Real-time monitoring of growing pigs Thomas Nejsum Madsen IQinAbox - PowerPoint PPT Presentation

Real-time monitoring of growing pigs Thomas Nejsum Madsen IQinAbox – www.iqinabox.com

IQinAbox IQinAbox • Bridges the gap between scientific research and modern pig production • Work with universities, research organizations and suppliers to the pig industry • Our aim is to increase productivity and animal welfare by better use of data • We provide sensors and software for monitoring pigs 2

IQinAbox Software in pig production Management Equipment Monitoring • Monitoring – general trends in industrial production • Experience from Industry 4.0 and Machine learning? • Challenges with biological data • Monitoring examples • Merging several data sources 3

IQinAbox Internet of Things – new opportunities 4

IQinAbox Industry 4.0 – the 4th industrial revolution 5

IQinAbox Machine learning in industrial production ”Machine learning: the science of making computers make decisions without being explicitly programmed to perform the task” Data that’s not used 6

IQinAbox Example: Vibrations on an industrial machine 7

IQinAbox Production unit - weaners 8

IQinAbox Water flowmeter 9

IQinAbox Water consumption per hour (7 days) 5 – 6 p.m. 12 a.m 4-5 a.m. 10

IQinAbox A change in drinking behaviour Change in behaviour 50 Liter vand pr time 40 30 20 10 0 7 8 9 10 11 Dag Day 11

IQinAbox Forecast vs. observation Change in behaviour 50 Liter vand pr time 40 30 20 10 0 7 8 9 10 11 Day Dag Observeret Model 12

IQinAbox Software for monitoring water and feed consumption - challenges • Pigs change drinking behavior as they grow • Drinking patterns vary between herds / housing systems • Research and modelling is based on data from very few test herds • Expensive equipment for data collection 13

IQinAbox Dynamic estimation of daily gain • Development project with Danish Crown • Launched in DK september 2019 Check it out… 14

IQinAbox Example – batch production Delivery date 15

IQinAbox Example of deviation weekly 16

IQinAbox Daily gain estimates converted to growth curves Age (days) 17

IQinAbox Examples with different deviation in growth rate(30-110 kg) • Feeding strategy • Delivery strategy • ‘Looser pigs’ 18 Age (days) Age (days)

IQinAbox Growth estimation based on feed intake • Investigation indicates that growth rate can be estimated based on feed intake • Need to know feed composition and continues feed consumption Growth curve 19

IQinAbox New joint development project with researchers • We use technologies from the production industry in combination with results from Herd Management research (e.g. PigIT) • IoT based sensors • Cloud based data analysis and Machine Learning 20

IQinAbox Modelling The concept Researchers Implement Data IoT hub Output Cloud IoT box 21

IQinAbox IoT – Farm monitoring, OEM IoT hub Output Cloud Feeding and ventilation equipment Feeding and ventilation User interface 22

IQinAbox Central alarm facilitation Centralized surveillance 23

IQinAbox Modular IoT-box 24

IQinAbox Trial farms • Establishment of 5 trial farms • Test and experimental work with various sensor types • Logbook on diseases and disorders • Playground for new ideas 25

IQinAbox Measuring dry feed in silos 26

IQinAbox Complex and difficult to mount load cells on existing silos 27

IQinAbox Strain gauge sensors • Sensors mounted on silo legs • Measure compression on silo legs • Converts signal to weight estimate 28

IQinAbox Load cells vs. strain gauge sensors Load cell data Strain gauge data 29

IQinAbox Growth estimation based on feed intake • Studies indicate that growth rate can be estimated based on feed intake. • Need to know the feed composition and the continuous feed consumption Growth curve 30

IQinAbox 31

IQinAbox Directional sound recognition Pen Cough 32

IQinAbox IQinAbox Modelling - cloud based data science Researchers Implement Data Output Cloud 33