Corrosion fatigue failure of tubes in water tube boilers

Introduction

Chemicals Industries Division of HSE is aware of three relatively recent failures associated with water tube boilers. Each of these appears to have been caused by corrosion fatigue of the outer water wall tubes. This alert, which is targeted in particular at the chemical industries sector, describes the incidents and outlines the action that should be taken by duty holders.

Note:  corrosion fatigue is caused by a combination of a corrosive mechanism and fatigue generated by mechanical or thermal stresses.

All three boilers were gas fired, of 'D' tube design (shown opposite) and around 30 years old with operating pressures at 80barg. In that time they have seen both continuous and intermittent service.

Background

Two of the three incidents resulted in a 'windowpane' type failure where a coupon of material was ejected from the boiler wall tube. These are shown in the photographs below. Such a failure can pose a very serious risk of injury, either from the coupon itself or from the resulting release of high pressure steam. In the third incident a longitudinal crack in the tube resulted in a release of steam within the boiler of sufficient energy to blow off a man-way door.

In each case the damaged tubes were positioned at the extreme ends of the boilers, close to the main framework. Two of the failures were in cold formed bends (one at a top bend and one at a bottom bend).  The third failure was in a straight section of tube.

Causes

Detailed inspection using video borescopes, angle probe UT and destructive testing revealed that corrosion fatigue cracking (though decreasing in magnitude) was evident up to around 12 tubes from the ends.

Incident investigations suggest that the more highly stressed parts of tubes (such as bends, or at other areas close to buckstays or supports) may be the most vulnerable. Additional thermal stresses could be imposed on the end tubes from the boiler frames.

It is believed that the precursor for the corrosion attack arises from unsatisfactory water-side controls during boiler outages rather than poor in-service water treatment (Reference 1). The corrosion penetrates the naturally forming magnetite layer forming pits. Defect morphology suggests that the fatigue cracking originates from the corrosion pits which then join to produce cracks.

The investigations revealed that although each of the boilers had an extensive inspection history (periodicities were 4 yearly and 3 yearly respectively) those inspections focussed predominantly on the steam and mud drums.  Only very limited inspections were undertaken on the tubes themselves, and these focussed predominantly on wall thinning caused by corrosion under insulation. Although some crack detection was referenced in the schemes of examination this did not take fatigue or corrosion fatigue in respect of the water wall tubes into account.

It also appears likely that a number of tube failures in the years leading up to these incidents could have been caused by corrosion fatigue but were wrongly attributed to failure due to overheating.

While it is doubtful that the condition of the manhole door fasteners associated with the third failure was contributory to the severity of this incident, the importance of properly establishing the condition and fitness for purpose of these, and other fastenings during outage inspections has been recognised.

At the time of writing it appears that videoscope inspection of tube internals plus angle-probe UT inspection from the outside are the only appropriate techniques. Research is underway to see if other techniques such as rotating eddy current methods (similar to IRIS) can be utilised.

The HSE investigations also revealed that those involved in carrying out technical reviews and risk-based integrity assessments, along with inspectors themselves, though experienced and skilled in conventional chemical process plant, were not specialists in boilers or fired heaters. 

Significantly, it became apparent also that corrosion fatigue, and indeed windowpane failure, in boilers, though not common, has been recognised in the power generation sector (Reference 2).

Recommendations

Duty holders should consider the following:

• whether they have water tube boilers which may be susceptible and pose a hazard to persons or other plant
• the scope of their schemes of examination and whether they include checks for fatigue and corrosion fatigue of tubes
• the adequacy of techniques employed to inspect tubes and how information on condition is recorded
• if they have formal acceptance/retirement and discard criteria for damaged tubes
• that schemes of examination recognise the importance of the condition of fastenings of spyholes, access doors, or other openings.
• whether those they employ to undertake boiler inspections or when carrying out technical reviews of written schemes are sufficiently experienced in this specialist field
• whether their water treatment and water-side protection of idle boilers is satisfactory. See Reference 1
• that, where applicable, they utilise the information given in guidance documents published by SAFed (Reference 3) or the American Petroleum Institute (Reference 4)
• that root causes of any tube failures are properly understood

References

BS 2486:1997 Treatment of water for steam boilers and water heaters' 
SaFed  'Pressure Systems – Guidelines on Periodicity of Examinations'
American Petroleum Institute – API RP 573 ' Inspection of Fired Boilers and Heaters'