Feeding through robots

AMES, Iowa — A relatively new piece of technology that aims to reduce feed waste is automatic feeding robots. Earlier this summer, the I-29 Moo University Dairy Webinar Series showcased the potential benefits of automated feeding robots. The seminar was hosted by Dr. Isaac Salfer, who discussed the potential benefits and challenges of implementing automatic feeding systems.

Salfer published an observational study comparing feed bunk nutrient consistency, cow behavior and milk components between herds with automated feeding robots and pair-matched herds with conventional feeding systems.

For the study, Salfer and his team used herds with robotic feeding systems that had been in use for over a year. These robotic feeding systems are battery-powered and self-propelled, making them different than earlier rail-guided systems or conveyor belt distribution systems. 

The robotic feeding systems can feed more than 12 times per day, which studies have shown improves productivity and reduces waste by managing refusals more efficiently. Currently, these systems are more effective for small- to medium-sized herds, ranging from 250-425 cows. 

“Some of the strengths the robotic feeding system provides are reduced labor and improved consistency of feeding,” Salfer said. “We can feed more frequently because we don’t need a person physically loading and driving a TMR (total mixed ration) mixer wagon. Most systems are also electronically powered, so they reduce energy costs substantially compared to a TMR mixer and tractor. The biggest limitations of these systems for larger herds are that they have smaller capacities than traditional TMR mixers and can’t deliver feed as quickly.”

The smaller batch size, as compared to a conventional TMR mixer, is by design. There is less demand for a large capacity mixer because of the automatic feeding robot’s ability to refill and deliver more frequently throughout the day.  

Another drawback to this smaller capacity is that an additional stationary mixer may be needed at times. To accommodate this, the staging area may also need to be adjusted.

With an automatic feeding robot, feed can be added directly, but most of the newer systems use a “feed kitchen” setup that allows automated or semi-automated addition of feed directly into the mixer. Different elevator systems could be needed to deliver feed to the mixer wagon itself. A majority of these systems have a conveyor belt that delivers feed to the mixer with an elevator. For many ingredients, especially silage, there is still a need to deliver feed from the bunk to the staging area. Certain systems also use robotic grabber arms to pick up bulky feeds, such as forages and deliver them to the robot.

“I have been to many farms with automated feeding robots, and I can tell you that not a single feed kitchen setup is the same between farms,” Salfer said.

With this system in place, feed is not required to be delivered as soon as prepared but can be delivered at any time of the day, depending on the setup of the program. Delivery is often based on the amount of feed left within a bunk. 

The theory behind the more frequent feedings is to allow for more productivity and milk butterfat production by stabilizing the pH of the rumen. Delivery of fresh feed is also a stimulus for cows to get up and eat; therefore, the more feedings that occur, the more meal consumption for the cattle and higher activity rates. 

This more frequent consumption over a larger window throughout the day could result in more stable introduction of feed within the rumen, more stable rumen fermentation and a more stable rumen pH. This also helps to control intake and reduce refusals. 

To test this theory, Salfer and his team worked with herds in Iowa, Minnesota and Wisconsin. They used a total of 16 herds, eight with conventional feeding, and eight with an automated feeding robot. The herds were primarily Holstein, with similar ingredients, and matched herds based on those factors. All farms also used a Lely Astronaut for milk harvesting, limiting as many variables as possible.

Feed bunks were sampled four times throughout the day: 5:00 a.m., 11:00 a.m., 5:00 p.m. and 11:00 p.m. Farms were paired across a 3-day window. Sampling times were consistent based on the lack of uniformity among feeding times, depending on the automatic feeding robot and conventional feeding. 

“What we found was that the herds with robots had slightly more consistent nutrient composition throughout the day for dry matter and fiber when compared to conventional herds,” Salfer said. “We didn’t see a difference in particle size distribution or variation in particle size between the two types.”

No differences in milk production were determined between herds with the two different feeding systems. However, this may partially be due to the fact that only a limited number of herds with feeding robots were available for study, and there may not have been enough data to determine a significant result. Salfer said, anecdotally, many of the farms he spoke with did not see major advantages in production after introducing the feeding robots. 

Salfer discussed the primary differences observed within the fatty acid profile. 

“Somewhat to our surprise, we saw a statistically significant increase in de novo fatty acid concentration in our herds that use the automated feeding robot,” he said. “An increase in de novo fatty acid concentrations is typically indicative of a more stable rumen pH and could indicate more consistent feed consumption.”

 While de novo fatty acids are typically indicative of improved rumen health, further research is needed to know for sure if rumen health is improved through robot feeding systems. 

“While we don’t have the data to completely support this, I expect that cows fed with robots are eating more frequently because food is being delivered more frequently,” Salfer said. “That’s stabilizing the rumen pH enough to get that small but statistically significant increase in concentration of de novo fatty acids that we observed. In the future, we would really like to study how feeding behavior changes in cows fed with robots to further understand this mechanism.”

De novo fatty acid concentrations are most commonly used as a way to determine if a cow or a herd is experiencing milk fat depression, as cows with milk fat depression will typically greatly decrease de novo fatty acid concentrations. De novo fatty acids are closely associated with increased milk fat yield and concentrations. If the frequent feeding allowed by robot feeding can increase de novo fatty acids across a herd, it could potentially suggest that cows fed with these systems could be more resilient to milk fat depression or have a greater potential to produce more milk fat. 

Another takeaway Salfer observed was that cows in herds that used feeding robots had an average of 0.54 more visits per day to the milking robot than herds with conventional feeding. Robot feeding herds also had greater milking robot refusals and lower feed allocation in the robot. These results all suggest that feeding robots increase cow traffic and stimulate cows to more easily visit the milking robot without the need to fetch them or increase feed within the robot milker.

“Our data on this is very clear, and I have also talked to several farmers who specifically discussed improvements in cow visits to the robot milker after they have added feeding robots,” Salfer said. “It seems like feeding more frequently gets cows up and moving more to eat, and they decide while they are up, they may as well go visit the milking robot.”