Agro-industrial co-digestion in biogas is the main strategy for diversifying the digester diet and raising productivity above single-substrate operation. But it is also the main source of operational instability.
For example: the seasonal variation of VS in alperujo can reach ±25%, silages lose BMP with storage time and fruit and vegetable waste concentrates peaks of simple sugars.
Correct management requires a minimum quarterly characterisation in triplicate and dynamic adjustment of the organic loading rate (OLR) to the real diet, not to the nominal diet in the contract.
Agro-industrial co-digestion is the most profitable and at the same time most unstable operational scenario in anaerobic digestion.
Combining two or three substrates in a balanced proportion raises the specific methane yield by between 15% and 30% against single-substrate operation, improves the C/N ratio, diversifies the supply of trace nutrients and reduces dependence on a single supplier.
The trade-off is that the digester diet stops being stable: it changes with the agricultural season, with the supplier batch and with the weather of the year.
This article describes why agro-industrial substrates change, how to characterise them to avoid OLR deviations and how to adjust the digester regime to a variable diet without entering the alert zone.
Why agro-industrial co-digestion is more volatile than is acknowledged
In the food and agri-livestock sector, substrates reach the digester with a composition that rarely matches the one declared in the supply contract. Three mechanisms explain that divergence:
- Variability of the operation over the year.
- Variability of the processing.
- Variability of storage: silages and other agri-food by-products progressively lose fermentable organic matter, and fresh waste degrades in the silo at rates that depend on ambient temperature.
In practice, a plant that characterises its substrates once a year has a typical deviation of 15-25% between the nominal OLR (calculated with contract data) and the real OLR (measured with VS triplicates of the current batch).
That deviation is the silent cause of most instability episodes attributed to biology. It is not biology: it is a significant variation in the feed.
Substrate characterisation: real VS versus declared VS
Volatile solids (VS) are the fraction of the substrate actually available for methanogenesis. Substrates with similar TS (total solids) can have very different VS: two batches of pig slurry with 7% TS can have VS of 5.8% and 4.1% respectively.
If the plant calculates its load with TS instead of VS, or uses the VS declared by the supplier instead of the one measured in the laboratory, the real OLR deviates systematically from the nominal value, with a direct impact on methane generation.
The Smallops operational rule is quarterly re-characterisation in triplicate as a minimum, and per batch for highly variable substrates: alperujo from a different mill, seasonal fruit and vegetable waste, silages near the end of the campaign, slurries, and so on.
Triplicates are essential because the spatial heterogeneity within the substrate itself generates an analytical dispersion of 5-10% in VS. Without triplicates, real dispersion cannot be separated from sampling error.
Alperujo: seasonality and lipid composition
Alperujo (the by-product of two-phase olive mills) is one of the most interesting and most volatile substrates in agro-industrial co-digestion. It has a high theoretical BMP (280-380 NmL CH₄/g VS), but that figure hides significant variability for two reasons.
First: the lipid fraction. It varies between 5% and 15% on VS depending on the harvest and the centrifugation regime. Lipids have an extremely high BMP (700-1,000 NmL CH₄/g VS), but they are also the most inhibitory fraction.
Above 100-200 mg/L of long-chain fatty acids in the digester, syntrophic β-oxidation slows down and propionic acid starts to accumulate.
Second: the polyphenol fraction. The polyphenols in alperujo, known for their natural antimicrobial character in preserving olive oil, also affect the methanogenic consortium.
Total polyphenol concentrations above 5 g/L in the digestate indicate excessive alperujo input and are associated with slower kinetics without explicit decoupling: FOS/TAC does not rise, but the CH₄ yield does fall. Operational control requires seasonal polyphenol analysis when the mixture includes more than 15% alperujo on VS.
Silages: the effect of prior fermentation on BMP
Maize silage is the reference energy substrate in co-digestion, thanks to its stable availability and predictable BMP (290-340 NmL CH₄/g VS). The lactic fermentation that occurs during ensiling converts soluble sugars into lactic acid, which improves preservation.
What is operationally critical is that the BMP of silage decreases with storage time:
- 0.5-1% per month under well-sealed conditions.
- 3-5% per month in silos with air ingress.
A plant that opens a silo in May with a BMP measured in November of the previous year may be working with a real BMP between 5% and 15% lower than assumed. If the OLR is kept constant, the real load is proportionally lower and productivity falls with no apparent biological cause.
Fruit and vegetable waste: seasonal peak and sugar peaks
Fruit and vegetable waste (by-products of packing centres, rejects from wholesale markets, juice processing residue) has an attractive BMP (350-450 NmL CH₄/g VS) and enters the mixture very easily because it is usually cheap or free.
But it has three particularities that make it dangerous in co-digestion without control:
- Concentration of simple sugars: fruit waste contains 40-70% fermentable sugars on VS. An abrupt input generates an acidogenesis peak that methanogenesis cannot absorb at the same rate, and FOS/TAC rises in a characteristic way, with low propionic and high acetic acid.
- Strong seasonality: in August and September there may be surpluses that tempt an increase in the fraction in the mixture; in February and March the waste dwindles and the plant must adjust the diet. If the OLR is kept constant, the sugar load varies even though the plant believes it is operating stably.
- Low inlet pH: much of it arrives at pH 3.5-4.5 from spontaneous fermentation during transport. If the fraction exceeds 15% without prior pH correction, the digester loses buffer capacity and FOS/TAC destabilises faster than the operator anticipates.
How to adjust the OLR dynamically in agro-industrial co-digestion
Dynamic adjustment of the OLR to the real diet requires a simple but disciplined system of four steps, repeated at every significant change of mixture.
1 · Measure VS and BMP of the new batch or campaign
Measure TS and VS of the new batch or campaign, and its BMP when the substrate is new or has changed origin. If the substrate is highly variable (alperujo, fruit and vegetable waste), repeat the characterisation for every significant batch.
The cost of three quarterly analyses is marginal compared with that of operating for a week with a miscalculated OLR.
If the substrate is genuinely new, the BMP falls short: it is a batch test that measures the maximum potential under ideal conditions. The semi-continuous test feeds a laboratory reactor continuously over several weeks and reveals the kinetics, the tolerance to load, the inhibitions and the acclimation time. With substrates like alperujo, which carry lipids and polyphenols, it is the only way to anticipate at what fraction they start to cause trouble.
2 · Recalculate the real OLR in kg VS/m³·day
With the real measured VS, recalculate the effective organic loading rate of the digester.
If the new OLR exceeds 110% of the nominal, reduce the flow or the percentage of the changing substrate until returning to the design OLR.
3 · Increase the frequency of sentinel variables
For 2-3 weeks after any significant substrate change, move the frequency of FOS/TAC and individual VFAs from weekly to twice weekly.
That reinforced vigilance makes it possible to detect incipient decoupling before it turns into an alert.
4 · Adjust the mixture, not the geometry
If the changing diet destabilises the digester, the correct response is not to modify the HRT or the mixing regime, but to adjust the mixture: reduce the fraction of the problematic substrate, compensate with a more stable substrate and add alkalinity if the pH falls.
The geometry of the digester is a given; the mixture is the operational variable.
Summary table: BMP ranges and operational particularities
| Substrate | Typical BMP (NmL CH₄/g VS) | Seasonal variability | Main operational risk |
|---|---|---|---|
| Alperujo (two-phase) | 280-380 | ±25% between campaigns | Lipids (LCFAs) and polyphenols |
| Maize silage | 290-340 | 5-15% depending on storage | BMP decreasing over time |
| Fruit and vegetable waste | 350-450 | High seasonality | Sugar peaks and low pH |
| Pig slurry | 180-280 | ±15% depending on animal diet | High TAN, free NH₃ |
| WWTP sludge | 150-350 | Stable (primary or secondary) | Low real BMP, low energy |
Operational case: 1 MWe agro-industrial plant with a seasonal diet
A 1 MWe plant (mesophilic, 37 °C) with a nominal diet of 50% pig slurry, 30% maize silage and 20% fruit and vegetable waste. Design OLR of 3.8 kg VS/m³·day and a historical specific production of 0.36 Nm³ CH₄/kg VS.
Seasonal symptom: every August and September productivity falls to 0.28-0.30 Nm³ CH₄/kg VS.
Diagnosis from the analytical characterisation
Re-characterisation in triplicate of the three substrates in August:
- The fruit and vegetable waste (peach and nectarine campaign) has VS of 12.5% against the 9.5% declared in the contract.
- Simple sugars account for 58% of the VS.
- Root cause: the real OLR during August and September is 4.5 kg VS/m³·day, 18% above nominal, with a very high fraction of fermentable sugars.
It is not seasonal biology: it is acidogenesis accelerated by hidden overload.
Operational intervention
Reduction of the fruit and vegetable fraction from 20% to 13% during the peak campaign months (July to September), compensating with an additional 7% of maize silage.
Reinforced frequency of FOS/TAC and individual VFAs throughout the season. Neither the HRT nor the mixing is modified: only the mixture.
Result after a full campaign
Specific productivity in August and September: 0.35 Nm³ CH₄/kg VS, against the historical 0.28-0.30 (+17%).
FOS/TAC in the stable zone (< 0.35) during 91% of the campaign period, with zero excursions into the crisis zone (> 0.5).
Cost of the intervention: four additional quarterly characterisations in triplicate (around EUR 2,400/year).
Economic recovery: the annual dip goes from 3-4 weeks to zero, with an estimated saving of EUR 28,000-35,000/year in production not lost. ROI ≥ 12x.
Frequently asked questions about agro-industrial co-digestion
Why does the BMP of the same substrate change between campaigns?
For three reasons. The biological variability of the crop (the composition of alperujo changes depending on whether the harvest is early or late). The variability of the processing (the centrifugation regime of the mill, for example, alters the lipid fraction). And the loss during storage, which in silages ranges from 0.5-1% monthly when well sealed to 3-5% monthly with air ingress.
That is why a reliable BMP under VDI 4630 must be measured every campaign, or at every significant change of supplier.
How is a seasonal substrate characterised correctly?
The minimum characterisation includes TS, VS, pH, elemental composition (C, N, P) and a reliable BMP under VDI 4630, with an ISR ≥ 2 and an acclimated inoculum.
Recommended frequency: quarterly as a minimum for stable substrates, and per batch for highly variable ones. Triplicates are mandatory to separate real dispersion from sampling error.
What should be done when a new substrate arrives at the plant?
Before introducing it into co-digestion, run three steps: full characterisation in triplicate (TS, VS, BMP, pH, composition), and a semi-continuous test if the substrate is unknown or problematic; gradual entry at 5-10% of the mixture over 2-3 weeks with reinforced FOS/TAC and VFA frequency; and validation of stability before raising the fraction to the final target.
Skipping this sequence is the most frequent cause of instabilities attributed to unpredictable biology.
What is the typical BMP of alperujo?
It sits between 280 and 380 NmL CH₄/g VS, with a seasonal variability that can reach ±25% between campaigns.
The lipid fraction (5-15% on VS depending on the harvest) raises the theoretical BMP, but also the risk of inhibition by long-chain fatty acids. Total polyphenols above 5 g/L in the digestate indicate excessive input and are associated with slower kinetics without explicit decoupling.
Why does a high alperujo BMP not translate into more methane?
Because the BMP is measured in batch mode, with a single load and under ideal conditions. There, the lipids in alperujo give an extremely high potential, and the polyphenols barely have time to take their toll.
In the real digester, fed every day, those two fractions accumulate: the LCFAs slow down syntrophic β-oxidation and the polyphenols slow the kinetics without FOS/TAC rising. A semi-continuous test, feeding the reactor for several weeks with the planned fraction, is what shows that cumulative effect before taking it to the plant.
Does your plant have a seasonal dip every year?
It is probably not biology: it is out-of-date analytical characterisation. Request an Operational Excellence Diagnosis and we will audit the analytical traceability of your co-digestion.
Normative and bibliographic references
Mata-Alvarez, J. et al. (2014). A critical review on anaerobic co-digestion. Renewable and Sustainable Energy Reviews, 36, 412-427. → doi.org/10.1016/j.rser.2014.04.039
Bouallagui, H. et al. (2009). Mesophilic biogas production from fruit and vegetable waste. Renewable Energy, 34 (1), 80-86. → doi.org/10.1016/j.renene.2008.04.007
VDI 4630 (2016). Fermentation of organic materials. Verein Deutscher Ingenieure.
Borja, R. et al. (2006). Anaerobic digestion of two-phase olive mill effluent. Process Biochemistry, 41 (6), 1268-1274. → doi.org/10.1016/j.procbio.2005.12.010