Process Analytical Technology puts a spectroscopic probe (IR or
Raman) directly in the reaction stream so composition is measured
continuously instead of by pulling and analysing offline samples.
In closed-loop mode, the live reading feeds a controller that trims
the feed valve every cycle to cancel the error between measured and
target concentration — the same logic real PAT-driven reactors use.
error = setpoint − measured
valve = valve + Kp·error + Ki·∫error dt
measured = true_conc + sensor_noise
- Setpoint concentration — the target the controller tries to hold.
- Base feed rate — the underlying reagent feed before any control trim is applied.
- Sensor noise — how noisy the spectroscopic reading is; higher noise makes closed-loop control jitter.
- Closed / open loop — toggle the feedback controller on or off to see drift without correction.