A steam methane reformer P&ID tags its instruments to ISA 5.1 like any other process drawing, so the reading skill transfers. What sets the sheet apart is the mix. Four things identify a hydrogen plant at a glance: a dense field of tube-skin temperature transmitters across the reformer radiant section, an analyzer count far beyond a typical process unit, an explicit steam-to-carbon ratio loop wired into the trip layer, and a PSA valve bank where every valve is on-off and driven by a sequencer. Around all of it sit hazardous-area callouts for the most severe gas group and a burner management layer with its own interlock logic.
This is a field guide to reading that drawing set: what each section of the SMR train carries, why the reformer dominates the tag count, where the safety layer sits, and how the whole thing lands when the tags become an I/O list.
The train, section by section
An SMR hydrogen plant is a chain of conversions, and each link carries a characteristic instrument signature. Once you know the signature, you can place an unlabelled sheet in the train within seconds.
| Section | What it does | Characteristic instrumentation |
|---|---|---|
| Feed gas compression | Raises feed to reforming pressure, typically 15 to 40 bar | PT and TT at suction and discharge of each stage, anti-surge FT (often Coriolis), ZT on the recycle valve, vibration and speed monitoring |
| Desulfurization | Hydrotreating plus ZnO beds hold sulfur below 0.1 ppm to protect the nickel catalyst | TT at bed inlet and outlet, PDT across each bed, H2S analyzer (AT) confirming no breakthrough |
| Pre-reformer | Adiabatic bed at 350 to 650 C cracks heavier hydrocarbons to methane, CO, and CO2 | Inlet and outlet TT plus multi-depth bed-profile TTs that track the reaction front as catalyst ages, PT, PDT |
| Primary reformer | Nickel-catalyst tubes in a fired radiant box, 700 to 1100 C | Tube-skin TTs by the dozen, pigtail and manifold TTs, radiant-box and flue-duct TTs, flue gas O2 and CO analyzers, fuel FT per burner group, combustion air FT, the steam-to-carbon ratio loop, draft PDT, outlet composition AT |
| Waste-heat recovery and steam drum | Recovers reformate heat into HP steam | Three-element drum level control (drum LT cascaded with steam FT and BFW FT), drum PT and TT, steam export FT |
| HT and LT shift | Converts CO with steam to CO2 and more hydrogen over Fe-Cr then Cu-Zn-Al catalyst | Inlet and outlet TT plus bed thermocouples, CO and CO2 analyzer at each converter outlet, PT, PDT |
| PSA | Purifies hydrogen to product spec, tail gas returns to the reformer as fuel | PT per adsorber, on-off switching valves with solenoids and ZSO/ZSC feedback, sequencer block, product purity AT, tail gas FT and H2 AT |
Blue hydrogen adds a CO2 removal train between the shift converters and the PSA: an amine absorber and stripper with lean and rich amine flow metering, overhead analyzers, then CO2 compression, dehydration, and custody metering on the CO2 export. That block reads like the amine unit on any gas plant, which is exactly the point. The distinctively hydrogen sections are the ones in the table.
The reformer radiant box, the most instrument-dense unit on the plant
No other process unit puts this many temperature tags on one sheet. The reason is that the tube-skin measurement does two jobs at once.
The reforming reaction is strongly endothermic and runs inside chromium-nickel alloy tubes hanging in a fired box. Wall temperature targets sit in roughly the 750 to 820 C band. Run the tubes hotter and conversion improves, but tube creep life falls off steeply with metal temperature. So the same Type K tube-skin thermocouple that feeds the firing control also feeds the tube-life record the metallurgists keep. That dual duty is why the radiant section carries skin TTs across rows of tubes, not one representative point. A worked example of the kind of tag you meet:
TT-1204Areads as temperature transmitter, loop 1204, position A, and on a reformer sheet a long run of consecutive loop numbers with position suffixes almost always marks the tube-skin field.AT-1310on the flue gas duct is the in-situ zirconia oxygen probe holding excess air at a low target for efficiency without risking substoichiometric firing.FFIC-1101is a flow ratio indicating controller, and on this sheet it is the steam-to-carbon loop.
The steam-to-carbon ratio loop deserves a moment, because it is the clearest example of control and safety interacting on one drawing. Steam flow and feed gas flow feed a ratio controller that holds the mix above roughly three moles of steam per mole of carbon. Below that, carbon lays down on the nickel catalyst and the tubes hot-spot. So underneath the ratio controller sits a low-low ratio trip, a safety instrumented function in its own right. Few other units draw a ratio computation directly into the shutdown logic, and seeing one is a strong hint you are looking at a reformer.
For the tag grammar itself, first letter for the measured variable and succeeding letters for the function, the base reference is the standard bubble-and-letters system covered in how to read P&ID instrument symbols. Everything on an SMR sheet obeys it.
Burner management and the reformer safety layer
The reformer is a fired heater, and it carries the interlock structure fired equipment demands. The burner management system appears on the P&ID as a cluster of trip-initiating instruments feeding a logic block, typically covering:
- high-high reformate outlet temperature
- low-low fuel gas pressure (and often high-high as well)
- flame failure, detected per burner or per burner group (flame detectors tag under first letter B in ISA 5.1, so expect BE elements and BSLL trip switches)
- high firebox pressure and loss of draft
- low-low steam-to-carbon ratio, as above
- loss of combustion air
The code landscape an engineer meets here: burner management practice draws on the fired-equipment codes, the NFPA 85 and NFPA 86 family in US practice, and CGA H-10, which addresses combustion safety for steam reformer operation specifically. The trip functions themselves are safety instrumented functions under IEC 61511, and reformer ESD functions commonly land at SIL 2. Treat that as a common LOPA outcome rather than a rule; the analysis is done per function, per plant.
What this means for reading the drawing is the same discipline that applies on any plant with a safety layer: the BMS and ESD instruments must be kept distinct from the basic process control loops when the tags are counted, because their signal handling, voting, and test requirements differ. The mechanics of that split on the I/O list are covered in SIL-rated I/O and BPCS separation. On an SMR the split matters more than usual because the safety-layer population is large: the BMS cluster, the ratio trip, and the hydrogen leak detection all sit on the SIS side.
The PSA, and what it does to the I/O count
The pressure swing adsorption unit purifies the shift effluent to product hydrogen, with the CO plus CO2 impurity spec typically held at or below roughly 10 ppmv, and it looks like nothing else on the plant. A row of identical adsorber vessels, each with its own pressure transmitter because the adsorption and depressurization steps are pressure-choreographed, connected by a lattice of switching valves. Every one of those valves is on-off. Full stroke, solenoid driven, position confirmed, commanded by a cycle sequencer that is usually drawn as a prominent logic block on the sheet.
The consequence shows up when you build the I/O list. Each switching valve contributes a discrete output for its solenoid and one or two discrete inputs for its open and closed limit switches, so a tag like XV-5203 on the drawing arrives on the list as a DO plus a ZSO-5203 and ZSC-5203 DI pair. Multiply by the valve count across the adsorber bank and the PSA singlehandedly skews the plant's signal mix toward discrete I/O. A typical process unit is dominated by analog inputs from transmitters. An SMR drawing set lands with an unusually heavy DI and DO block from the PSA, a large AI block from the reformer TT field and the analyzers, and a modest AO count from the modulating valves. If a signal-class summary off a hydrogen plant does not show that discrete bulge, something was missed. How each bubble maps to AI, AO, DI, or DO is covered in signal classes explained.
Hydrogen-specific measurement problems on the sheet
Beyond the train itself, hydrogen as a fluid forces choices that leave visible marks on the P&ID.
| Measurement problem | Why hydrogen makes it hard | What you see on the drawing |
|---|---|---|
| Leak detection | H2 is buoyant and disperses upward, and it is IEC group IIC / NEC group B, the most severe gas group | Point detectors (catalytic bead or thermal conductivity) placed at ceiling and roof high points, not at grade; detector tags in the SIS scope |
| Flame detection | A hydrogen flame is nearly invisible and produces no soot, so IR-biased hydrocarbon detectors miss it | UV or combined UV/IR flame detectors around the reformer and at flare and vent points |
| Area classification | Group IIC equipment requirements apply across the unit | Ex d and Ex ia callouts on instrument specifications throughout the sheet notes |
| Pressure measurement in H2 service | Hydrogen permeates and embrittles standard diaphragm materials over time | Materials callouts on pressure instruments, including gold-coated diaphragms on transmitter specifications |
| High-temperature points | Reformer wall and reformate temperatures sit far above ordinary thermowell duty | Type K tube-skin thermocouples, mineral-insulated elements, pigtail and manifold TTs with metallurgy notes |
| Composition and purity | Product spec is measured in ppm, and every conversion step needs its own confirmation | The analyzer chain: H2S after desulfurization, CH4 slip at reformer outlet, CO and CO2 after each shift, purity AT after the PSA, zirconia O2 in the flue gas |
| Custody and inventory metering | Hydrogen's low density punishes volumetric metering | Coriolis meters on feed, product, and CO2 export where mass or molar accounting matters |
The analyzer row is worth restating as its own tell. A typical process unit carries a handful of AT tags. An SMR carries them at every conversion boundary because each catalytic step is confirmed by composition, not inferred. When the tag census off a hydrogen plant is compared against other units, the AT density stands out immediately.
What the I/O list off an SMR set looks like
Pulling it together, the register an engineer builds from a hydrogen plant P&ID set has a recognizable shape:
- A large AI block from the reformer temperature field, the analyzer chain, and the ordinary transmitter population, with the tube-skin TTs often outnumbering every other instrument type on the reformer sheets.
- A discrete-heavy PSA block, DO for every switching valve solenoid and DI for every position switch, cycling under sequencer control.
- A clearly separated SIS population: the BMS trip instruments, the low-low steam-to-carbon function, hydrogen leak and flame detection, kept apart from BPCS loops with voting annotations preserved.
- Materials and area-classification metadata carried on the instrument index, because the H2-service callouts on the drawing are load-bearing for procurement, not decoration.
- On blue hydrogen, an amine and CO2 metering annex that reads like a gas-treatment unit appended to the train.
Building that register by hand across a full drawing set is the slow part of every hydrogen project's controls scope, and it repeats at every revision. For teams turning SMR and electrolyzer drawing sets into I/O lists and instrument registers, that extraction is the work Tagsight does for hydrogen plants.
Further reading
- How to read P&ID instrument symbols, the bubble and letter-code grammar every tag on these sheets follows.
- AI, AO, DI, DO signal classes, where the PSA's discrete skew and the reformer's analog bulk land on the list.
- SIL-rated I/O and BPCS separation, the discipline that keeps the BMS and leak-detection scope distinct from process control.