeders were calibrated and data was collected using the same procedures as Exp. 1, except individual feed drops were collected 8 times per feeder throughout lactation. Average percentage difference across all feeders ranged from 3.8 to 13.4% (P<0.001, SEM 1.5). There was no evidence (P<0.542) of difference between computer recorded total lactation feed and actual weight of feed delivered (124.8 vs. 121.8 kg; SEM 1.8). Individual feeders had recorded total feed delivery ranging from 90.4% to 106.4% of actual weight delivered. Overall, this study shows the new model was less variable in feed drops and total feed delivery.