An industrial pipe corrosion monitoring system is useful only when each measurement leads to a defined inspection or maintenance decision. Start with the piping circuit, plausible damage mechanism, inspection history and accountable engineer. Then choose sensors, locations and data paths. This guide explains the difference between wall-thickness monitoring and corrosion coupons or probes, followed by baselines, temperature compensation, alert review, procurement and acceptance testing. Any example sequence here is a planning illustration, not a prescribed inspection interval or a determination that a pipe is fit for service.
Corrosion monitoring and leak detection answer different questions
Leak detection looks for fluid already escaping or related pressure and flow abnormalities. Corrosion monitoring looks for evidence of metal loss or a change in the corrosive environment before a leak. A floor water sensor cannot tell you the remaining thickness at a pipe elbow. A stable reading at one thickness sensor cannot rule out damage elsewhere. If the project also needs leak detection, define it separately; our factory cooling-water leak detection guide covers that use case.
The output of corrosion monitoring is evidence for inspection planning, not automatic proof of safety. API describes API 570 as covering inspection, rating, repair and alteration of in-service piping, while RP 574 addresses inspection practices and the selection of inspection locations. Applicability depends on the piping and governing requirements. The plant’s designated inspection authority must decide what the readings imply for the particular circuit. A dashboard alert and a decision to continue operating are not equivalent. See API’s announcement on API 570 and RP 574; use the authorized standards themselves where they apply.
Define the piping circuit and damage mechanism first
Build a circuit register with service, material, diameter, design and operating pressure and temperature, nominal and previously measured thickness, coatings, insulation, fluid chemistry, maintenance history and access restrictions. Distinguish internal general thinning, local pitting, erosion, external corrosion and corrosion under insulation. One monitoring instrument rarely detects all of them. For example, a sensor on an exposed straight run cannot represent hidden damage beneath wet insulation at another location. A single-point average also cannot exclude a nearby local pit.
Rank candidate circuits by likely damage, consequences, poor visibility under existing inspections, access difficulty and exposure during manual work. Do not move the measurement point to an easy radio location without revisiting what the point represents. Check area classification, hot surfaces, fall protection and work permits against the actual device approvals and site rules. A manufacturer’s statement that a device measures thickness continuously refers to its installed measurement location, not the entire pipe. Periodic scans and other examinations may still be required.
Wall thickness versus coupons and probes
Ultrasonic testing measures the pipe wall at a selected location. Portable UT, permanently mounted UT and area mapping differ in coverage and workflow. A fixed sensor improves repeatability at one location and can reveal a changing trend between manual visits. Its suitability still depends on surface condition, coating, curvature, temperature and mounting. Emerson’s ultrasonic thickness monitoring catalog shows several device families; evaluate the exact model and piping condition rather than treating catalog capabilities as universal.
A corrosion coupon is a test specimen exposed to the process and later recovered to estimate corrosion behavior from its loss. An electrical-resistance probe tracks the resistance change of an exposed sensing element as material is lost. Both provide information about the process environment; neither directly measures how much wall remains at every point of the actual pipe. AMPP’s technical article explains coupons and ER probes. They can complement fixed UT: a probe may respond to a chemistry change before pipe-wall loss is visible, while a previously thinned pipe remains vulnerable even when current coupon results look mild.
When these signals disagree, inspect the context before calling one reading wrong. Check fluid samples, temperature, operational changes, coupon exposure, probe condition, sensor mounting and a repeat portable-UT measurement. Define that escalation in the procurement scope. A low coupon corrosion rate must never erase an established wall-thickness finding.
Select locations that change a decision
Review historical thickness readings and drawings alongside flow changes, elbows, branches, drains, low points, injection points, weld vicinity, material transitions and paths for external water ingress. These are candidates rather than a universal checklist; the relevant locations depend on fluid and damage mechanism. For each proposed point, ask: if this reading falls, which inspection or maintenance decision changes? The piping inspector should review representativeness and the need for circumferential or nearby scans.
Record the circuit and spool IDs, drawing reference, photograph, flow direction, mounting angle, sensor ID, surface preparation, insulation restoration, area classification and responsible team. “Under the elbow” is insufficient if a replacement crew could mount the sensor on a different side. A replaced spool should start a new asset history even if its tag stays the same. Keep rejected candidate points in the register with the reason, alternative inspection method and owner. Otherwise the risk quietly disappears from the project record.
A pilot should expose the design to both ordinary and difficult conditions: representative suspected and reference locations, accessible and difficult mounts, stable and changing temperatures. The pilot must not postpone necessary inspection of a high-risk pipe. Treat a location at which valid measurement cannot be obtained as a meaningful finding, not as a failure to hide.
Establish a baseline and handle temperature explicitly
Keep measurement variation distinct from metal loss
During baseline work, record repeated readings at the same spot, differences after removing and refitting the sensor, changes across operating temperatures and systematic differences from portable UT. Repeatability and closeness to actual wall thickness are different properties. Assess repeatability, reference agreement, temperature dependence and the rate of invalid readings separately. Do not headline a change comparable to measurement variation as confirmed corrosion.
Record product, flow, fluid temperature, cleaning, shutdown and restart states alongside each reading. A baseline is a reading plus its operating context, not one context-free number. If several points fluctuate together, investigate the shared gateway, clock, power, temperature and software before concluding that every pipe corroded at once. If one point changes, check mounting and surface condition and request a nearby portable-UT scan when needed. Run the site’s safety response in parallel whenever the indication may be serious; instrument diagnostics do not replace safety action.
Record an inspector-approved reference measurement at the same mapped location when installing a fixed sensor. Do not substitute nominal wall thickness for the baseline: manufacturing tolerance, prior loss and surface condition matter. Store time, operating condition, metal temperature, instrument, calibration record, technician, exact location and reasons for accepting or rejecting repeat readings. Avoid turning the first device reading into an unquestioned “zero corrosion rate.”
Temperature changes can shift ultrasonic measurement conditions. The Emerson ET210 data sheet describes a thermocouple and compensation for the device’s thickness readings. The capability and applicable temperature range are model-specific. Check the circuit’s minimum, typical and maximum conditions and startup changes against the selected equipment. Flag values outside validated conditions rather than plotting them as equally reliable observations.
Before treating an apparent drop as metal loss, examine temperature, signal quality, battery and communication status, mounting, coating changes and recent maintenance. This verification must not become an excuse to delay a serious safety response. Define who can request repeat inspection, restrict access, change operating conditions or stop equipment under site procedures. No generic number of minutes or millimeters can replace that circuit-specific engineering decision.
Transport the measurement and its quality together
Each record needs more than thickness: location and asset version, timestamp and time zone, metal temperature, device health, signal quality, communication state and maintenance events. Missing data is not zero wall thickness; estimated values should be visibly different from observations. Synchronize clocks so the order of process changes and thickness readings is reliable. Confirm that exported units match the dashboard units and that replaced sensors do not silently join incompatible data series.
An architecture may pass from sensor to gateway, an OT-side collection point and only then to authorized IT or cloud consumers. If monitoring requires no control action, keep interfaces read-only and preserve existing control behavior during a communication outage. NIST SP 800-82 Rev. 3 addresses OT security with performance, reliability and safety constraints, including segmentation. Map communications, user permissions, remote support, logs, backups and updates to the plant’s OT policy.
Long-term analysis also needs a durable tag and event model. Our industrial historian selection guide discusses collection intervals, retention and access. For corrosion monitoring, preserve spool replacements, sensor moves, reference readings and engineering approvals beside the trend. A polished graph without the original measurements and decision trail is poor evidence for future inspection.
Set alerts through engineering review
Separate device failure, missing observations, abrupt apparent change, longer-term trend, inspection rescheduling and an urgent operational response. A single red light for all these cases encourages alert fatigue. Minimum required thickness and fitness for service depend on the design, material, pressure, temperature, fluid, defect geometry, applicable standards and inspection history. ASME’s description of B31G identifies guidance for evaluating metal loss in pressurized pipelines and piping systems; it is not permission to apply one formula indiscriminately to every factory pipe.
First route data-quality alerts to instrumentation staff and credible metal-loss indications to the piping inspection authority. Link any decision to isolate or stop a circuit to the site’s approved procedure. Record receipt, acknowledgement, repeat measurement, engineering decision and corrective action. An automatically delivered email is not completed response. A simple slope calculated from two thickness points may be dominated by measurement uncertainty, temperature or a pipe replacement. State the assumptions and uncertainty of any projection. Do not use a dashboard’s “days remaining” estimate as an operating permit without an authorized assessment.
Check the real retrofit conditions
Non-intrusive installation does not mean zero site work. Confirm surface preparation, coating or insulation handling, curvature, mounting hardware, fall protection, replacement access, battery work and chemical exposure. Some products support measurement through specified external coatings; this does not mean all coating types or insulation thicknesses are acceptable. Ask the supplier for a model-specific applicability assessment for material, diameter, curvature, coating and temperature.
Steelwork, racks, tanks and moving equipment affect wireless communication. Survey candidate mounts and gateway positions during normal operation. Measure missing-data rates and recovery, not just signal strength. If radio coverage is poor at an important point, consider a relay, wired path, different technology or manual inspection rather than shifting to a low-risk location. Verify that monitoring cannot disrupt existing control or protective systems.
Work on operating piping follows the plant’s permit and safety process. Hot surfaces, pressure, chemicals, confined access and hazardous atmospheres require local controls. If insulation is opened, restore its weather protection and document the work; a poor repair could introduce external corrosion. Record both installation and future removal conditions.
Write an RFP around deliverables and accountability
The RFP should state the target circuits, damage hypotheses, intended decisions and relationship to existing inspections. Request a location rationale, instrument operating limits, accuracy and repeatability conditions, mounting drawings, temperature compensation, reference-measurement procedure, data schema, communications, OT controls, maintenance, local support and an acceptance plan. Verify each numeric performance claim against the selected product’s formal data sheet and actual installation conditions.
Distinguish responsibility for measurement quality from responsibility for piping integrity. The system supplier can deliver measurement and transport; the designated inspector or engineering authority determines what a result means for inspection and operation. Define how unsuitable candidate sites are replaced, how insulation is restored, how unavailable data is handled and how histories are exported at contract end. Retain the reasons for rejected measurement points and their alternative examination plans.
Compare total cost rather than sensor price alone: survey, scaffolding, permits, surface preparation, insulation work, gateway, OT changes, storage, licensing, repeat inspections, battery or device replacement and local support. More points are not always better if the inspection team cannot act on alerts. Include exclusions and recurring costs in the same comparison. Specify a manual fallback for network, gateway or supplier outages so the existing inspection program continues.
Use FAT and SAT to test the decision chain
Factory acceptance testing should verify values, temperature, quality flags, time, IDs, alarms and export with representative inputs. Simulate loss and recovery of communications, duplicate or stale data, time drift, power interruption, sensor replacement and user permissions. A correct number displayed under the wrong pipe tag is a dangerous failure. Link each test to a requirement, expected and actual result, evidence file, owner and retest.
Site acceptance testing checks mounting position, reference UT comparison, insulation restoration, operating temperature, wireless performance and alert delivery to the people who will respond. Do not create an unsafe process condition to prove an alarm. Use test inputs to verify routing and keep test data distinct from the production record. After testing, confirm production thresholds, recipients and permissions. Classify each point as accepted, accepted with constraints or rejected with an alternative inspection method.
Train the technician who checks measurement quality, the inspector who orders repeat examination and the engineer who approves a piping decision. Verify that records can be exported with point IDs and asset history when software or vendor changes. An observation period after SAT can validate communications and workflow; it does not prove long-term corrosion behavior.
Phase the rollout and define stop conditions
Move from circuit screening and damage review to field survey, reference thickness, pilot, design revision and staged expansion. Before the pilot, define what better decision it must enable: faster inspection prioritization, useful repeat-UT requests or workable after-hours escalation. Choose point count and duration for the circuit’s operating cycle; a short technology trial cannot prove long-term integrity.
Pause or redesign if mounting is unsafe, the baseline disagreement cannot be explained, valid data is unavailable at required temperatures, alarms overwhelm staff, inspectors reject point representativeness or the OT connection conflicts with protective design. A clear finding that the device cannot answer the question is valuable. Reassess baselines and alerts after spool, material, chemistry, process or sensor changes. Software updates must preserve the meaning of units, compensation and timestamps across historic data.
For each alert, retain one case record from operator receipt through instrument-quality review, field thickness remeasurement and approval by the piping authority. Name deputies and contact routes so a shift change or holiday does not break the response. Closing a case should require the cause, repeat finding, action and preventive step, rather than only a return to the previous displayed value. The record lets the next team review why the earlier decision was made.
Summary
A factory pipe corrosion monitoring system should connect fixed wall-thickness trends, environmental evidence from coupons or probes and repeat inspection under a clear engineering authority. One sensor and one predictive graph cannot establish integrity for an entire circuit. The practical design work is choosing representative points, trustworthy baselines, temperature and quality handling, OT data flow, reviewed alerts, safe installation and documented FAT/SAT.
If you are still selecting circuits, evaluating sensor approaches or drafting an RFP, you can discuss the scope with TOMAS TECH. We can work with your inspection team to fit monitoring around the plant’s existing piping program.
Frequently asked questions
Does a corrosion monitoring system replace periodic UT inspection?
No automatic replacement follows from installing fixed sensors. They monitor their installed locations and may miss other circumferential or nearby damage. The piping authority should decide how the observations complement, rather than erase, required examinations.
Should I choose a coupon or an ultrasonic thickness sensor?
Use thickness measurement to follow the actual pipe wall at selected points. Use coupons or probes to understand the process environment’s corrosive behavior. Combining them may help explain disagreements and guide repeat inspection.
What wall-thickness alarm value should we enter?
There is no universal millimeter value. A qualified engineering review must account for design, material, loading, temperature, fluid, defect shape, applicable code and history. A device-quality warning is a different threshold from a decision about continued operation.
Can a sensor read through paint or insulation?
It depends on the exact device and installation. The ET210 documentation describes measurement through certain external coatings, but that must not be generalized to every coating or insulated pipe. Check the specified configuration and compare it with a reference field measurement.
What should be included in the cost comparison?
Include site survey, access, installation, insulation restoration, gateway, networking, storage, license, repeat inspection, replacements and support. Keep capital and recurring items separate and compare suppliers on the same scope.
Should an abnormal reading automatically stop the line?
That is a safety-engineering decision, not a default setting for an information system. Verify data quality and define escalation against the existing protective design before any automation is considered.
