PigSys - Improving pig system performance through a whole system approach
Aim of the project
PigSys has adopted a multidisciplinary, system level approach to pig production systems with the purpose of generating a deeper understanding of the interactions between the animals, technology and environmental conditions in pig rearing. A model of mass and energy flows and Decision Support system, sensor networks, soft sensors (sensor data fusion) as well as novel building climate control systems, have been developed to underpin sustainable improvement in system performance and increase animal welfare. Furthermore, in depth data analysis of the impact of environmental conditions on animal welfare and performance was conducted and a Life Cycle Analysis conducted
Introduction
There are many approaches for the improvement of individual aspects the system animal-technology-environment. To date, none provide effective whole system considerations for pig production. PigSys addresses these issues by adopting a multi-disciplinary, system level approach.
PigSys will provide the sector with (a) a whole system model of energy and mass flows and decision support system (DSS); (b) sensor systems and control devices for improved barn climate control and animal welfare; (c) ‘big data’ to support barn and control system design; (d) soft sensors (sensor data fusion based on animal and environment related data); (e) LCA and LCCA; (f) increased animal welfare and performance; (g) increased sustainability; (h) reduced emissions, waste and carbon footprint; (i) improved public perception of the sector; (j) increased competitiveness of the sector
During the project period, the DSS has been established and utilised to integrate data and analytical models, information from field tests has been integrated into recommendations for practitioners and disseminated to stakeholders.
What: By taking a cross-scale, multi and trans-disciplinary approach including the direct integration of pig farmers in the project, the project has ensured that all aspects relevant for the development of sustainable, socially acceptable and economically viable pig production systems are adequately addressed.
Why: Current EU pig production is characterised by suboptimal resource utilisation, resulting in unnecessarily high emissions and wastes. At the same time, animal welfare is of increasing concern, farmers are struggling to maintain economic competitiveness, the public image of the sector is poor and legislative pressure is increasing.
Where: The research has been conducted by eight academic centres located in Germany, Denmark, France, Latvia, Sweden and the United Kingdom in collaboration with five commercial (DE, DK, SE) and two institutional (FR) farms to ensure relevance for the sector and direct impact on the partner farms.
Furthermore, sensors systems and control devices for improved understanding of animal welfare and performance as well as to improve barn climate control for increased animal welfare were developed. The developed soft sensors (sensor data fusion) directly link animal based data with environmental data. Other outputs include ‘big data’ to support barn and control system design, sound LCA and LCCA, increased animal welfare and performance, and increased sustainability of production through increased resource efficiency. The goal of PigSys has been to reduce emissions, waste and carbon footprint, and improve the public perception of the sector through a better understanding of interactions between the animals, the technology and environmental conditions. Optimisation based on this information aims to help decreased production costs while increasing animal welfare and thus increase the competitiveness of the sector.
Research findings
Data warehouse (DW) is considered as a cloud-based data storage and processing unit, with capabilities to combine unlimited data sources like other existing systems and available on-farm generated data. The developed DW follows best practices in distributed and asynchronous data processing by utilizing multi-agent techniques in conjunction with a real-time data warehousing approach.
The DW architecture consists of various main components to ensure that information is stored for further analysis. After the data has been entered into the DW, the next step has to be the data analysis and decision making for getting added value from the data. Several potential models can be integrated into the DW to conclude on different aspects of pig status and variation of indoor conditions. Pig status models can include growth model and overall pig health model. Pig behaviour models can include criteria such as pig laying and standing behaviour model, aggressiveness and feeding behaviour. Description of indoor conditions is mainly based on a barn microclimate model. The DW idea and concept allow to integrate models based on the data collected within the DW or provide an API for other external system to get the data from the warehouse and then store the model output back to the warehouse.
From a survey performed in 6 partners’ countries, some common characteristics across pig systems as well as some differences have been identified at the fattening room scale. So, a list of criteria has been defined to be used as inputs in a modelling approach implemented to optimise the pig production system under different outdoor conditions.
The prototype of the DSS has been successfully launched and data from several field trials are included. INRAE and IFIP have successfully merged the animal and building models. The growth-bioclimatic model called Thermipig simulates the thermal balance (with a precision close to 0.5°C) resulting from the outdoor conditions and the heat produced by the animals at the room scale, depending on the characteristics of the room, the animals, the feeding strategy and the management of equipment. It can be used to investigate the impact of any change in one or more of these components on the multi-criteria performance. Therefore, it is possible to assess the interactions between the thermodynamic behaviour of a building with the performance of the pigs in terms of productivity in dependence of building design, breed of pigs and diet used under different climates.
The field trials in Denmark, France, Germany and Sweden have produced valuable information on climatic conditions and its distribution for numerous building designs, weather conditions and animal age.
The Danish field trials have been completed successfully. However, challenges occurred with respect to both ensuring stable data flows and the camera-based weighting system in one of the Danish herds. The latter, resulting in a new system with a 3D camera being used instead of the initially planned. The new system estimates live weight of each individual pig in a pen. Each pig wears an RFID ear tag with a unique ID. ID is read by a scanner when the pig drinks water. The 3D camera for weight estimation is placed above the drinking bowl and the weight of each pig is paired with the ID of the ear tag. [Figure X] shows weight curves of all pigs in two individual pens. Each line represents a pig. In the left plot variation within the pen increases around day 42 (3 days after new feed mixture) indicating suboptimal feed management. In the right plot multiple pigs stop growing at the same time around day 42 and 60, indicating health or welfare issues to be checked up upon.
Data from both Danish test herds have been monitored and weekly reports sent to the farmers. This has increased the farmers’ engagement and is the direct cause of optimization of climate control parameters and feeding strategies in both farms, resulting in increased productivity.
An optical system using deep learning along with machine vision techniques has been developed by UniKassel and TLLLR to provide the possibility to monitor lying, standing and activity levels of pigs in barns and provide early warning in case of problems. This delivers a valuable tool to improve welfare and health of the animals, and thus economic competitiveness and social acceptability.
Sensor data analysis and fusion show unsteady patterns and spatio-temporal distribution of barn climate data and levels of noxious gases which have impact on pigs’ performance and welfare.
Reduced ammonia emission and improved hygiene are crucial for environment protection and societal acceptability, creating improved economic competitiveness. Pig cooling with showers over the slatted area, as well as increasing air velocity at pig lying area, were cost-effective options for mitigation of ammonia emissions and heat stress. Showers performed better overall, both from an environmental and economic perspective. Models to assess the trade-offs between environmental and bio-economic impacts of different strategies for pig housing and manure management were developed. The project website along with dissemination activities is regularly updated and available publicly (http://pigsys.eu/).
A Life Cycle Assessment (LCA) framework able to evaluate pig production systems through a holistic approach and capture the effect that variations in pig housing have on the system environmental impacts has been developed. This has been supplemented by an environmental abatement cost analysis framework, allowing both the environmental and economic impacts of different impact abatement measures in European pig production to be estimated. Different measures in pig housing and manure management were then evaluated through the framework. Results indicated that anaerobic digestion was the most cost-effective investment for global warming mitigation, and slurry acidification for acidification and eutrophication mitigation. Measures for mitigation of global warming potential and abiotic depletion required higher investments than for acidification and eutrophication, but also generated profit. Pig cooling with showers significantly reduced system environmental impact and mitigated heat stress, in a cost-effective way. The framework can be used in future policy and practical considerations to facilitate decision making in farm investments that aim to mitigate the farm environmental impact.
Project consortium
Coordinated by: Dr. Barbara Sturm - University of Kassel (GERMANY)
DENMARK: SEGES Pig Research Centre
FRANCE: INRAE; French Research Institute for Agriculture, Food and Environment and IFIP; French Technical Institute for Pork
GERMANY: TLLLR; Thuringian State Institute for Agriculture and Rural Development
LATVIA: LLU; Latvia University of Life Sciences and Technologies
SWEDEN: SLU; Swedish University of Agricultural Sciences
UNITED KINGDOM: UNEW; Newcastle University, UK
Funded by: BMEL, DAFA, ANR, VIAA, Formas and DEFRA as part of the ERA-NET Cofund SusAn through a virtual common pot model including EU Top-Up funding from the European Union´s Horizon 2020 research and innovation programme (grant agreement no 696231).
More information
The PigSys project started on 1 September 2017 and runs until 31 December 2020.
Aim of the project
PigSys has adopted a multidisciplinary, system level approach to pig production systems with the purpose of generating a deeper understanding of the interactions between the animals, technology and environmental conditions in pig rearing. A model of mass and energy flows and Decision Support system, sensor networks, soft sensors (sensor data fusion) as well as novel building climate control systems, have been developed to underpin sustainable improvement in system performance and increase animal welfare. Furthermore, in depth data analysis of the impact of environmental conditions on animal welfare and performance was conducted and a Life Cycle Analysis conducted
Introduction
There are many approaches for the improvement of individual aspects the system animal-technology-environment. To date, none provide effective whole system considerations for pig production. PigSys addresses these issues by adopting a multi-disciplinary, system level approach.
PigSys will provide the sector with (a) a whole system model of energy and mass flows and decision support system (DSS); (b) sensor systems and control devices for improved barn climate control and animal welfare; (c) ‘big data’ to support barn and control system design; (d) soft sensors (sensor data fusion based on animal and environment related data); (e) LCA and LCCA; (f) increased animal welfare and performance; (g) increased sustainability; (h) reduced emissions, waste and carbon footprint; (i) improved public perception of the sector; (j) increased competitiveness of the sector
During the project period, the DSS has been established and utilised to integrate data and analytical models, information from field tests has been integrated into recommendations for practitioners and disseminated to stakeholders.
What: By taking a cross-scale, multi and trans-disciplinary approach including the direct integration of pig farmers in the project, the project has ensured that all aspects relevant for the development of sustainable, socially acceptable and economically viable pig production systems are adequately addressed.
Why: Current EU pig production is characterised by suboptimal resource utilisation, resulting in unnecessarily high emissions and wastes. At the same time, animal welfare is of increasing concern, farmers are struggling to maintain economic competitiveness, the public image of the sector is poor and legislative pressure is increasing.
Where: The research has been conducted by eight academic centres located in Germany, Denmark, France, Latvia, Sweden and the United Kingdom in collaboration with five commercial (DE, DK, SE) and two institutional (FR) farms to ensure relevance for the sector and direct impact on the partner farms.
Main project activities
A whole system model of energy and mass flows and decision support system have been developed and are accessible.
Furthermore, sensors systems and control devices for improved understanding of animal welfare and performance as well as to improve barn climate control for increased animal welfare were developed. The developed soft sensors (sensor data fusion) directly link animal based data with environmental data. Other outputs include ‘big data’ to support barn and control system design, sound LCA and LCCA, increased animal welfare and performance, and increased sustainability of production through increased resource efficiency. The goal of PigSys has been to reduce emissions, waste and carbon footprint, and improve the public perception of the sector through a better understanding of interactions between the animals, the technology and environmental conditions. Optimisation based on this information aims to help decreased production costs while increasing animal welfare and thus increase the competitiveness of the sector.
Research findings
Data warehouse (DW) is considered as a cloud-based data storage and processing unit, with capabilities to combine unlimited data sources like other existing systems and available on-farm generated data. The developed DW follows best practices in distributed and asynchronous data processing by utilizing multi-agent techniques in conjunction with a real-time data warehousing approach.
The DW architecture consists of various main components to ensure that information is stored for further analysis. After the data has been entered into the DW, the next step has to be the data analysis and decision making for getting added value from the data. Several potential models can be integrated into the DW to conclude on different aspects of pig status and variation of indoor conditions. Pig status models can include growth model and overall pig health model. Pig behaviour models can include criteria such as pig laying and standing behaviour model, aggressiveness and feeding behaviour. Description of indoor conditions is mainly based on a barn microclimate model. The DW idea and concept allow to integrate models based on the data collected within the DW or provide an API for other external system to get the data from the warehouse and then store the model output back to the warehouse.
From a survey performed in 6 partners’ countries, some common characteristics across pig systems as well as some differences have been identified at the fattening room scale. So, a list of criteria has been defined to be used as inputs in a modelling approach implemented to optimise the pig production system under different outdoor conditions.
The prototype of the DSS has been successfully launched and data from several field trials are included. INRAE and IFIP have successfully merged the animal and building models. The growth-bioclimatic model called Thermipig simulates the thermal balance (with a precision close to 0.5°C) resulting from the outdoor conditions and the heat produced by the animals at the room scale, depending on the characteristics of the room, the animals, the feeding strategy and the management of equipment. It can be used to investigate the impact of any change in one or more of these components on the multi-criteria performance. Therefore, it is possible to assess the interactions between the thermodynamic behaviour of a building with the performance of the pigs in terms of productivity in dependence of building design, breed of pigs and diet used under different climates.
The field trials in Denmark, France, Germany and Sweden have produced valuable information on climatic conditions and its distribution for numerous building designs, weather conditions and animal age.
The Danish field trials have been completed successfully. However, challenges occurred with respect to both ensuring stable data flows and the camera-based weighting system in one of the Danish herds. The latter, resulting in a new system with a 3D camera being used instead of the initially planned. The new system estimates live weight of each individual pig in a pen. Each pig wears an RFID ear tag with a unique ID. ID is read by a scanner when the pig drinks water. The 3D camera for weight estimation is placed above the drinking bowl and the weight of each pig is paired with the ID of the ear tag. [Figure X] shows weight curves of all pigs in two individual pens. Each line represents a pig. In the left plot variation within the pen increases around day 42 (3 days after new feed mixture) indicating suboptimal feed management. In the right plot multiple pigs stop growing at the same time around day 42 and 60, indicating health or welfare issues to be checked up upon.
Data from both Danish test herds have been monitored and weekly reports sent to the farmers. This has increased the farmers’ engagement and is the direct cause of optimization of climate control parameters and feeding strategies in both farms, resulting in increased productivity.
An optical system using deep learning along with machine vision techniques has been developed by UniKassel and TLLLR to provide the possibility to monitor lying, standing and activity levels of pigs in barns and provide early warning in case of problems. This delivers a valuable tool to improve welfare and health of the animals, and thus economic competitiveness and social acceptability.
Sensor data analysis and fusion show unsteady patterns and spatio-temporal distribution of barn climate data and levels of noxious gases which have impact on pigs’ performance and welfare.
Reduced ammonia emission and improved hygiene are crucial for environment protection and societal acceptability, creating improved economic competitiveness. Pig cooling with showers over the slatted area, as well as increasing air velocity at pig lying area, were cost-effective options for mitigation of ammonia emissions and heat stress. Showers performed better overall, both from an environmental and economic perspective. Models to assess the trade-offs between environmental and bio-economic impacts of different strategies for pig housing and manure management were developed. The project website along with dissemination activities is regularly updated and available publicly (http://pigsys.eu/).
A Life Cycle Assessment (LCA) framework able to evaluate pig production systems through a holistic approach and capture the effect that variations in pig housing have on the system environmental impacts has been developed. This has been supplemented by an environmental abatement cost analysis framework, allowing both the environmental and economic impacts of different impact abatement measures in European pig production to be estimated. Different measures in pig housing and manure management were then evaluated through the framework. Results indicated that anaerobic digestion was the most cost-effective investment for global warming mitigation, and slurry acidification for acidification and eutrophication mitigation. Measures for mitigation of global warming potential and abiotic depletion required higher investments than for acidification and eutrophication, but also generated profit. Pig cooling with showers significantly reduced system environmental impact and mitigated heat stress, in a cost-effective way. The framework can be used in future policy and practical considerations to facilitate decision making in farm investments that aim to mitigate the farm environmental impact.
Project consortium
Coordinated by: Dr. Barbara Sturm - University of Kassel (GERMANY)
Funded by: BMEL, DAFA, ANR, VIAA, Formas and DEFRA as part of the ERA-NET Cofund SusAn through a virtual common pot model including EU Top-Up funding from the European Union´s Horizon 2020 research and innovation programme (grant agreement no 696231).
More information
Presentations and posters
Research articles
Video