A Thailand factory planning wireless steam trap monitoring should not buy sensors on the assumption that installing them automatically saves energy. Start with a verified trap register, definitions of leakage and blockage, the condensate return layout, radio coverage and the maintenance response after an alarm. Then compare acoustic and temperature measurement, ask vendors to answer the same RFP, and preserve evidence in factory and site acceptance tests. This guide is for maintenance, utilities and energy teams responsible for steam distribution and process heat.
The first decision: monitor both live steam loss and condensate backup
A steam trap discharges condensate and non-condensable gas while retaining useful steam. A trap failed open can pass live steam into the return system. A trap failed closed can back up condensate in a line or heat exchanger, impair heating and contribute to water hammer. The US Department of Energy’s Steam System Survey Guide explains that visual, acoustic and thermal assessment each has limits and that different failure modes have different consequences. An alarm is therefore a reason to investigate, not an automatically verified mass-flow loss.
Define the deliverable as an asset record that follows each trap ID from status and reason for concern through field inspection, repair and retest. Wireless data can help prioritize inaccessible or widely distributed traps. It does not eliminate competent field assessment under every operating condition.
| Topic | Decide before buying | Acceptance evidence |
|---|---|---|
| Fault classification | Definitions of leaking, blocked, healthy and indeterminate | Observe state changes with safe test inputs or known assets |
| Asset register | ID, model, duty, location, pressure and return circuit | Physical tag matches the screen |
| Radio network | Steel obstructions, height, outdoor conditions and gateways | Log missing readings and recovery in production |
| Maintenance | Alarm owner and site response | Work order and post-repair measurement are traceable |
| Benefit | Baseline fuel, steam and operating hours | Separate estimates from measured changes |
Failure modes a steam trap monitoring system must distinguish
Failed open: live steam passing
An open failure or leaking seat can allow live steam into the condensate return. This may increase fuel, water and chemical demand and change conditions at return vessels. Normal flash steam can look similar to live steam. DOE notes that visual inspection can struggle to distinguish them and that acoustic assessment depends on background noise and trap design. Treat a leak alarm as a prioritized field check. Plant-authorized personnel decide isolation, repair and return to service.
Failed closed: condensate retained
A blocked trap, clogged strainer, incorrect installation, insufficient differential pressure or return-line backpressure can restrict discharge. A flooded heat exchanger may deliver less heat; condensate in a header raises concern during startup. Low temperature alone does not prove blockage. A stopped process, light load or change in ambient temperature also changes surface temperature. Link each classification to operating mode, pressure, load and trap type, and retain the evidence used for the alarm.
Healthy also needs a context
Intermittent and continuous-discharge trap types have different sound and temperature patterns. Startup, steady operation and shutdown differ too. Do not apply one threshold to the whole population without grouping assets by duty, type, pressure band and operating mode. State when a device cannot classify an asset. Lowering sensitivity merely to make the alarm count look better defeats the purpose.
Steam trap sensor selection: temperature, acoustic or combined
Contact temperature is comparatively simple and can show inlet or outlet trends, but insulation, ambient conditions, mounting point, surface material and steam pressure matter. DOE specifically notes that a thermostatic trap may subcool condensate during correct operation, so temperature must be interpreted with pressure and the trap’s operating characteristic. One handheld temperature reading cannot certify every trap design.
Acoustic and ultrasonic measurement can reveal continuous passing, discharge cycles and change over time. Adjacent valves, pumps, pipe flow and structure-borne vibration may contaminate the signal. Do not convert sound level straight into leakage mass flow. Fix the mounting position and establish a baseline for each duty. DOE also cautions that flash steam and live steam can sound similar.
Combined temperature and acoustic data may provide complementary evidence, but two measurements are not a guarantee of correct classification. Yokogawa’s Thailand application note describes a battery-powered wireless device using acoustic and temperature sensing with a LoRaWAN gateway and monitoring platform for leaks and blockages. Armstrong describes temperature and ultrasonic fluctuation measurement in its AIM and SteamEye products. IMI’s November 2025 announcement describes its NEON Sonic acoustic and temperature sensor. These are manufacturer descriptions of capabilities; accuracy and economic effect at your site require an acceptance test.
| Method | Useful for | Main limitation | RFP question |
|---|---|---|---|
| Manual survey | Asset register, repair checks, small populations | Changes between rounds | What skills and record format are required? |
| Contact temperature | Temperature trends and possible blockage | Pressure, insulation and ambient effects | Where is it mounted and how is the baseline set? |
| Acoustic/ultrasonic | Passing, cycling and condition changes | Noise, neighboring assets and flash steam | How are false alarms tested on site? |
| Temperature plus acoustic | Prioritization using two types of evidence | More installation and analysis conditions | How is an indeterminate result explained? |
| Wired or integrated | Critical assets where wiring is practical | Construction and shutdown coordination | Are cabinets, cabling and upkeep priced? |
The Spirax Sarco UK Product Handbook 2026 lists the Spiratec R16C wired monitor for up to 16 traps. It is an example of a wired option appearing in a 2026 handbook; the listing does not mean it is a newly launched wireless product. Compare wired and wireless alternatives over the same scope: installation, power, communications, integration and support.
Survey the conditions of a Thailand factory first
Match the register to the plant
Walk down the P&ID and the physical trap. An installed model may have changed while a drawing did not. Record the trap ID, equipment, supply and return pressure, type, size, orientation, insulation, indoor or outdoor location, hazardous-area designation, access restrictions, fault history and criticality. In a closed condensate return, discharge is hard to see, so note test points and safe inspection access.
A full rollout is not always the first step. Prioritize elevated traps, hazardous locations, remote buildings and heat exchangers where a fault affects production. If inspection finds a misapplied trap or blocked return piping, correct the mechanical cause first. Monitoring does not fix a poor piping design.
Measure radio performance where the sensor will be installed
Steam piping, steelwork, tanks, cabinets and building materials affect propagation. A protocol name or catalogue range is not a site gateway design. Temporarily mount trial devices at candidate points. Record reception and missing samples under normal operation, shutdown and representative outdoor conditions. For hazardous locations, verify the exact product certification, the local installation rules and how batteries may be replaced. Certification depends on model, geography and installation details.
Trace the condensate return
Condensate may pass through a return header, receiver and pump before recovery. High return pressure can reduce differential pressure and destabilize discharge even in a functioning trap. Receiver vent steam can include normal flash steam and live steam from a failed trap. DOE treats excessive or changing vent steam as an investigation clue. When one sensor alarms, review other traps on the same return circuit and the receiver.
Architecture and data fields of a steam trap monitoring system
Specify the complete path: sensor, gateway, data store, alarm, work order, field check, repair and retest. Ask whether the node or gateway buffers data during an outage, how retransmission avoids duplicates and how timestamps are synchronized. A sensor can transmit correctly while the fault remains unresolved because nobody owns the work order.
At minimum, retain trap ID, sensor ID, gateway ID, measurement time and time zone, raw readings or features, state and reason, missing-data flag, battery and link condition, alarm acknowledgement, field result, repair and post-repair baseline. Raw-data availability differs by platform. Include export format, API, retention and data ownership in the RFP. A dashboard-only view is different from data the plant can retrieve at contract end.
Separate alerts for probable leakage, possible blockage, lost communication, low battery and indeterminate classification. Never display lost communication as healthy. Link an alarm to asset criticality, inspection history and response owner. Automatic valve action or a safety interlock should be outside a basic monitoring project unless separately engineered and assessed.
RFP clauses that make vendor offers comparable
Assign owners to each alarm stage
The response speed for the same leak candidate depends on who receives it first. For example, maintenance can acknowledge the signal, the utilities team can assess its effect on the steam circuit, and production can arrange a safe shutdown window. Specify night and holiday escalation. Broadcasting every minor change to everyone creates alert fatigue and can hide an important event. Give the alarm states such as received, acknowledged, field investigation, repair approval, action and retest. Record who handed the case to whom and when.
If the platform connects to CMMS or ERP, decide which system owns the asset register. The trap model can change while sensor ID remains, or a sensor can move to another trap ID. An outdated mapping attaches valid readings to the wrong asset. A successful API response is insufficient: in SAT, confirm that ID, timestamp, responsible person and action match across screens and reports.
Evaluate security and maintenance updates
A wireless node may stay installed for years. Cover key enrollment and revocation, disabling a lost device, gateway updates, time synchronization and administrator changes, not just initial encryption. For cloud data, request a diagram of transmission, storage, readers, access removal when staff leave, and contract-end export. Identify the connection between plant control and monitoring networks. The practical question is who detects missing, altered or misrouted trap condition data and which records allow recovery.
Give every bidder the same asset list and test conditions. “Detects leaks” is too vague for acceptance. Define the target, exceptions, communication, data integration, maintenance workflow, responsibility and tests. Site-specific numeric limits belong in the surveyed specification, not in a generic brochure.
| RFP item | Required answer and evidence |
|---|---|
| Scope | Applicability by trap ID, excluded duties and piping changes |
| Classification | Definitions and evidence for leak, blockage, healthy and uncertain; false-alarm evaluation |
| Measurement | Sensor position, interval, time accuracy, calibration and access to data |
| Communications | Survey method, gateway design, outage indication and retransmission |
| Power | Battery-life assumptions, safe replacement, full replacement cost |
| IT/OT | Connectivity diagram, authentication, encryption, roles, logs and update process |
| Workflow | Notification, work orders, inspection records and reset after repair |
| Data | CSV/API, retention, export and contract-exit procedure |
| Support | Thailand site survey, installation, training, service and spares |
| Acceptance | FAT/SAT cases, criteria, corrective actions and retests |
Score offers on site fit, explainable classification, communications, maintenance integration, total cost and local support. Treat mandatory hazardous-area and temperature ratings as gates, not scores to offset with price. If a bidder advertises AI, ask about training data, verified fault labels, reconfiguration after replacement and conditions that produce “uncertain.”
Steam trap FAT and SAT: two different acceptance stages
A factory acceptance test (FAT) checks a configured system under controlled conditions. A site acceptance test (SAT) checks the installed system in its real environment. Do not create dangerous live steam failures for a demonstration. Use safe test fixtures, recorded data, simulated vendor inputs and assets with known conditions. Isolation, bypass and startup follow the plant permit and safety procedure. FAT success does not prove radio coverage or classification amid site noise.
FAT checks
Verify sensor-to-trap ID mapping, time, transitions, notification, roles, CSV/API export, indication of power and communication loss, and data consistency after reconnection. Show healthy, likely leaking, likely blocked, indeterminate, low battery and missing data. Verify who can change thresholds and that the changes are logged. Retain test input, expected output, observed output, logs and sign-off, not screenshots alone.
SAT checks
Inspect mounting, neighboring assets and gateway reception. Measure under relevant operating modes and site noise. Confirm field investigation of an alarm and state after repair. Deliberately include representative trap types and difficult locations, not only the best signal points. For elevated and hazardous sites, check access permits, installation and safe battery replacement. Agree on an observation period for missing readings and recovery rather than accepting one successful transmission. Freeze pass/fail criteria in the RFP.
Design a proof of concept that can support rollout
An easy pipe run alone will conceal rollout problems. Stratify the PoC by trap type, pressure band, duty cycle, loud machinery, elevation and metal obstruction. Compare sensor classifications with competent conventional inspection. For discrepancies, investigate whether the reference inspection, mounting, operating mode or logic was responsible. Do not report a “detection rate” without a credible reference diagnosis.
Useful metrics include detection among verified faults, false alarms among verified healthy traps, indeterminate classifications, missing data, time to field check, time to repair, recurrence and inspection effort. State the denominator: assets, observed hours and confirmed cases. An idle asset must not quietly enter the healthy denominator. Report initial settings separately from tuned settings.
End the PoC with explicit adopt, redesign or reject decisions. Weak reception may require a gateway move or a wired option. Blockage false alarms may require operating-mode context. A trap type that the method cannot classify may need another method or exclusion. Before full deployment, confirm that the plant can own routine monitoring and repairs.
Separate estimated savings from measured benefit
Economics cannot be certified by multiplying a generic loss per trap by alarm count. The escaping mass depends on orifice, pressure difference, degree of failure and operating hours. DOE observes that accurate loss quantification for a failed trap can be difficult. Build a baseline with fuel price, boiler feedwater, measured steam, condensate return, hours, makeup water, treatment chemicals and inspection effort. Normalize before and after repair for output, product mix, season and operating pressure.
Include sensors, mounting, gateways, connectivity, platform fees, batteries, calibration, training, field service and integration in cost. Keep verified fuel savings, lower inspection effort and possible downtime avoidance as separate benefit categories. Avoid double-counting and label scenario values. Payback should use your plant’s baseline and measurements, not an unverified advertised figure.
Hold the comparison period constant
A simple month-before versus month-after comparison mixes production and weather changes into the result. Where possible, compare steam consumption per unit of output or equipment operating hour, separating startup and idle periods. Record trap replacement dates alongside valve and pressure changes. When multiple traps are repaired together, reconcile individual estimates against measured circuit-level changes. If meters are missing, publish a range with assumptions rather than a falsely precise value. Trap condition monitoring locates repair candidates; a steam or fuel meter tests what changed across the system.
For a broader utilities framework, see factory energy-saving measures in Thailand. Compressed-air leak detection also illustrates distributed-asset maintenance, but its thresholds and ROI must not be transferred to a steam trap: condensate discharge is a distinct function.
Close the loop from alarm to repair
Rollout order and handover criteria
After the PoC, stage expansion by duty or piping circuit so causes of false classification remain traceable. First confirm the asset register; then install gateways and priority sensors; then operate the full chain from alarm to work order, repair and retest. Start another area only after reviewing overdue alarms, missing data, false positives and maintenance workload in the previous one. A sensor count alone can conceal installed points that nobody monitors.
At handover, provide configuration files, trap-to-sensor mapping, gateway locations, radio survey records, classification-rule versions, access roles, battery procedures, spare parts and fault contacts as one set. A replacement team member must be able to reconstruct why a threshold changed. Vendor remote configuration changes need plant approval and a before-and-after check. Include how to revert to manual survey during system downtime.
Prioritization after a fault is found
Likelihood of leakage is only one dimension. Main steam headers, critical heat exchangers, possible water hammer, inaccessible elevated equipment and clusters of alarms on one return circuit must be assessed for operational impact and inspection difficulty. Escalate urgent cases through existing safety procedures. Plan ordinary repairs with parts and a suitable shutdown opportunity. Record confidence in the sensor classification, asset criticality and time needed for field confirmation separately; this makes the reason for each response explainable.
The useful result is not the number of alarms but the number of verified issues repaired and retested. Prioritize by asset criticality and evidence strength. At the site check, record trap type, pressures, sound, temperature, return line, strainer, bypass valve and nearby equipment. Choose replacement, cleaning or piping correction based on the confirmed cause.
After work, measure again under comparable conditions and confirm condensate discharge, not merely the disappearance of an alarm. Update the installed model and its baseline. Keep false alarms in the record, categorized as shutdown, noise, mounting, communications or logic, so the next adjustment has evidence. Review overdue alarms, repeat faults, missing data, battery status and failed exports monthly. The register should become the common record for inspection, monitoring and maintenance.
FAQ: introducing wireless steam trap monitoring
Does a steam trap monitoring system measure the leak rate directly?
Usually it classifies a condition from acoustic and temperature observations; it does not necessarily measure steam mass flow. If a platform estimates leakage, request its pressure, orifice, failure and operating-hour assumptions and uncertainty. Check the estimate against plant steam and fuel data. A displayed currency value is not a verified saving.
Can temperature alone select a steam trap sensor?
Temperature can be useful for trends in a limited, known duty. It is harder to interpret where traps intentionally subcool, insulation is heavy or processes frequently stop. If both leakage and blockage matter, compare acoustic-plus-temperature and field inspection in a site trial.
What determines wireless range at a Thailand factory?
On-site reception from proposed mounting points to gateways. Include steel pipework, tanks, walls, outdoor conditions and other radios. Log missing readings and reconnection. Also verify certification and local installation requirements for hazardous areas.
Is FAT or SAT alone enough?
They answer different questions. FAT tests function and integration in controlled conditions; SAT tests mounting, radio, site noise and operations. Agree on evidence and pass/fail criteria before purchase. Use safe simulation when producing an actual failure would be unsafe.
Should every trap be instrumented from day one?
Start with a reliable register and a PoC covering critical, elevated, hazardous and hard-to-access duties. Expand after the communications and maintenance response work. The value lies in closing faults, not merely counting installed devices.
Summary
A credible wireless steam trap project covers both steam passing and condensate backup. It interprets acoustic and temperature data against trap type and operating mode, tests radio performance at the Thailand site, and makes alarms actionable through maintenance. A comparable RFP plus documented FAT and SAT turns a sensor proposal into an auditable deployment decision.
Even if the trap register is incomplete or you are still comparing measurement methods, contact TOMAS TECH to structure the site survey, radio trial, RFP and FAT/SAT around your plant’s operating procedures.
Primary sources
- US Department of Energy / Oak Ridge National Laboratory, Steam System Survey Guide, section 6.2.2 (published 2002; failure modes and survey limitations).
- Yokogawa Thailand, Monitoring Steam Trap Conditions (2025 application note).
- Yokogawa Thailand, wireless steam trap monitoring product page (current Thailand listing).
- Armstrong International, Steam System Monitoring (manufacturer account of AIM and SteamEye).
- IMI, NEON Sonic launch announcement (25 November 2025).
- Spirax Sarco, UK Product Handbook 2026, volume 2, p.23 (wired Spiratec R16C listing, not a claim of a new 2026 launch).
Product capabilities are manufacturer claims. Verify suitability and benefit with the exact specification, plant survey, FAT/SAT and operating data.
