Inside the HRSG: The High‑stakes Art of Finding Tube Failures Before They Find You

Think the Heat Recovery Steam Generator (HRSG) — the boiler that turns gas‑turbine exhaust into steam in a combined‑cycle gas turbine plant (CCGT) — leads an easy life? The data says otherwise. Modern multi‑pressure, often reheat HRSGs are pushed hard in flexible, cycling duty, not just baseload, and they accumulate damage in ways that surprise operators (tetra-eng.com) (tetra-eng.com).

Hot-end superheaters and reheaters are the bullseye, accounting for nearly 50% of tube failures; creep and fatigue drive ≈40% of failures, while multiple corrosion modes add ≈23% — and flow‑accelerated corrosion (FAC) sits near 8% (tetra-eng.com) (tetra-eng.com).

The cost of neglect is real: tube leaks trigger forced shutdowns and lost megawatts (tetra-eng.com). At CCPP Belawan in Indonesia, a 28‑year‑old HRSG’s thinned tubes caused recurring leaks that forced the steam turbine to derate by ~50%, slashing availability (researchgate.net).

Well‑planned preventive programs work. Industry reviews put the penalty for poor maintenance at 5–20% lost capacity, while digital/predictive approaches can lift availability by ~5–15% and cut maintenance spend by ~18–25% (ncbi.nlm.nih.gov). GE adds that proactive HRSG upkeep reduces risk and improves reliability and availability — by catching corrosion, fouling, fatigue, and creep early (gevernova.com) (gevernova.com).

Damage mechanisms and hot‑end exposure

Tetra Engineering’s surveys show the dominant mechanisms and where they strike (Figure 1): high‑temperature creep/fatigue at ≈40% of tube failures — concentrated in superheater/reheater bundles and nearby headers under thermal cycling and high pressure — and corrosion at ≈23% (tetra-eng.com) (tetra-eng.com). Corrosion includes under‑deposit attack in economizers and stress‑corrosion cracking at welds or lap joints; ammonium bisulfate deposits (from SCR exhaust) and iron oxide are common culprits.

FAC (~8%) thins low‑alloy steels on the steam/water side under high velocity, but after years of mitigation it’s a smaller share than often assumed (tetra-eng.com). Erosion and foreign object damage are rarer in gas‑fired units, though gas‑path fins and tubes can still suffer if particulates or condensate are present. Mechanical overload and vibration can fatigue tube‑to‑header welds.

The takeaway: hot‑end steam sections (HP/IP superheaters and reheaters) see nearly half of failures and warrant the closest surveillance (tetra-eng.com). No single inspection tool finds everything, so multi‑method NDT is mandatory.

Non‑destructive testing toolkit (NDT)

There is no one‑size‑fits‑all NDT method; selection depends on materials and defects (researchgate.net). Visual Testing (VT) via borescopes and cameras spots cracks, deposits, and distortion through hand‑holes and hatches, and even simple visual checks materially improve reliability (tetra-eng.com) (tetra-eng.com). Surface replica techniques preserve micro‑crack evidence for later microscopy.

Ultrasonic Testing (UT) measures remaining wall thickness in tubes, headers, drums, and piping. Phased‑array UT images welds; the Internal Rotary Inspection System (IRIS — an in‑service rotating ultrasonic probe using a liquid couplant) provides full‑circumference tube profiles that reveal localized pitting or erosion.

Eddy Current Testing (ECT) induces currents to flag wall loss and cracks, with Remote‑Field ECT (RFECT) suited to carbon‑steel tubing. Pulsed eddy current (PEC) systems such as Eddyfi Lyft can screen for inner‑diameter corrosion from the outside, even over fins or insulation, detecting sub‑2 mm defects and producing thickness maps with minimal disassembly (onestopndt.com). Operators use EC scanners to rapidly triage bundles.

Magnetic Particle (MT) and Penetrant Testing (PT) reveal hairline surface cracks on ferromagnetic and non‑ferrous components, respectively — a staple for header and tube‑end welds in shutdowns. Radiography (RT) via X‑ray/gamma images thick welds and attachments when a suspect joint appears.

Acoustic leak monitoring deploys microphones to detect the high‑frequency hiss of steam leaks in real time, with systems like Mistras AMS trending signals and alarming on threshold exceedance; field deployments have flagged reheater/superheater leaks minutes before any visible plume (ccj-online.com) (ccj-online.com).

Infrared thermography can expose hot‑spots on casings and ducting that hint at internal fouling or leaks. Laser profilometry and robotic crawlers carrying UT/EC probes assess tight bends, returns, and whole‑bundle bowing. Software now matters too: tools like Tetra’s PipeVue collate and trend borescope imagery across outages for condition assessment (tetra-eng.com).

Inspection cadence and outage planning

Routine internal inspections are standard practice. Visual checks occur every outage; pressure‑boundary examinations for drums and headers are often mandated by insurers or regulation — in the US, annual inspections are common (tetra-eng.com). Early‑life inspections establish a baseline, and cycling/two‑shift units merit more frequent looks because cycling amplifies damage (tetra-eng.com) (tetra-eng.com).

During planned outages, full internal inspections — visual, UT, EC — target economizers, evaporators, superheaters, reheaters, boiler drums, and headers. Cleaning improves access for NDT; professional chemical washes are often scheduled alongside an HRSG/boiler cleaning service to restore thermal performance.

Major overhauls every few years refresh the asset: boiler washes, retubing the worst sections, and replacing attemperators or valve packs. At Belawan, tube replacement lifted the reliability factor by +0.137 (0–1 scale) and availability by +0.016, with a financial analysis showing an IRR of ~7.5% and positive NPV for full retubing (researchgate.net).

Online monitoring and performance flags

Continuous monitoring between outages — vibration, temperatures, emissions, and acoustic leak sensors — catches drift early. A rising gas‑turbine back‑pressure or unexpected feedwater flow can point to economizer fouling and prompt targeted checks (gevernova.com). Analytics that track pressure‑part tube thinning let operators time superheater/reheater repairs just ahead of a planned outage, maintaining steam output and avoiding unplanned trips (gevernova.com) (ccj-online.com).

Chemistry control and deposit prevention

Water/steam chemistry is a front‑line defense. Deviations in dissolved solids or corrosion products can signal FAC or under‑deposit corrosion; drain monitoring is a useful early indicator. Operators commonly maintain feedwater pH with a neutralizing amine program to limit carbonic acid attack in condensate return.

Removing dissolved oxygen remains critical to prevent pitting; oxygen scavenging is typically part of the regimen, supported by an oxygen scavenger matched to operating pressure and temperature.

Stable alkalinity helps maintain protective magnetite layers; boiler chemists often standardize on an alkalinity control approach aligned to the plant’s AVT/OAVT philosophy.

Chemical precision matters, and metering is handled with an accurate dosing pump to avoid overfeed or underfeed that can accelerate corrosion or fouling.

Feedwater purity underpins all of this. Plants relying on demineralized makeup often deploy a demineralizer upstream of the HRSG to minimize dissolved solids that drive under‑deposit attack.

To push silica and ionic impurities even lower, many facilities add a mixed‑bed polisher to achieve [less than 20 ppb silica](https://beta.co.id/en/blog/the-ppb-that-trips-a-turbine-inside-hrsg-steam-purity-separation-and-realtime-alarms) and very low total dissolved solids.

Where continuous operation without chemical regeneration is preferred, a continuous EDI system produces ultra‑pure water and stabilizes conductivity at boiler‑grade levels.

Condensate return streams are routinely cleaned via a [condensate polisher](https://beta.co.id/en/blog/the-hidden-workhorse-keeping-hrsgs-clean-condensate-polishing) to remove corrosion products before they deposit on economizer and evaporator tubes.

Risk‑based inspection and record‑keeping

Risk‑Based Inspection (RBI) turns inspection data into decisions. Tracking thickness readings, weld indications, and defect locations over time allows targeted replacements just before critical thresholds, aided by digital logs and trending tools (onestopndt.com).

GE notes HRSGs are too often serviced only when availability issues emerge or at the annual, leaving them suboptimal for long periods; shifting to analytics‑driven inspections with smart sensors and AI keeps issues from maturing into outages (gevernova.com) (gevernova.com) (ncbi.nlm.nih.gov).

NDT execution best practices

Focus on high‑risk areas: gas‑path tubes and welds in superheater/reheater regions, and water/steam‑side high‑velocity circuits where FAC can bite; tube‑bundle faces on the gas side often show early stains or corrosion (tetra-eng.com).

Access and lighting drive outcomes. Scaffolds or manipulators that bring cameras/probes within roughly 60 cm of tube surfaces, plus high‑intensity illumination and mirrors, reduce blind spots (tetra-eng.com).

Combine casing and tube checks. External casing and stack corrosion can hint at internal conditions, and borescopes pay dividends on attemperators, burners, and exhaust outlets.

Use redundant methods to confirm anomalies: validate visual weld cracks with UT; corroborate UT thinning via IRIS or PEC; and keep a database of thickness/defect trends to plan replacements (onestopndt.com).

Qualification matters. ASNT‑caliber training and careful interpretation improve accuracy. Documenting findings and repairs builds a knowledge base for RBI and budget cases. Early detection enables targeted interventions — from cleaning or silicate treatment of feedwater to arrest economizer corrosion, to weld overlays or reinforcement on headers — turning a “hidden” HRSG into a monitored asset (tetra-eng.com) (gevernova.com).

Preventive maintenance payback

Reliability and availability move in the right direction under proactive programs: at Belawan, retubing lifted reliability by 0.137 and availability by 0.016, and industry reviews put digital maintenance gains at ~5–15% availability improvement (researchgate.net) (ncbi.nlm.nih.gov).

Unplanned outages fall when leaks are found faster. Acoustic systems have allowed operators to repair days sooner than conventional detection, avoiding hours of forced downtime (ccj-online.com).

Costs bend lower too: optimized maintenance trims spend by roughly 20% in aggregated analyses, and Belawan’s tube replacement — about IDR ~360 billion — delivered positive NPV and ~7.5% IRR in its feasibility case (ncbi.nlm.nih.gov) (researchgate.net) (researchgate.net).

Performance retention follows naturally: deposit removal and timely repairs avoid rises in back‑pressure on the gas turbine that shave megawatts and erode cycle efficiency; neglected HRSGs can lose several efficiency points, while cleaning and replacement put that capacity back on line (gevernova.com).

Safety and compliance are reinforced, too. In many jurisdictions (and likely in Indonesia under boiler safety regulations), periodic internal pressure‑part inspections are mandatory; preventive programs satisfy insurer and regulatory expectations and reduce catastrophic risk.

Industry trend lines and adoption

Risk‑based inspection models, condition monitoring, and asset‑management software are rising across combined‑cycle plants, with HRSGs seeing more online acoustic leak detection and fiber‑optic RBI (Figure 2). Analysts put digital maintenance gains at material availability and cost improvements in new projects (ncbi.nlm.nih.gov). As more aging HRSGs reach critical condition in Indonesia, systematic annual schedules and formal outage planning are expected to be emphasized.

Bottom line: a robust HRSG program blends fixed‑interval NDT, continuous monitoring, and timely repair. Plants that prioritize HRSG maintenance endure fewer forced outages and hold higher availability than those that do not (tetra-eng.com) (gevernova.com). As one practitioner put it, the need to maintain the HRSG “is often only realised in hindsight” after a tube failure trips the unit (tetra-eng.com).