For cassava wastewater biogas, a covered lagoon is usually the safer choice when there is room for it. It is stable, simple to run and can take the mill’s fibrous filter-cake as well as the liquid. A UASB reactor is worth considering when land is tight, but it needs more skill to run. In a proposal for a 300 tonnes-per-day cassava starch mill in West Africa, a covered lagoon was put forward as the main option, with UASB and CSTR designs compared against it.
The challenge: a strong, variable waste stream
Starch extraction uses a lot of water. In this case the mill was expected to produce about 18 m³ of wastewater for every tonne of starch, which is roughly 5,400 m³ a day at 300 tonnes a day. The wastewater is warm (30 to 35°C) and rich in organic matter. Added to that, the mill generates around 300 tonnes a day of wet filter-cake, the fibrous root residue left after starch is separated.
The strength of the waste is measured as chemical oxygen demand (COD). Here the wastewater carried about 57 tonnes of COD a day, and the filter-cake carried a similar amount, about 61 tonnes. Together that is roughly 117 tonnes of COD a day. Dividing the wastewater COD by its volume gives roughly 10 kg of COD per cubic metre, which is why untreated cassava effluent is such a difficult discharge and such a good biogas feedstock.
The approach: three digester types compared
The proposal compared three anaerobic digestion technologies on the same mill. All three can work. The right one depends on land, skills and how much the plant will be pushed.
| Covered lagoon | CSTR (stirred tank) | UASB | |
|---|---|---|---|
| Footprint | Large | Medium | Small |
| Digestion rate | Slow, very stable | Moderate | Fast, sensitive to shocks |
| Biomass | Grows its own from local bacteria | Grows its own | Needs granular sludge, possibly imported |
| Gas yield | Highest, longest retention | Higher with biomass recycle | Lowest, some COD bypassed |
| Operator skill | Trained mill technician | Trained mill technician | Twice-daily volatile fatty acid tests, tight pH control |
| Alkali dosing | Not normally needed | Some | Often a major cost |
| Gas buffer | Six to twelve hours under the cover | Reactor roof or separate vessel | Separate gasholder usually needed |
Why the lagoon came first
The lagoon’s long retention time means even stubborn COD is converted, so it produces the most gas from a given feedstock. It also has a large biological inertia. If the feed doubles, the lagoon copes with little loss in performance. The proposal cited a comparable plant in South-East Asia whose flow rose from 400 to 700 tonnes a day without any change to the lagoon. If the lagoon is upset, the volume of biomass in it allows the process to recover.
Crucially, only the lagoon was judged able to use the mill’s full production of filter-cake over the long term. Adding the cake roughly doubles the biogas available, but it needs a sludge-recycle facility and a lamella plate separator to manage the fibres without disturbing digestion.
Where the other two fit
A UASB can suit a site with limited space, provided the cake and other fibres are taken out first. A CSTR suits sites where footprint matters or where the rainy season would stall lagoon construction, because its civil works are confined to a tank foundation.
The design that came out of it
The covered-lagoon design was sized around the following:
- Lagoon: a single lined and covered lagoon of about 80,000 m³.
- Gas output: about 62,000 m³ of biogas a day at 60 to 70% methane, which is roughly 37,000 m³ of methane a day.
- Power potential: up to about 6 MW of electricity if both wastewater and filter-cake are digested. The wastewater alone supports roughly half of that.
- pH control: a carbonic acid unit manages pH without chemical additions.
- Gas train: a two-vessel hydrogen sulphide bio-scrubber (2,000 ppmv in, 200 ppmv out), a gas dewatering unit, a booster delivering about 300 mbar, a final filter and a standby flare rated at 3,000 Nm³/h.
- Controls: a SCADA system with data logging, remote monitoring and alarm notification.
The gas engines were left out of scope and could be added later. The supplier recommended that the owner contract directly for the engines.
Why it works, and what it assumes
A covered lagoon trades footprint for stability. The slow, warm digestion tolerates the spikes that cassava mills produce, and it does not need a specialist chemist on shift. The proposal estimated about ten months from order to commissioning, depending on when the rainy season falls.
The design also rests on assumptions the owner has to confirm. Without a soil survey, it assumed there was enough space, no piling, good access and services at the plant boundary. Piling, utilities, buildings, lightning protection and import duties were all excluded. Acceptance was by a five-day proving trial in which the plant had to meet its process guarantees on five consecutive days.
What to consider for your own project
- Measure the flow and COD first. Everything else follows from litres per tonne of starch and kilograms of COD per day.
- Decide what to do with the filter-cake. Digesting it doubles the gas but adds equipment and handling.
- Check land and ground conditions. A lagoon needs a lot of space and a soil survey before the design is fixed.
- Be honest about operator skill. A UASB rewards a well-staffed plant. A lagoon forgives a thinly staffed one.
- Plan for the rainy season. Earthworks stall in the wet, so the programme should allow for it.
- Treat the gas properly. Biogas from any digester needs hydrogen sulphide removal and drying before an engine sees it. Guidance on explosive atmospheres under DSEAR is a useful starting point for hazardous-area design.
For a related look at a palm oil mill, see our post on sizing a covered lagoon for palm oil mill effluent.
FAQ
How much biogas can cassava wastewater produce?
In this design, about 62,000 m³ of biogas a day from a 300 tonnes-per-day starch mill when wastewater and filter-cake were both digested. The wastewater alone gives roughly half.
Is a covered lagoon better than a UASB for cassava?
It is better where land is available and operator skill is limited, because it is more stable and cheaper to run. A UASB is more compact but more sensitive.
Can cassava filter-cake go into the digester?
Yes, but it needs fibre management, such as sludge recycle and a lamella plate separator, so the fibres do not disturb digestion.
How long does a plant like this take to build?
The proposal allowed about ten months from order to commissioning, longer if the rainy season interferes.
Planning a biogas plant for a starch mill?
Tell us about your flow, COD and available land and we will outline which digester suits your site.