part 1 – introduction to lumut power plant & flow assisted...

OMMI, Vol.4, Issue 2, August 2007 www.ommi.co.uk

MANAGING HRSG INSPECTION & REPAIR WORKS EFFECTIVELY

Rahimi Md. Sharip & Khairul Nizam Jasman, Lumut Power Plant, Teknik Janakuasa Sdn. Bhd., Malaysia Abstract Flow Assisted Corrosion or Flow Accelerated Corrosion (FAC), or sometimes also called Erosion Corrosion (EC), is a serious issue for Combined Cycle Power Plants, especially ones having HRSG configuration of dual pressure, vertical tube arrangement such as in the Lumut Power Plant where the existence of high velocity flow and the likelihood of attack can be further compounded if boiler water chemistry is not properly controlled. From a total of 6 HRSGs in the plant, 4 units have experienced FAC induced failures in year 2002 after operating for almost 6 years. This paper will outline the typical repair activities carried out on one of the units, the scheduling process and cooperation between the client and repair contractor which were very essential to ensure work proceeded smoothly, and actions taken to prevent the recurrence of FAC that eventually enabled us to maintain the Availability Factor (12 months rolling average of Equivalent Availability Factor) of the plant overall at above 92%, as targeted by the Company. Keywords: Combined cycle power plants, HRSG, corrosion Part 1 – Introduction to Lumut Power Plant & Flow Assisted Corrosion

http://www.ommi.co.uk

Managing HRSG inspection & repair works effectively, Rahimi Md. Sharip et. al.

2

1. Plant Introduction Lumut Power Plant (LPP) consists of 2 x 651.5MW combined-cycle blocks, namely Block 1 and 2. Each block has 3 units of ABB 13E2 model gas turbines, 3 units of ABB CE HRSGs, and 1 unit of ABB condensing steam turbine with seawater once-through cooling condenser. All HRSGs are exactly identical – they are unfired, bottom supported, dual pressure system with vertical arrangement tubes, and natural circulation type. In addition, the plant is designed for continuous daily base load operation, as well as partial load according to the requirements of the Malaysian grid system. The commercial operation date of Block 1 was back in July 1996, whereas Block 2 was one year later.

HRSG Data Design Case (GT Load 100%, Gas Fuel, TAmb 32 deg C, RH 85%)

Unit HP Steam LP Steam Mass Flow Kg/s 60.3 15.9 Pressure Bar abs 67.3 5.9

Temperature Deg C 508 Saturated

Table 1: HRSG Design Data 2. Flow Accelerated Corrosion (FAC) in LPP – Brief Introduction Flow Accelerated Corrosion (FAC) is one of the great concerns for HRSG users nowadays. It can lead to undesirable forced outages if factors contributing to FAC attack are not controlled properly, especially the feedwater chemistry. The main influence is the localized removal of the protective surface film (magnetite), which then leads to the accelerated corrosion of the base metal. Major parameters that influence FAC are: PH, turbulence (flow geometry and velocity), temperature, oxygen concentration and material composition. The following are brief explanations of the factors that may have contributed to FAC occurrences in the LPP HRSGs. 2.1 Temperature It is well known that with improper water chemistry and other combined factors, there is a certain temperature window which may make the solubility of the protective oxide layer (magnetite) become significant, consequently exposing the metal and making it vulnerable to erosion and corrosion. In the steam water system, it is also well documented that FAC is mainly observed in the operating temperature range between 80-230 deg C, with a maximum in the temperature range around 150-180 deg C. Indeed, all FAC induced failures that occurred within LPP HRSGs have been observed to fall within the mentioned operating temperature region. Table 2 provides a summary of tube failures due to FAC for the year 2002.


3

Unit Date of

Repair No. of Days

Location No. of Tubes

Replaced

Original Tube

Material & WT (mm)

New Tube Material & WT (mm)

HRSG 11 22nd Feb – 2nd March

9 HP Econ 4 4 SA178 C / 2.8 SA213 T22 / 3.5

HRSG 13 7th June – 15th June

9 HP Econ 2 8 SA178 C / 2.8 SA213 T22 / 3.5

HRSG 12 28th June – 15th July

18 HP Econ 1C & 2

32 SA178 C / 2.8 SA213 T22 / 3.5

HRSG 23* May (Scheduled

Outage)

~ 3 HP Econ 5 1 SA178 C / 2.8 SA213 T22 / 3.5

HRSG 23**

26th July – 31st July

6 HP Econ 1C

1 SA178 C / 2.8 SA213 T22 / 3.5

HRSG 21 19th Sept – 4th Oct

16 HP Econ 1C & 2

11 SA178 C / 2.8 SA213 T22 / 3.5

HRSG 13 (2nd failure)

25th Sept – 2nd Oct

8 HP Econ 2 7 SA178 C / 2.8 SA213 T22 / 3.5

Table 2: Summary of Tube Failures for Year 2002 HRSG 11,12,13 are in BLOCK 1. HRSG 21,22,23 are in BLOCK 2. * HRSG 23 tube replaced in May was due to weld defects (excessive penetration & root concavity) observed during inspection; it was not related to FAC. ** HRSG 23 tube leak in July was due to weld failure, and also not FAC related. 2.2 PH & Oxygen Concentration As mentioned previously, pH and oxygen play an important role in contributing to FAC attack inside the HRSG. Below is the feedwater chemistry guideline from the OEM that was followed by LPP since commissioning of the units.

Parameter Units Normal Operating value

Conductivity after Cation Exchanger µS/cm < 0.2 Specific conductivity µS/cm 3 – 11

PH value - 9.0 – 9.6 Silica ppb < 20 Iron ppb < 20

Copper ppb < 3 Oxygen ppb < 10

Table 3: Feedwater chemistry at HRSG inlet


4

In general, there were no problems with the pH value and other parameters for the LPP HRSGs. However, when FAC started to occur there was concern about the use of hydrazine when the limit of dissolved oxygen (DO2) concentration in the feedwater was very low. Because this was never stated directly by the OEM in the operating instruction manual from the beginning, as a result, the DO2 level for all LPP HRSGs was consistently kept by the operator at a very low level, which was 1 ppb on average, since commissioning. In fact, sometimes the DO2 level was being maintained at less than 1 ppb. As it is well known that hydrazine acts as a reducing agent (oxygen scavenger), and there is potential for magnetite dissolution if being used excessively, the dosage amount for HRSGs has been limited not only at many plants, but also now at LPP. 2.3 Turbulence Vertical configuration tubes provide the high velocity flow profile at the inlet headers, as shown in Figure 1 (indicating typical flow for LPP HRSGs tubes). Previously, with the magnetite being dissolved due to the presence of excessive hydrazine, the turbulent flow at the inlet subsequently increased the mass transfer of magnetite, and eventually impaired the tubes protection from erosion and corrosion.

Specifically, for LPP HRSGs, most of the tubes affected were the ones located at the inlet header nozzle areas where high velocity flow was expected and faster removal of protective layers on tube walls (compared to formation of the layers).


5

2.4 Material Composition As can be seen in Table 1, the original tube material installed was low carbon steel SA 178 Grade C, without the presence of chromium (Cr) and molybdenum (Mo) alloy elements. Both elements, Cr and Mo at least 2% and 1% respectively, are known to make material more resistant to FAC or EC. 3. Lesson Learned and Actions Taken to Mitigate FAC in LPP HRSGs Among the four main factors that influence FAC, obviously the variables that could be controlled are only water chemistry and material composition. Therefore, considering the best practical solutions recommended by the OEM and third party consultants, hydrazine dosing in LPP HRSGs has now been stopped completely. The DO2 level is being continuously maintained between 5 to 10 ppb by throttling the air evacuation valve at the feedwater tank. Total Organic Carbon (TOC) is also monitored in the make-up water produced from the demineralised water plant, in order to prevent any organic contamination that may suppress the pH value, due to presence of organic acids. Injection of Poly Alumina Chloride (PAC) into raw water is also being carried-out to reduce the TOC level. In addition, all tubes at the affected areas have been replaced with a better erosion-corrosion resistant material. Thus, SA 213 Grade T22 with Cr and Mo compositions of 2.25% and 1% respectively has been selected as the new tube material, and tubes with greater wall thickness are being used in order to mitigate the FAC occurrences (refer Table 1). Although, initially, one of the suggestions made was to reduce the flow rate and hence velocity at the header inlet nozzles (by changing the existing inlet nozzle design from 1 inlet to 2 inlets), nevertheless, this recommendation was not chosen due to the massive fabrication required and, possibly, long work duration and machine outages. Obviously, on the temperature profile, nothing much can be done since the HRSGs have been designed by the OEM based on the performance required. This is because changes to the operating conditions (such as temperature related parameters) on the HRSG-side could mean that changes in Gas and Steam Turbine operating conditions may become necessary as well. In summary, apart from the alterations on the water chemistry-side and replacement with better FAC resistant tube material, scheduled inspections in line with Gas Turbine planned major overhauls are now already in place. Internal inspections (as far as accessible) are being carried out using a newly purchased remote control robotic crawler, which allows more extensive coverage of tubes compared to normal practice by borescopic equipment (which was the equipment initially being used). Hopefully, through the inspections, by detecting the FAC symptoms at an early stage, or at least before a particular tube fails, immediate replacement work can be carried out during the ongoing planned outage, thus reducing forced outages and subsequently maintaining the Company’s targeted Availability Factor.


6

In fact, the first planned inspection has already taken place, and was carried out on HRSG 23 by an external third party inspection company. This was initiated after 3 FAC induced tube failures had occurred, one after another, in less than 5 months. Subsequently, a planned inspection of HRSG 21 (which will be discussed in the next part of this paper) was carried out by LPP’s own plant personnel using borescopic equipment (this was before the purchase of the robotic crawler).

Availability for year 2002

91.0

92.0

93.0

94.0

Dec Nov Oct Sept Aug Jul Jun May Apr Mar Feb Jan

Month

Ava

ilabi

lity

(%)

Figure 2: Availability Factor for Year 2002

(Note that the failures in June and October almost caused the availability to go down below 92%) Part 2 – Managing HRSG Inspection & Repair Works Effectively 1. Introduction This report concentrates on the borescopic inspection and tube replacement work conducted on HRSG 21 recently. Basically, this report will outline how all tasks were managed in order to achieve completion without any delay. Based on the number of forced outages for Block 1 HRSGs due to FAC problems, any inspection finding and immediate rectification work were considered critical in order to maintain the plant availability. This report will thus describe the whole process of HRSG 21 inspection and repair work, which indirectly will also illustrate the typical routine activities to be executed in any LPP HRSG repair work in case of tube leak failure.


7

2. HRSG 21 – Preliminary Preparation Due to the FAC problems experienced by Block 1 HRSGs, a condition-monitoring program, mainly in the form of borescopic inspection, has been carried out for Block 2 HRSGs. Taking the opportunity during GT 21 C-Inspection maintenance outage, HRSG 21 was the second unit to be inspected by plant personnel from TSG department. The first unit, which was HRSG 23, had been inspected by an external party during GT 23 C-Inspection in April 2002. From this first work experience, the scope of inspection work for HRSG 21 was reduced and concentrated only on the 6” opening inlet nozzle areas of HP Economizers 1 to 5. Based on the failure location pattern for Block 1 HRSGs, these areas were believed to be the potential ‘FAC–prone’ areas due to the existence of two main factors for FAC attack: temperature gradient and turbulent flow. Similarly as for other normal work, initial preparation entailed the application of Permit to Work (PTW) for the main work, attached with hot work and confined space entry permits, wherever and whenever applicable. Indeed, safety is paramount in any boiler tube repair such as this, since hot work in confined space represents the main percentage of the total work involved (as shown in Figure 1). As the contractor would work around the clock in 2 shifts (with 12 working hours during each shift), prior to start of work in every shift, renewal of hot work and confined space entry permits were required to be requested by the contractor supervisor, and to be issued by the plant Shift Charge Engineer on duty.

Fig. 1 - Portion of Work Scope

NDT

Jacking

Inspectionothers HOT WORK -

Cutting, Grinding &

Welding


8

To further ensure safety compliance, contractor personnel were also given a compulsory safety induction talk prior to work commencement. All electrical tools to be utilized were also required for inspection and approval by LPP Electrical Department. In addition, as part of local statutory regulations (Factories & Machinery Act 1967), notification and submission of detailed work procedures to the Department of Occupational Safety & Health (DOSH), had to be made by the plant owner and contractor, before starting any boiler repair work. For fast detection of “almost-failed tubes”, hydrotesting (which was also one of the NDT methods) was carried out for HRSG HP and LP circuits, once GT fast cooling stopped at approximately TAT 45 deg C. Besides confirming the integrity of pressure parts, especially on weld attachments, the main purpose of the hydrotest was to check the condition of nearly failed tubes (whether FAC induced or due to weld defects). This is because, during previous after-repair hydrotest for HRSG 12, tubes at other locations were also found leaking, which subsequently extended the forced outage duration. Hence, the draining process took over only after the hydrotest was completed. 2.2 Inspection Work The work began with oxy-acetylene cutting of all 6” SA106B nozzles that had already been initially identified. Cutting of the joints was necessary in order to provide inspection access for the 6 mm borescope camera (refer Figure 4). A total of 32 joints had to be cut off and this took about 12 hours for completion (refer to the 16 inlet nozzles at bottom header areas, highlighted in red in Figure 2 below).

Fig. 2: Location of 6” inlet nozzles at bottom header section (Top-view)


9

Fig. 3: Configuration of a 6” inlet nozzle and a 6” inter-connecting pipe for 1 harp row

Fig. 4: Tubes internal viewed from one of the 6” inlet nozzle openings

Each economizer header was also fabricated with a handhole. These handholes are located between the 6” interconnecting pipes and the bottom header partition plates. Depending on its position (whether near the inlet or outlet flow), a total of 8 handholes were selected and cut off to create additional inspection access.

Fig. 5: Tubes internal viewed from one of the handhole openings

6 “ inlet nozzles

6” inter-connecting pipe (between Side-A & Side-B harps)


10

Subsequently, the main task of inspection personnel was to manually manoeuvre the borescope flexible cable into the 6” openings, and visually inspect the tube internal condition (especially at each inlet side) on the borescope monitor. The coverage area of borescopic inspection was only as far as reachable by hand, and mostly concentrated about 4 rows of tubes from each side of the 6” openings (refer Figure 3). Nevertheless, these were the areas where turbulent flow was expected to be at the highest magnitude due to their close vicinity to the inlet flow, and hence the susceptibility to FAC would be higher than at other locations. Borescopic inspection through the handholes access, however, assisted the plant personnel to confirm the tubes condition at the inlet flow downstream areas for selected headers. 3. Inspection Findings The deterioration condition of tube internals was captured by still photographs from a connected printer. However, the evaluation of the deterioration level was solely based on ‘visual judgment’ from the monitor, photographs and discussions among related personnel. The only measurable indicator of failure rate that could be obtained was by checking the tube wall thickness through ultrasonic thickness gauge at the tube bend area, if headers are lifted to provide access (refer photo in Figure 10). However, the available thickness gauge transducer was unsuitable to be utilized due to its large tip diameter. This prevented a full flat contact surface, and was thus unable to give an accurate and non-fluctuating reading. Summary of the inspection is tabulated in Table 1 below. Harp 17A – HP Econ 1 High severity with at least 6 tubes showing FAC signs. Tube

replacement required Harp 18A – HP Econ 1 Medium severity with at least 4 tubes showing FAC signs.

Tube replacement still required Harp 16A – HP Econ 2 Medium severity (similar to Harp 18A). Tube replacement

required Harp 18B – HP Econ 1 Medium severity but FAC signs present at few tubes. Tube

replacement required Harp 16B – HP Econ 2 Medium severity (similar to Harp 18B). Tube replacement

required Harp 9B – HP Econ 3 Low severity with minimal FAC signs at few tubes.

Monitoring required (to re-inspect during next outage)

Table 1: Inspection Result Summary (Note: Harp A is on the left hand-side, whereas Harp B is on the right hand-side, if facing against hot gas flow direction. Each harp consists of 72 tubes. Further reference can be found in drawing “HRSG General Arrangement Pressure Part Section ‘D-D’, Maintenance Manual, H, Vol. 1”)


11

Fig. 6: FAC findings in one of the headers inspected, showed by still photographs

In summary, a FAC problem was present within HRSG 21. This fact was confirmed with the clearly visible signs such as ‘sand dunes’ and ‘horse shoe’ marks found prominently at the internal surface of the tube inlets (areas where tubes being inserted inside the header boreholes). Therefore, immediate corrective actions were necessary to be taken on the tubes that had been identified having the worst FAC signs or potential failures. Obviously, this could cause forced outages and affect plant availability should the affected tubes fail in future, if no action was taken. From the inspection results, there were 5 harps that were ‘fit’ for repair for tube replacement (refer Table 1). Due to time constraint, proper planning for the repair work was very important. This had to be arranged such that the progress would be in line with GT 21 overhaul and commissioning work. A decision was then made based on tube condition severity obtained from plant personnel evaluation, as well as considering the harps position and configuration. As a result, Harp 17A was chosen to be repaired together with Harps 16B and 18B. The reason was both harps were concluded to have more FAC damaged tubes compared with Harps 16A and 18A. This selection would then permit the repair work to proceed concurrently.


12

4. Repair Work The harp jacking method is the most effective way to repair tubes for HRSGs with vertically installed harps. In this particular case (for Side-A harps), 2 adjacent harps - 16A and 18A - were lifted to provide headroom (access) for cutting and welding works at Harp 17A.

Fig. 7: Jacking of Harps 16A & 18A to create access for Harp 17A repair work

Whereas (for Side-B harps) in order to perform tube replacement work at Harps 16B and 18B, three harps had to be jacked up, namely Harps 15B, 17B and 19B.


13

Fig. 8: Jacking of Harps 15B, 17B & 18B to create access for Harps 16B & 18B repair work Each harp weighs around 16t. Hence, the main equipments used in harp lifting were 2 hydraulic jacks with at least 10t capacity (but due to unavailability of tools, in this case, 20t capacity hydraulic jacks were utilized instead). Other main equipments were 1” and ½” mild steel shims to provide the shimming, and together with I-beam supports and bracket supports for supporting the hydraulic jacks during the lifting process. Fabricated saddles were also used between the flat jacks surface and the curved bottom half of the headers. In addition, 2 lever chain blocks with 5t capacity were also used in order to guide the upper harps as additional safety measures.


14

Fig. 9: Main equipments used for harps lifting – 20t hydraulic jacks & 5t lever chain blocks

The process to lift a harp would take around 3-4 hours when there were no stuck tube fins. When a header movement got stuck, sometimes the process took almost the whole day, since the specific header had to be lowered and lifted several times to free it from rubbing fins. Once completed, the lifted harp looked as shown in Figure 10. Note also the I-beams and shims that were used to provide support for the harp.

Fig. 10: I-beams, brackets and shims – used to support header in lifted position


15

As mentioned previously, borescopic inspection (through 6” inlet nozzle openings and handholes) was required to identify specific tube positions that needed to be marked, which eventually would be cut-off for replacement. In order to limit the usage of 240V electrical tools inside the confined space area, cutting of the marked tubes was done using pneumatic grinders as a safety precaution. Removal of failed tubes also took some time, since the portion of existing fins needed to be removed, as well as the remaining tube walls left inside the header boreholes. The header boreholes then also became extra access points once the identified defected tubes had been removed. Through these new 2” openings, another borescopic inspection was carried out again to confirm whether adjacent neighbouring tubes also required replacement. This indeed assisted the inspection personnel to reach further tube positions that were not reachable by borescope camera when accessing through the 6” inlet openings or handholes, initially. These tasks were repeated for each repaired harp. As mentioned earlier, hot work was the main activity for this repair work. Once all inspection had been completed (and total affected tubes finalized), the welding of new tubes then started continuously after surface preparation had been done. In general, the new tubes were butt-welded to the existing tubes at the upper joint, whereas fillet-welded to the header at the bottom joint (set-through connection – refer also Figure 11 below).

Fig. 11: Typical newly replaced tubes

The next task was to carry out the Non-Destructive Testing (NDT) on the completed welds. In this case, a Radiography Test (RT) was conducted on butt-weld joints, while a Penetrant Test (PT) was done on fillet weld joints. Basically, NDT time management, specifically RT, was very crucial and needed to be well planned in order not to disrupt the on-going overhaul work at the gas turbine side, as well as the activities of other operation personnel who were operating the


16

other running machines. In addition, the repair work for HRSG 13, which was having a forced outage during the same period (also due to FAC induced tube failures), was not to be disrupted. This was because the unit was required to be dispatched to the national grid, as soon as possible. Given restricted time of about 2-3 hours per day for the RT session, the repair task activities were prioritized and sometimes re-arranged, such that the whole work would make progress and continue smoothly. As soon as all NDT works on tubes were completed, preparation for Post Weld Heat Treatment (PWHT) then took place. PWHT was required as in compliance to the ASME Boiler & Pressure Vessel Code Section 1, due to the wall thickness of the headers, which was greater than ¾ inch. Normally, the PWHT machine could take up to 12 weld joints per session. For this case, PWHT consumed about 9 hours for completion, with the soaking temperature around 600-650 deg C. The next task after PWHT had been completed was lowering down the lifted harps. Once preparation had been made, the lowering process took place. Similarly as the lifting process, lowering each harp would take around 3-4 hours when there were no stuck tube fins. Subsequently, alignment and pipe fit-up for the main 6” nozzle joints were carried out. Note that surface preparation for these joints was already done earlier whenever time allowed, as the task was independent to other work. Once pipe alignment was satisfactory, welding of the 6” nozzle joints commenced. During previous repair work on other units, the 6” nozzle weld joints were tested by RT. However, upon further review of ASME BPV Code Section 1, Ultrasonic Test (UT) was agreed and selected for this repair work since it still conformed to the code requirement. This indeed had significantly reduced the duration of the overall repair work. At all stages, the repair contractor supervisor and the client’s person in-charge played an important role in planning and managing the sequences of work. During certain times, fast decisions were required to be made so that there would be no delay or disruption to the schedule. 5. Hydro-Testing As part of local statutory regulations, after any repair (or hot work) hydrotesting is required to be done. Hydrotesting was basically done in 2 stages. The first stage was after PWHT had been completed. Usually, hydrotesting would be performed for a single harp, unless the harps involved in repair work are in line together (for example, Harps 16A & 16B). However, for this particular case, since all harps were not in the same rows, hydrotesting had to be done separately and individually. Fabricated blind flanges would be welded on the necessary openings of the selected harp. A hose would also be fitted to one of the blind flanges at the bottom header, to be used for filling with demineralised water. While another hose would be connected at one of the modified vents at the top header for overflow monitoring and to release air locks. Once harp already full, the demineralised water supply point would then be replaced with an external pump at the bottom header, whereas the modified vent


17

would be installed with a local gauge to monitor the pressure readings. The harp was then pressurized until it reached about 90 barg, and the pressure was held for at least 20 minutes. After the hydrotest was successful with no leak observed, repair work was continued by lowering the harps (as described previously). The second stage of hydrotesting was done after all repair work had been completed. This was carried out using the system boiler feed pumps, once all 6” nozzle butt-weld joints were confirmed satisfactory by UT, and after all handhole fillet-welds and all drain and vent socket-weld joints were confirmed by PT to be satisfactory. The whole HP circuit was then tested at 90 barg, whereas the whole LP circuit was tested at 10.5 barg. Each pressure was held at least for 30 minutes, being witnessed by a DOSH Inspector, or at least by the plant 1st Grade Steam Engineer (on behalf of the DOSH Inspector if he could not be present). The machine was eventually handed over to the Operation Department, once satisfactory results had been achieved. 6. Conclusion The nature of repair work consists of a large proportion of ‘manual-driven work’, hence the vital influence to a successful repair work of an HRSG depends on good communication and cooperation between the repair contractor and the client. Both need to plan the sequence of work tasks carefully so that delays can be avoided and eventually meeting the objective of maintaining the plant’s availability. They also need to have close rapport with their sub-contractors, namely the NDT and PWHT services to ensure their work flows smoothly. The client has to monitor and ensure continuous work progress on meeting the deadline, but without tolerating the safety and health of workers, including other plant personnel.

Fig. 12: Typical “sand dunes” marks on tube wall attacked by FAC

part 1 – introduction to lumut power plant & flow assisted...

Documents