A process that never stops is steered by a measurement that never stops.
Digester, biomass dryer, recovery plant: the flow is continuous, the incoming material varies from one delivery to the next, and the decision is taken in minutes. The laboratory result is the reference; the in-line measurement serves the running of the plant.
On a digester the question is always the same: will today’s ration produce biogas, or acidify the digester? FOS/TAC, volatile fatty acids, ammonium: what near infrared informs, and what it does not reach.
These sites rarely have a chemometrics team, and the sensor lives outdoors or in a classified area. Hence the first question before buying: a measurement with a model, or without.
A probe on the digester feed? A measurement installed with no one on site to maintain its model. Before buying anything new, the diagnosis: often the hardware is sound, and it is everything else that is missing. What to do with an instrument that has fallen out of use.
Where to start, depending on your role
- You run the site: start from the decision you take on judgement today. Scoping.
- You follow the digester’s biology: what the measurement tells you about FOS/TAC and VFAs. Running a digester.
- You own safety: ATEX zone, admissible temperature and pressure. Safety.
- You decide the investment: with a model or without, two different economics. Measurement families.
- You want examples: eight real situations, presented without names. Experience.
Two families of measurement, two different economics
Moisture in divided solids and biomass
Chips, sludge, pellets, digester feedstock: the water you evaporate is energy you have bought. Measuring moisture on arrival and during drying stops you drying beyond what is needed, and lets you dose a feed on its real dry matter. What a moisture measurement actually covers.
Identifying a material on receipt
Checking a delivery as it arrives, recognising a material as the one declared, detecting something unwanted in an incoming stream. On a digester feedstock — silage, biowaste, effluent — the question goes further: will this load produce methane, and how much? Published work shows that methane potential is predicted better from the spectrum than from chemical composition. What can be known about a feedstock before it is fed.
To those add continuous composition monitoring in a liquid medium: the content of one constituent, how far a transformation has gone, the drift of a bath. It is a common case, and the one where the calibration burden decides the budget. Where the composition is known, finite and declarable, a pure-component model decomposes the spectrum and deploys quickly. Where it is broader, as with a digestate, an effluent or biomass, a model calibrated on samples does the work, with a sample campaign and reference values behind it. What each approach costs.
Where the measurement sits, operation by operation
| Unit operation | What is measured | What it changes |
|---|---|---|
| Receiving and characterising the feedstock | Dry matter, organic matter, methane potential, presence of something unwanted | Refusal or routing is decided on the weighbridge. What enters the digester becomes documented instead of assumed. |
| Silage taken from the face | Dry matter, from one day to the next | The ration is dosed on the day’s dry matter, not on the silo average. |
| Running the digester | Dry matter and ammonium measured; volatile fatty acids followed as a trend; FOS/TAC predicted, titration remaining the reference. The detail. | An imbalance that builds over a few days shows while it builds, not at the next sample. The laboratory confirms before the ration is touched. |
| Digestate out | Dry matter, organic matter, nitrogen | Agronomic value is read on the digestate actually spread. |
| Biomass drying | Residual moisture, incoming dry matter | Following moisture lets you stop at the right moment, and the margin you no longer pay for shows on every tonne. |
ATEX is a selection criterion, and it comes first. Some instrument families are certified for explosive atmospheres, with a documented protection method. On other families, and on in-line particle sizing instruments in particular, certification is worth confirming configuration by configuration. Asked at the start, the question steers the choice of technology and the budget is built on the right one.
It holds in three points. Which area classification applies at the exact point of measurement, which protection method exists for that instrument family, and what becomes of the certification if you add a mechanical adaptor or move the measurement away through a fibre. The third is the one worth writing down.
Where admissible temperature and pressure are settled
In this field, admissible temperatures and pressures are assembly characteristics, which depend on the sensor material, the seal and the geometry of the process connection. They are settled through a user requirement specification: a document that records the process, the material and the ranges of temperature, pressure and concentration, then calls for a traceable technical answer.
That is good practice, and the client’s best protection: a range is quoted for your assembly, at your point of measurement, and it is traceable to whoever wrote it. This specification is the first deliverable of scoping, before any trial.
The two questions that decide the project
ATEX, asked at the first meeting
Certification is available on some instrument families and in some configurations, and the list is short enough to check in five minutes. Asked at the first meeting, that question steers the choice of technology and the feasibility work follows a single route. One question at the start saves redoing the trials.
Designing for the place the sensor will actually live
A result obtained indoors, on a stabilised sample, is a first step. Three factors are added at the design stage to make it hold outdoors.
- Temperature, which shifts the absorption bands and biases a model built at ambient.
- Vibration and flow movement, which the mounting is designed to absorb.
- Fouling of the window, whose slow drift reads like model drift until you separate the two.
All three are known, and all three are handled by design: compensation, mounting, and a cleaning regime. What makes an in-line measurement hold.
Tell us the measurement point, its zone, and what you have to decide. We will tell you what holds.
Forty-five minutes is enough to know whether your need calls for a measurement with a model or without, what the area classification allows, and what the specification should say before you go out to tender. Your existing plant belongs in that conversation: what is already installed, who will look after the measurement, and how far the laboratory is.