5 Things SyngaPure Measures Every Second, and Why Each One Matters

You cannot control what you cannot measure in real time. SyngaPure's sensor layer reads five signals every second and feeds them into one control model — so the cleaning train reacts before the operator has to.

Most gas cleaning trains run on separate controllers arguing with each other. The reformer holds its setpoint, the filter pulses on a timer, the sorbent bed swaps on a calendar. Tar spikes and the reformer does not know. Breakthrough starts and the operator finds out when the bed is spent. The five signals below are the ones SyngaPure closes that loop on.

1. Tar concentration (mg/Nm³)

The hardest signal in the whole system, and the most valuable. It is the difference between reacting to a breakthrough and preventing one. A typical spec limit sits around 10 mg/Nm³. If the measured trend bends upward four hours before it crosses the limit, the control layer raises the reformer temperature now — not when the alarm fires. Seeing the curve bend hours early turns a shutdown into a setpoint change.

2. H₂S and NH₃ load (ppm)

Sulfur and ammonia loads tell the control layer how much sorbent capacity is left before saturation. Maintenance stops being a date on a calendar. The bed is swapped when the bed is spent — not before, not after. Swap threshold around 85% capacity used, and the trigger is condition, not schedule.

3. Differential pressure across the filter (mbar)

Rising pressure drop across the filter is the earliest honest warning of blinding. It is the cheapest signal to read, and the most expensive one to ignore. Left alone, the curve crosses the trip point and the plant shuts down. If the cycle is shortened at 62 hours, the line continues. The filter tells you it is blinding long before it fails. Acting on that line is the cheapest maintenance there is.

4. Reformer bed temperature profile (°C across zones)

Bed temperature is the lever that keeps tar conversion stable when the feedstock gets wetter or heavier. It has to move before the tar does. The conversion window is roughly 820–900 °C across zones. A wetter batch is coming up the line; the bed is already +18 °C hotter by the time it arrives. The setpoint shift is the pre-emptive action. Wait for the tar to climb and the conversion window closes with it.

5. The control layer itself

All five signals feed one locally installed control model that adjusts catalyst temperature, sorbent flow and filter cycles together — not one at a time. One decision across the whole train, every second — not three controllers arguing with each other. No cloud dependency. No latency. No data leaving site. The model runs on the hardware in the cabinet.

Why each one matters

Tar is the spec the gas must meet. H₂S and NH₃ are the sorbent's remaining life. ΔP is the filter's remaining life. Bed temperature is the converter's remaining margin. The control layer is what makes the four sensors act together instead of in sequence. If one of these signals is missing, the other four lose most of their value. A tar reading without a sorbent reading does not tell you when to swap. A ΔP reading without a control layer does not tell you when to shorten the cycle. The signals are not a dashboard. They are inputs.

The honest caveat

Bench validation of the full five-signal stack is still running. Chemistry is proven; industrial validation of the integrated layer is next. We are not pretending otherwise.

The question for operators

Which signal would you trust least in your current plant, and why? Mean time between interventions is the number nobody publishes. The five signals above are the ones we are designing against.