When biogas production drops, the question is not what happened, but in what order to look. Almost all drops come down to seven causes, and most can be told apart with data the plant already has.
The first three (an uncharacterised diet change, real-versus-nominal OLR overload and free NH₃ inhibition) explain most cases. The next four are less frequent, but they are diagnosed just as quickly if you know what to measure.
This article ranks the seven by probability and proposes a diagnostic tree to reach the root cause in 48-72 hours.
It is the most frequent call any biogas process engineer receives: “we have been producing 15 % less for three weeks and we do not know why”.
The problem is rarely a lack of information. The problem is looking in the wrong order: starting with the most sophisticated (the biology, the consortium, the additives) when the cause is usually in the most basic (what enters the digester and how much).
This article proposes the reverse order: from most likely to least likely, and from cheapest to check to most expensive.
Why biogas production drops: the diagnostic problem
A production drop is a symptom, not a diagnosis. And it is a symptom shared by very different causes, which have to be assessed quickly to reach a solution.
The real difficulty is that several of those causes produce the same pattern in the variables the plant watches daily. A 15 % drop with stable pH can be a hidden overload, an incipient inhibition or simply a worse substrate.
What separates them is not measuring more, but measuring in the right order. A structured diagnosis reaches the root cause in 5-10 working days; trial and error can take months and do damage along the way.
Two principles before starting:
- First rule out the physical and the analytical. If the starting data is bad (poorly measured VS, degraded sample), everything that comes after is noise.
- One cause does not exclude another. In plants with persistent drops, the usual finding is a combination of two: a main one and one that aggravates it.
Cause 1 · Uncharacterised diet change
It is the most frequent cause, by far. And it is almost never perceived as a change.
You do not need to change supplier to change diet: it is enough for the same supplier to deliver a different batch. A silage from another season, a residue from another campaign, a slurry from animals on another feed.
The mechanism is simple: the plant keeps feeding the same cubic metres, but the volatile solids content has changed. If the VS drop by 20 %, the real load drops by that same percentage and production with it, without anything being “wrong”.
How to check it: triplicate characterisation of the current batch (TS, VS) and comparison with the value being used in the load calculation. It costs little and is resolved in 48 hours.
Characteristic signal: the specific production (Nm³ CH₄/kg VS fed) holds, but the absolute production falls. If the yield per kilo is normal, the problem is how much comes in, not how it digests.
Cause 2 · OLR overload (real versus nominal)
It is the flip side of the previous one and almost as frequent.
Here the VS of the substrate have risen without anyone knowing, and the plant is overloading the digester while believing it operates on spec. The typical divergence between the nominal OLR (calculated with contract data) and the real OLR (measured with triplicates of the current batch) is 15-25 %.
The overload does not lower production immediately: first it raises it. Then the VFA accumulate, the pH falls and the system blocks.
How to check it: recalculate the OLR with the real measured VS. If it exceeds 110 % of the nominal one, there is your explanation.
Characteristic signal: the FOS/TAC rises before production falls, and propionic rises before acetic. The detail of this reading is in stabilise the anaerobic digester.
Cause 3 · Free NH₃ inhibition
The third in frequency, and the first in plants with protein-rich substrates or pig slurry.
What inhibits is not the total ammonia nitrogen (TAN), but its free fraction (NH₃), which depends on the TAN, the pH and the temperature. That is why a plant can have stable TAN for months and enter inhibition because its pH has risen.
The reference thresholds:
- 150 mg N-NH₃/L in a non-acclimated consortium: incipient inhibition.
- 700 mg N-NH₃/L: collapse, unless previously acclimated.
How to check it: measure the TAN and calculate the free NH₃ with the real pH and temperature of the digester. The calculation is detailed in free ammonia.
Characteristic signal: falling production with moderate FOS/TAC and elevated propionic. Ammonia inhibition slows methanogenesis without triggering acidogenesis.
Cause 4 · Dissolved sulphide inhibition
It appears when the substrate brings sulphates: vinasses, some food-industry residues, slurries with sulphated waters.
Sulphate-reducing bacteria compete with the methanogens for the same substrate and win. On top of that, the sulphide they produce is toxic to the consortium and precipitates the trace metals the archaea need as cofactors.
The operational reference threshold is 200 mg S/L of dissolved sulphides.
How to check it: dissolved sulphides in digestate and H₂S in biogas. An H₂S that rises while production falls is a fairly clear signature.
Characteristic signal: a production drop accompanied by rising H₂S and, often, problems in the cogeneration engine.
Cause 5 · Loss of digester temperature
The easiest to check and the most often overlooked, because the probe reads what it reads and nobody questions it.
The critical range in mesophilic operation is narrow: ±2 °C around 37 °C. A sustained deviation of 3-4 degrees is enough to slow the kinetics measurably. That said, most of the time it is recoverable when this is the cause.
The usual causes are mechanical: a fouled or scaled heat exchanger, a failed trap, deteriorated insulation, or simply cold substrate entering in winter that the heating system no longer compensates.
How to check it: the temperature history and, above all, checking the probe against an independent thermometer. A drifted probe is a surprisingly common root cause.
Characteristic signal: a gradual, seasonal drop that worsens in the cold months.
Cause 6 · Entry of antibiotics or disinfectants
Specific to livestock plants, and the hardest to see because it leaves no conventional analytical trace.
Pharmacological residues enter with the slurry as a bolus, not in constant flow, coinciding with the farm medication campaigns. Tylosin inhibits methanogenesis above 50 mg/L; glutaraldehyde, a common disinfectant, above 100 mg/L.
Below those thresholds the effect is sublethal: it slows the kinetics without decoupling the process. The plant sees a stable FOS/TAC and production falling 8-15 %.
How to check it: not with an analysis, but with the farm calendar. Cross the drop dates with the medication campaigns. If they coincide twice, it is no longer a coincidence.
Characteristic signal: recurring drops that recover on their own in 2-3 weeks, without intervention. The detail in biogas from slurry.
Cause 7 · LCFA inhibition
The last in frequency, but very likely if the plant has raised the fat fraction.
Long-chain fatty acids adsorb onto the membrane of the microorganisms and physically block substrate transport. It is not toxicity: it is a surface phenomenon and it is reversible.
It starts to show from 100-200 mg/L of total LCFA in digestate, and appears when the lipid fraction exceeds 5-8 % in VS.
How to check it: review whether the fat fraction in the diet has changed and measure LCFA in digestate.
Characteristic signal: the most misleading of all. The FOS/TAC stays moderate while methane falls, because the inhibition slows β-oxidation without accumulating light acids. The full mechanism, in LCFA inhibition.
Diagnostic tree in 48-72 hours
The order matters because each step rules out causes and makes the next one cheaper.
| Step | What is measured | What it rules out or confirms |
|---|---|---|
| 0 · First hours | Checking the temperature probe against an independent thermometer | Cause 5. It is free and rules out the silliest one |
| 1 · Day 1 | TS and VS of the current batch in triplicate; recalculation of the real OLR | Causes 1 and 2, the two most frequent |
| 2 · Day 1-2 | FOS/TAC and individual VFA (acetic and propionic) | Separates a load problem from an inhibition problem |
| 3 · Day 2 | TAN, pH and temperature → free NH₃ calculation | Cause 3 |
| 4 · Day 2-3 | Dissolved sulphides and H₂S in biogas | Cause 4 |
| 5 · Day 3 | Farm medication calendar and changes in the lipid fraction | Causes 6 and 7 |
The combined reading of steps 1 and 2 already orients the diagnosis in most cases:
- Normal specific production, low absolute production → less substrate is coming in than believed (cause 1).
- High FOS/TAC and high acetic → overload (cause 2).
- Moderate FOS/TAC and high propionic → inhibition: ammonia, sulphides or LCFA (causes 3, 4 or 7).
- Everything normal and low production → look at temperature, pharmacology or physical problems such as crusts and sediments.
This tree is the summarised version of the biogas plant diagnosis over 14 variables.
Key quantified data
| Concept | Value |
|---|---|
| Typical divergence between real and nominal OLR | 15-25 % |
| Free NH₃ inhibition threshold (non-acclimated) | 150 mg N/L |
| Dissolved sulphide inhibition threshold | 200 mg S/L |
| Critical thermal range in mesophilic operation | ±2 °C around 37 °C |
| LCFA inhibition threshold | 100-200 mg/L in digestate |
| Tylosin threshold | 50 mg/L |
| Average time of a structured diagnosis | 5-10 working days |
Frequently asked questions about biogas production drops
Why has biogas production dropped at my plant?
In most cases because of one of these three: an uncharacterised diet change (the substrate has fewer volatile solids than the ones being used in the calculation), an overload of real organic load versus nominal, or free ammonia inhibition.
The other four usual causes are dissolved sulphide inhibition, loss of digester temperature, entry of antibiotics or disinfectants with the slurry, and inhibition by long-chain fatty acids. The key is to rule them out in order, starting with the cheapest to check.
How is a production drop diagnosed quickly?
With a fixed sequence. First the temperature probe is checked against an independent thermometer, which is free. Then the total and volatile solids of the current batch are characterised in triplicate and the real organic load is recalculated, which resolves the two most frequent causes.
Next, FOS/TAC and individual VFA are measured, which separate a load problem from an inhibition one, and only then do you go to the specific analyses: free NH₃, sulphides and LCFA. With that sequence the root cause is reached in 48-72 hours.
What are the most frequent causes of production loss?
The ones related to the substrate, not the biology. The uncharacterised diet change and the organic load overload explain most cases, and both have the same origin: the load is being calculated with a volatile solids figure that no longer corresponds to the batch coming in.
It is a problem of analytical method, not of process. That is why a plant that characterises its substrates every quarter has far fewer episodes than one that does it once a year.
How long does the digester take to stabilise after the intervention?
It depends on the cause. If it was load, the recovery shows in 1-2 weeks after correcting the feeding. If it was ammonia or sulphide inhibition, the consortium needs between 3 and 6 weeks, because the inhibitor has to be brought down and the archaea given time to recover activity.
If it was LCFA inhibition, an acclimated consortium recovers in 5-15 days with no more intervention than easing off the fat. And if the origin was pharmacological, recovery is spontaneous in 2-3 weeks, but it will happen again as long as there is no warning protocol with the farm.
Smallops and structured diagnosis
The difference between a plant that recovers production in two weeks and one that drags the problem for six months is not in the technology: it is in having an order.
At Smallops we apply the full tree over 14 variables and deliver the exact location of the limiting factor, not a list of suspicions. And when the cause is a new substrate or a diet change, we validate it in the laboratory with a semi-continuous assay before touching the plant: feeding a reactor for weeks with the real load reveals the kinetics and the tolerance that a batch BMP cannot show.
Is your plant producing less and you do not know why?
Request a Smallops Operational Excellence Diagnosis: we apply the diagnostic tree over 14 variables and tell you which of the seven causes is yours, with data and a correction plan.
References and standards
Chen, Y. et al. (2008). Inhibition of anaerobic digestion process: a review. Bioresource Technology, 99 (10), 4044-4064.
Rajagopal, R. et al. (2013). A critical review on inhibition of anaerobic digestion. Bioresource Technology, 143, 632-641.
Mussoline, W. et al. (2013). The anaerobic digestion of rice straw: a review. Critical Reviews in Environmental Science and Technology, 43, 895-915.
Drosg, B. (2013). Process monitoring in biogas plants. IEA Bioenergy Task 37.