Food & beverage

Industries

Food & beverage

Raw material arrives and waits for lab approval. Every batch is traced, tested and approved. And some of it is still judged by a trained human panel on smell and taste — which is expensive, slow and not reproducible.

The problem, in a plant manager's words

At our ARTC demo session a visitor described their process exactly: a range of liquids with different pH, viscosity and colour; samples drawn and carried to the lab for approval before raw material can be released from the warehouse; the same again after mixing; every batch traced, tested and approved; and comparisons that depend on smell or taste, which need a trained panel and are human-dependent.

Two costs sit inside that. Material waits — sometimes days — while the lab works through the queue. And the judgements that a panel makes cannot be logged as a number, so they cannot be trended, audited or automated.

What a spectrum can and cannot replace

  • It can give you a reproducible number for composition, logged automatically and stored as an immutable record.
  • It can flag an incoming lot that differs from the one you approved last month, in seconds rather than days.
  • It can be calibrated against the panel: if the attribute correlates with composition, the model learns the relationship and scores it consistently.
  • It cannot taste anything. Where a sensory attribute has no compositional correlate, the panel stays — but it stops being the first line of screening.
  • Coloured and opaque liquids are not a problem for a transmission measurement at a 20 µm path; suspended solids need an inline filter of 15 µm.

Tell us the measurement you cannot make today

Describe the stream, the chemistry and the decision it blocks. Where it makes sense we test the technology on your own process samples before you commit.