Final Report Summary - CROPGEN (Renewable energy from crops and agrowastes)
The overall objective of the CROPGEN project research was to produce from biomass a sustainable fuel source that can be integrated into the existing energy infrastructure in the medium term, and in the longer term will also provide a safe and economical means of supplying the needs of a developing hydrogen fuel economy. The work was based on the use of anaerobic digestion (AD) as a means of producing methane from biomass, including energy crops and agricultural residues.
The technology for methane generation by biochemical means is well established: the breakthrough to a cost-effective and competitive energy supply will come from engineering and technical improvements that can be made to increase biomass conversion efficiencies, and from reductions in the cost of biomass. In addition to developing process technology aspects of AD, the research determined how the technology can best be applied to provide a versatile, low-cost, carbon neutral gaseous biofuel within an environmentally sound and sustainable agricultural framework.
An extensive review was undertaken of the available literature on potential crops, including non-food crops, energy crops and other plant species not currently grown as crops. Data collection concentrated on those species suitable for growing in European locations and included some that are currently regarded as energy crops and natural vegetation that could be used for energy production. It was concluded that a number of criteria were useful in assessing the value of crops as feedstock for anaerobic digestion. These included biomass/biogas yield; input energy requirements; alternative/secondary crop uses; land use/availability; and ecological and environmental factors.
The energy required in crop production can be divided into two types: direct energy - fuel used in field operations and labour; and indirect energy - energy required to produce equipment and materials used in the production of the crop. Indirect energy includes that required for fertilisers and pesticides, and the manufacture, delivery and maintenance of agricultural equipment. Information on crop yields (where known) climate and fertiliser requirements, potential methane yields, cultivation operations was included in a crop production database (Microsoft Access format) using average values reported in the literature.
Other crops, for example white mustard (Sinapis alba) and stubble turnips (Brassica rapa), can be used as 'catch' crops and provide nutrient capture, an energy source and soil cover at times when land might otherwise be fallow, thus increasing overall biomass yield per hectare within the agricultural year. Mixed cropping, crop rotation, double cropping and intercropping are all possible ways of incorporating energy crops into more effective production cycles to maximise net energy gain per unit of land area. Crop residues left on the land also form another important source of biomass and include cereal straws and vegetative biomass from root crops.
The biochemical methane potential (BMP) test has been widely used to assess the digestibility of organic matter, including plant material. The project has established a database of more than 700 BMP values from both laboratory determinations and the literature. Additional experiments were undertaken by other partners to assess factors that could influence the BMP value, in an effort to standardise the procedure and facilitate interpretation. These included the method of preparation of the material, particle size, source of inoculum; substrate to inoculum ratio, and the use of buffers.
From an economic viewpoint it is important to identify and minimise any energy losses during storage. Experiments were conducted with grasses (mixtures of timothy and meadow fescue, timothy and clover and ryegrass) and sugar beet tops by partners at JyU and for maize (whole crop and grains) at BOKU.
Different pretreatment methods including size reduction, alkaline treatment and thermal pretreatment (steam explosion) were evaluated. No significant effects on methane production were observed by reducing particle sizes of maize from a distribution typical of forage feeds down to 1 mm. Thermal pre-treatment at varying temperatures and reaction times was used on grass and whole crop maize silage and showed a considerable difference in effect between the two crops in terms of making sugars available. While this kind of treatment appears suitable for fibrous material like grasses, it cannot be recommended for substrates with high starch content like maize.
Silage additives based on homo and heterofermentative lactic acid bacteria were compared with untreated silage. The addition of amylase, silage stored under semi-aerobic conditions, and silage spoiled with Clostridium tyrobutyricum were also tested. Conventional silage additives did not improve the biogas yield, but the results on storage stability in a full-scale silo indicate that they may reduce silage losses. Spoilage with C. tyrobutyricum, however, showed a significant improvement in methane yields as a result of the hydrolysis of starch and the formation of butyric acid; but the addition of C. tyrobutyricum was not a pre-requisite as Clostridium strains were predominant in the buffered silages, even without pre-inoculation.
Work carried out by J y U and Metener looked at ryegrass and timothy-clover stored in the field in plastic covered bales at ambient temperatures and also stored for 2 - 6 months under controlled conditions in laboratory studies. Effects of additives and moisture content were studied in detail and showed that methane yield in digestion tests was unaffected if timothy-clover was stored without pre-wilting and ryegrass stored after 48 h of pre-wilting, although actual harvest conditions (moisture) may affect pre-wilting requirements.
Experiments were carried out in bench, pilot and full-scale digesters on a variety of different substrates. Trials on sunflower flour at CSIC showed good operational stability up to loadings of 2 kgVS m-3 d-1 with specific methane yields close to the maximum observed at hydraulic retention time (HRT) > 25 days. At a loading of 3 kgVS m-3 d-1 the specific methane yield decreased considerably although the maximum volumetric yields were reached at this loading with a HRT of 30 days. Digestion at this loading was unstable, however, with high volatile fatty acid (VFA) concentrations and a poor alkalinity ratio.
Co-digestion trials at full scale (150 m3 digester) were undertaken by Metener using the same mixes in which the silage was chopped with a diet feeder before mixing with the slurry in a pre-mix tank and feeding to the digester.
Dealing with relatively novel feedstocks such as energy crops requires consideration of whether better reactor designs and operating regimes are possible. The research has looked at a range of systems from low-cost static bed reactors to complex multi-phase systems with uncoupling of solids and liquid retention times. Trials with static permeating bed systems were carried out at Soton using ensiled whole crop fodder maize as both the substrate and bed medium. The maize showed very rapid acidification reaching low pH values making the use of a single-phase batch reactor, with or without liquid recirculation, impractical due to the very low rates of solids destruction.
The work therefore focused on two-phase systems in which the bed liquors were removed and replaced by clean water or by recycled liquid after treatment in a second stage methanogenic reactor. The research investigated the effects of changing the flush rate, inoculums to substrate ratio in the bed, the effect of introducing buffering, and the interaction between phases. Physical and chemical factors relating to the flush water had some effect on the solids destruction and theoretical methane yield from the hydrolysis and acidification products generated.
Major improvements in solids destruction, however, depended not just on replacement of the bed liquor to flush out acid products, but on recycling of the effluent from the second phase reactor to provide a constant source of re-inoculation with microorganisms. This also had the effect of allowing the first stage reactor to become methanogenic, even at fairly short solids retention times, which helped to maintain the pH and buffering of the system leading to improved rates of solid destruction. Work at J y U also showed the necessity of a two phase system when using leach beds for the digestion of crops such as grass silage and sugar beet that contain a proportion of very readily degradable material.
Work by Soton and OPL on plug flow digestion of ryegrass indicated the importance of optimising the recycle ratio by retaining solids. This has the advantage of both providing an inoculum and obtaining any residual energy from the substrate. There are thus similarities between an optimal plug flow approach for high solids substrates and the conditions required for successful operation of a leach bed reactor.
BOKU-IAM developed a software tool for simulation of AD in the form of a virtual laboratory for processing and interpretation of data, formulation of 'fit for purpose' mathematical models, training purposes and dissemination of results. The heart of the software is an adapted version of Anaerobic Digestion Model 1 (ADM1). Outputs from the model are the gas production, methane content, pH, volatile fatty acid concentration, proprionic acid concentration, acetic acid concentration and the COD Reduction. The model has so far been calibrated for nine substrates (blue, white and yellow lupin, maize, soy, sunflower, rape, rye and wheat), and up to four of these can be mixed.
A fuzzy logic control tool was also developed by BOKU-IAM using as input either the total fatty acid concentration, the pH or the proprionic acid concentration, with the methane content and the gas production and organic loading rates. The output gives the organic loading rate to be applied for the following day. The tool was tested with ADM1 to further refine the program against a number of appraisal factors including gas production, methane content, concentration of acetic, proprionic and total volatile fatty acids, COD reduction and pH. The fuzzy algorithm was then incorporated into a graphical user interface and tested on laboratory reactors. Control by this means showed an improved methane yield and more stable reactor conditions.
The cultivation of energy crops requires inputs of time and energy that have to be taken into account when modelling both the net energy productivity of a biomass-to-energy scheme and the economic costs. Questionnaires were used to collect data on cultivation, fertilisation, harvest and transport from farmers supplying the biogas plant in Strem, Austria. The required working time and fuel consumption of each step were recorded with the system border defined as the unloading point of the digester feedstock.
The research has produced a number of case studies for crop-based AD energy systems, and typical values of energy inputs into a commercial scale energy crop digester are given in table 11 with the probable biogas yield and thermal energy value. Thus the net crop energy yield is 127 GJ ha-1. The real energy gain depends on how the methane is used, for example, conversion to electricity without heat recovery would only give a net gain of 17 GJ ha-1 (4722 kWh ha-1) whereas conversion to vehicle fuel yields a net gain of 77 GJ ha-1 (equivalent to 2150 l of diesel).
The efficiency can be improved, for example by replacing fossil fuel based fertilisers with digestate; increasing the efficiency of electricity generation; reducing heat loss from the digester (or using self heating materials). The energy yield per hectare of crop used can also be improved by, for example: increasing the crop yield per hectare; reducing fertiliser inputs (e.g. by using legumes); reducing the number of crop operations; supplementing the crop based feedstock with animal slurries and agrowastes.
Overall, the project has shown that biogas production on farms is technically feasible using conventional technology and can be applied to a wide range of feedstocks ranging from dedicated energy crops to agricultural residues. When using energy crops for methane production the net energy gain per hectare of land it is very favourable compared to first generation liquid biofuels, the chief advantages being the ability to use whole crop material either as the main substrate or as a co-digestate, and the possibility of retaining nutrients and returning these to the land. Despite the high-energy returns the economics of using only cultivated energy crops remain marginal due to the costs associated with crop cultivation.
The technology for methane generation by biochemical means is well established: the breakthrough to a cost-effective and competitive energy supply will come from engineering and technical improvements that can be made to increase biomass conversion efficiencies, and from reductions in the cost of biomass. In addition to developing process technology aspects of AD, the research determined how the technology can best be applied to provide a versatile, low-cost, carbon neutral gaseous biofuel within an environmentally sound and sustainable agricultural framework.
An extensive review was undertaken of the available literature on potential crops, including non-food crops, energy crops and other plant species not currently grown as crops. Data collection concentrated on those species suitable for growing in European locations and included some that are currently regarded as energy crops and natural vegetation that could be used for energy production. It was concluded that a number of criteria were useful in assessing the value of crops as feedstock for anaerobic digestion. These included biomass/biogas yield; input energy requirements; alternative/secondary crop uses; land use/availability; and ecological and environmental factors.
The energy required in crop production can be divided into two types: direct energy - fuel used in field operations and labour; and indirect energy - energy required to produce equipment and materials used in the production of the crop. Indirect energy includes that required for fertilisers and pesticides, and the manufacture, delivery and maintenance of agricultural equipment. Information on crop yields (where known) climate and fertiliser requirements, potential methane yields, cultivation operations was included in a crop production database (Microsoft Access format) using average values reported in the literature.
Other crops, for example white mustard (Sinapis alba) and stubble turnips (Brassica rapa), can be used as 'catch' crops and provide nutrient capture, an energy source and soil cover at times when land might otherwise be fallow, thus increasing overall biomass yield per hectare within the agricultural year. Mixed cropping, crop rotation, double cropping and intercropping are all possible ways of incorporating energy crops into more effective production cycles to maximise net energy gain per unit of land area. Crop residues left on the land also form another important source of biomass and include cereal straws and vegetative biomass from root crops.
The biochemical methane potential (BMP) test has been widely used to assess the digestibility of organic matter, including plant material. The project has established a database of more than 700 BMP values from both laboratory determinations and the literature. Additional experiments were undertaken by other partners to assess factors that could influence the BMP value, in an effort to standardise the procedure and facilitate interpretation. These included the method of preparation of the material, particle size, source of inoculum; substrate to inoculum ratio, and the use of buffers.
From an economic viewpoint it is important to identify and minimise any energy losses during storage. Experiments were conducted with grasses (mixtures of timothy and meadow fescue, timothy and clover and ryegrass) and sugar beet tops by partners at JyU and for maize (whole crop and grains) at BOKU.
Different pretreatment methods including size reduction, alkaline treatment and thermal pretreatment (steam explosion) were evaluated. No significant effects on methane production were observed by reducing particle sizes of maize from a distribution typical of forage feeds down to 1 mm. Thermal pre-treatment at varying temperatures and reaction times was used on grass and whole crop maize silage and showed a considerable difference in effect between the two crops in terms of making sugars available. While this kind of treatment appears suitable for fibrous material like grasses, it cannot be recommended for substrates with high starch content like maize.
Silage additives based on homo and heterofermentative lactic acid bacteria were compared with untreated silage. The addition of amylase, silage stored under semi-aerobic conditions, and silage spoiled with Clostridium tyrobutyricum were also tested. Conventional silage additives did not improve the biogas yield, but the results on storage stability in a full-scale silo indicate that they may reduce silage losses. Spoilage with C. tyrobutyricum, however, showed a significant improvement in methane yields as a result of the hydrolysis of starch and the formation of butyric acid; but the addition of C. tyrobutyricum was not a pre-requisite as Clostridium strains were predominant in the buffered silages, even without pre-inoculation.
Work carried out by J y U and Metener looked at ryegrass and timothy-clover stored in the field in plastic covered bales at ambient temperatures and also stored for 2 - 6 months under controlled conditions in laboratory studies. Effects of additives and moisture content were studied in detail and showed that methane yield in digestion tests was unaffected if timothy-clover was stored without pre-wilting and ryegrass stored after 48 h of pre-wilting, although actual harvest conditions (moisture) may affect pre-wilting requirements.
Experiments were carried out in bench, pilot and full-scale digesters on a variety of different substrates. Trials on sunflower flour at CSIC showed good operational stability up to loadings of 2 kgVS m-3 d-1 with specific methane yields close to the maximum observed at hydraulic retention time (HRT) > 25 days. At a loading of 3 kgVS m-3 d-1 the specific methane yield decreased considerably although the maximum volumetric yields were reached at this loading with a HRT of 30 days. Digestion at this loading was unstable, however, with high volatile fatty acid (VFA) concentrations and a poor alkalinity ratio.
Co-digestion trials at full scale (150 m3 digester) were undertaken by Metener using the same mixes in which the silage was chopped with a diet feeder before mixing with the slurry in a pre-mix tank and feeding to the digester.
Dealing with relatively novel feedstocks such as energy crops requires consideration of whether better reactor designs and operating regimes are possible. The research has looked at a range of systems from low-cost static bed reactors to complex multi-phase systems with uncoupling of solids and liquid retention times. Trials with static permeating bed systems were carried out at Soton using ensiled whole crop fodder maize as both the substrate and bed medium. The maize showed very rapid acidification reaching low pH values making the use of a single-phase batch reactor, with or without liquid recirculation, impractical due to the very low rates of solids destruction.
The work therefore focused on two-phase systems in which the bed liquors were removed and replaced by clean water or by recycled liquid after treatment in a second stage methanogenic reactor. The research investigated the effects of changing the flush rate, inoculums to substrate ratio in the bed, the effect of introducing buffering, and the interaction between phases. Physical and chemical factors relating to the flush water had some effect on the solids destruction and theoretical methane yield from the hydrolysis and acidification products generated.
Major improvements in solids destruction, however, depended not just on replacement of the bed liquor to flush out acid products, but on recycling of the effluent from the second phase reactor to provide a constant source of re-inoculation with microorganisms. This also had the effect of allowing the first stage reactor to become methanogenic, even at fairly short solids retention times, which helped to maintain the pH and buffering of the system leading to improved rates of solid destruction. Work at J y U also showed the necessity of a two phase system when using leach beds for the digestion of crops such as grass silage and sugar beet that contain a proportion of very readily degradable material.
Work by Soton and OPL on plug flow digestion of ryegrass indicated the importance of optimising the recycle ratio by retaining solids. This has the advantage of both providing an inoculum and obtaining any residual energy from the substrate. There are thus similarities between an optimal plug flow approach for high solids substrates and the conditions required for successful operation of a leach bed reactor.
BOKU-IAM developed a software tool for simulation of AD in the form of a virtual laboratory for processing and interpretation of data, formulation of 'fit for purpose' mathematical models, training purposes and dissemination of results. The heart of the software is an adapted version of Anaerobic Digestion Model 1 (ADM1). Outputs from the model are the gas production, methane content, pH, volatile fatty acid concentration, proprionic acid concentration, acetic acid concentration and the COD Reduction. The model has so far been calibrated for nine substrates (blue, white and yellow lupin, maize, soy, sunflower, rape, rye and wheat), and up to four of these can be mixed.
A fuzzy logic control tool was also developed by BOKU-IAM using as input either the total fatty acid concentration, the pH or the proprionic acid concentration, with the methane content and the gas production and organic loading rates. The output gives the organic loading rate to be applied for the following day. The tool was tested with ADM1 to further refine the program against a number of appraisal factors including gas production, methane content, concentration of acetic, proprionic and total volatile fatty acids, COD reduction and pH. The fuzzy algorithm was then incorporated into a graphical user interface and tested on laboratory reactors. Control by this means showed an improved methane yield and more stable reactor conditions.
The cultivation of energy crops requires inputs of time and energy that have to be taken into account when modelling both the net energy productivity of a biomass-to-energy scheme and the economic costs. Questionnaires were used to collect data on cultivation, fertilisation, harvest and transport from farmers supplying the biogas plant in Strem, Austria. The required working time and fuel consumption of each step were recorded with the system border defined as the unloading point of the digester feedstock.
The research has produced a number of case studies for crop-based AD energy systems, and typical values of energy inputs into a commercial scale energy crop digester are given in table 11 with the probable biogas yield and thermal energy value. Thus the net crop energy yield is 127 GJ ha-1. The real energy gain depends on how the methane is used, for example, conversion to electricity without heat recovery would only give a net gain of 17 GJ ha-1 (4722 kWh ha-1) whereas conversion to vehicle fuel yields a net gain of 77 GJ ha-1 (equivalent to 2150 l of diesel).
The efficiency can be improved, for example by replacing fossil fuel based fertilisers with digestate; increasing the efficiency of electricity generation; reducing heat loss from the digester (or using self heating materials). The energy yield per hectare of crop used can also be improved by, for example: increasing the crop yield per hectare; reducing fertiliser inputs (e.g. by using legumes); reducing the number of crop operations; supplementing the crop based feedstock with animal slurries and agrowastes.
Overall, the project has shown that biogas production on farms is technically feasible using conventional technology and can be applied to a wide range of feedstocks ranging from dedicated energy crops to agricultural residues. When using energy crops for methane production the net energy gain per hectare of land it is very favourable compared to first generation liquid biofuels, the chief advantages being the ability to use whole crop material either as the main substrate or as a co-digestate, and the possibility of retaining nutrients and returning these to the land. Despite the high-energy returns the economics of using only cultivated energy crops remain marginal due to the costs associated with crop cultivation.