Back to search

FFL-JA-Forskningsmidlene for jordbruk og matindustri

Precision feeding of sows. Feed for need.

Alternative title: Presisjonsfôring av purker

Awarded: NOK 3.6 mill.

Project Number:

269074

Project Period:

2017 - 2022

Partner countries:

The main goal of the precision feeding of sows (PFOS) project has been to use new and available technology for large scale data collection of feed intake, weight, back fat, and other production traits in sows. In this project we have worked on automating and digitalizing the collection of feed data from two different feeding systems and linking this data to new and existing data platforms. Thus, it is now possible to collect a more continuous large scale data flow of feed intake, weight, backfat and body condition throughout the sows’ different productions cycles. This is data that can be used to improve feed composition, feed advice and herd management. During this project we have made a data platform for receiving and sending of feed data from two different feeding systems. The main job has been to automate the sending and receiving of such data by developing different API’s (application programming interface), structure and handle the data collected, and make a cloud solution for storage of the data. We have developed similar systems for water consumption data. Feed- and water data can now be transferred to ingris and be linked to relevant production results. For now, this system is operational for our two test herds. The aim is that more farms can use the system, but to do so contracts for the regulation of data rights and usage must be in place. This project has also resulted in a new feed module in ingris which can receive feed and water data on pen or valve level. The feed module also allows for automated registration of feed invoices from Landbrukets Dataflyt, or the producers can choose to manually add feed status and feed purchase to get a simple overview over their productions feed costs. Another feature of the module is that if the producers add an overview of their different production units into ingris, they can track all transfers of their animals throughout their systems. Individual weight, back fat and body condition can now also be registered in ingris, and a production report showing weight and back fat development over time is available. As part of the project, we have performed two feed trials where feed and water data were collected automatically through the systems we have set up. This meant that less manual labor was needed to perform the trials, and it was easier to run the trials for a longer period. In the first trial the main goal was to compare the effect of using a specially formulated transition feed during late lactation and until farrowing had occurred, with a control group that went straight from a gestation feed to a lactation feed. The results showed that using a transition feed gave a significantly higher average daily gain (+ 300g per day per litter) and higher average weaning weights (+ 16kg per litter). Although not significant, using a transition feed also showed a numerically lower number of stillborn piglets and higher birth weights. In this trial we also looked at the effect of water usage on production performance. We found that gilts use around 32 liters of water per day during the last two weeks of gestation, while sows use on average 37 liters per day. This is higher than the current recommendations of 12-15 liters per day. The water usage increase towards time of parturition, then it drops for about 24 hours before it increases rapidly again in the days after farrowing. During lactation the water usage was on average 38 and 43 liter per day for gilts and sows respectively. We also found a moderate correlation between water and feed intake during lactation. In the second trial we compared a gestation feed with a higher energy and protein for Landrace sows, to an ordinary gestation feed. The hypothesis was that our pure breed animals have a higher energy and protein need during gestation because they are leaner then crossbred animals. The trial only resulted in small numerical differences in weight and back fat gain during gestation between the test groups, and it was concluded that from an economic standpoint the extra feed costs was higher than the added production value. This project has shown that it is doable to set up systems for automated transfer of feed and water data from available systems in Norway, and to transfer this information automatically to ingris. This will make it less time consuming to perform feed trials, and it gives the feed producers a better overview of their feed related production costs. Such data are also valuable for improving diets, feed advice and farm management. It is a big advantage that such data can be connected to other production data in ingris, so that the farmers must relate to only one digital platform to get a complete overview over their production. However, this project has also uncovered a lot of issues with regards to data rights and data usage. Big data is big business, and proper contracts must be in place to ensure both suppliers and buyers the necessary data access.

I prosjektet har vi hatt søkelys på å innhente fôrdata fra to ulike fôrsystemer (Nedap og Big Dutchman) og et vannmålingsanlegg fra Skov. Det var viktig at anleggene vi valgte hadde et systemoppsett som muliggjorde enkel overføring av data gjennom en skyløsning. Vi har fått på plass avtaler om utveksling av slik data for tre ulike besetninger i prosjektperioden, og vi har bygd opp en struktur for mottak, overføring og lagring av slik data hos Norsvin. Dette er erfaring og kompetanse som vi kan nyttiggjøre oss av i andre prosjekter. Animalia har som vår samarbeidspartner i prosjektet utviklet en ny fôrmodul i ingris. Denne fôrmodulen kan hente data fra databasen som er bygd opp hos Norsvin. Dataene kan hentes ut helt ned på fôrventilnivå, og er satt i produksjon for våre testbesetninger. Vi vet at flere besetninger er interessert i å få koblet seg opp til dette systemet, dermed vil det i etterkant av prosjektet jobbes videre med å få på plass dataavtaler med leverandører av ulike fôringsanlegg for at dette skal bli mulig. Den nye fôrmodulen har også flere funksjoner som allerede er tilgjengelig for ingris brukere, og vi forventer at når flere resultater blir kommunisert ut vil det gjøre at flere benytter seg av funksjonaliteten i modulen. Felleskjøpet fôrutvikling har som vår samarbeidspartner i dette prosjektet gjennomført to fôringsforsøk i to ulike besetninger og med gjentak over flere innsett. Fordi slike forsøk ofte krever mye manuell registrering av daglig fôropptak, samt andre registreringer, blir de ofte dyre, kortvarige og involverer få dyr. I dette prosjektet ble fôringsforsøkene gjennomført ved at data på vann- og fôrforbruk ble hentet ut daglig ved automatisk overføring. Det betydde at det ble mindre jobb både for både produsent, og for fôrfirma, og forsøkene kunne gjentas over flere innsett. Får vi flere produsenter knyttet opp til systemene vi har utviklet vil det også bety at man kan kjøre forsøk hos flere produsenter, over lengre tid og som er mindre ressurskrevende enn tidligere. Dette gjør at man kan gi bedre rådgivning og holde bedre takt med den genetiske utviklingen i produksjon som vi ser hos besetningene. En viktig effekt av dette prosjektet er at vi har satt søkelys på en del utfordringer rundt data- og bruksrettigheter. Bondens data er dessverre ikke alltid lenger bondens data. Når en produsent investerer i et avansert fôringsanlegg sender dette anlegget fra seg langt mer data enn det produsenten har mulighet til å hente ut selv. Det vi har jobbet mye med i dette prosjektet er nettopp å få tilgang til disse dataene, overføre de til ingris og knytte de opp mot andre produksjonsresultater tilgjengelig i ingris. All data og kunnskap ervervet gjennom dette prosjektet brukes blant annet til forbedringer av ingris, rådgivning og oppdatering av fôringsanbefalinger både nasjonalt og internasjonalt (gjennom Topigs Norsvin).

The main objective of this project is to use current and new technology to develop methods for collecting, automating, handling and analysing large-scale feed and performance data from litter producing sows, and use this data to develop more precise feed and management advise, and to improve our breeding goal. We attend to achieve this by investigating the different technology available for cost efficient data collection within a commercial setting, and establish agreements with farmer for installation of feeders, videos and other equipment needed to allow for a more automated process of data collection. The next step will be to develop a platform for data management and handling of the data collected and link this with already existing software, before we test the results by running a phase-feeding trial aimed at getting more detailed information regarding feed intake, body development and performance data from our litter producing sows (both Landrace and TN70). The next step includes developing and/or optimizing statistical tools for handling of such big data and link the data with already existing pedigree data and other relevant information from Norsvin database. Finally, we want to implement the results by developing new feed manuals and feed advise for the national and international market, allow for more detailed registering in the national litter recording system (Ingris), and including new traits in the breeding goal to increase factors related to feed efficiency in our Norwegian maternal breeds.

Funding scheme:

FFL-JA-Forskningsmidlene for jordbruk og matindustri