[分享] yamal现场部分图片

原作者：吴兆蕾

Observations from the visit 16th January 2017
Site investigations were carried out with Mike Murphy (STL) and Sergey Medviev (STL).  Keith Proctor (Senior Field Engineer for Passive Protection), Robin Wade (Chartek Segment Manager), Lev Leviev (Key Account Manager AkzoNobel) and Valeriy Khramykh (Site Technical Service AkzoNobel)
Visual inspections were carried out on the following structures.  Discussions regarding the observations were only conducted for structures where COOEC, CPOE, SOE and PJOE were responsible.  There was an instruction not to discuss the observations from 715-PAU-001 from QMW.  There were two distinctly difference failure modes for the Passive Fire Protection and Cold Spill Protection systems.

图片:1.png

Passive Fire Protection
After a review of the findings by the joint investigation team, Mike Murphy (STL) summarised the findings as follows:
The majority (~90%) of defects on the modules were found on the webs of beams (Defect type A) – especially deep beams under modules – where surface cracking and signs of disbondment were observed (where in some cases the material had fallen off) (Figure 1).  There were numerous modules affected – from all yards which have similar arrangement.   The remaining (~10%) defects on modules were all localized and may be due to variety of different causes (Defect type B).

图片:2.png

Further observations
Cracking was observed propagating from the connections of secondary and primary structural steel where there had only been “point contact” connection and the gap filled with Chartek.  This was a common failure on all modules (Figure  2).  Generally the Chartek was tightly adhering around the cracks for these areas.

图片:3.png

Figure  2 – Cracking from infill on point connections

There were also signs of cracking from high stress points potentially due to movement.  It was not possible to rule this out being a contributory factor for the disbondment between the first and second layers of Chartek for the “internal” structures accessible from ground level (large plate girders) (Figure 3).  There was significant variation in the cracking observable at ground level between some modules from the same fabrication yard.

图片:4.png

Figure 3 – Potential cracking due to movement on structures at ground level
Where there was disbondment from the first layer of Chartek the base layer was glossy in appearance and it was not uncommon to observe a mirror image of markings made on the first coat of Chartek prior to the application of the second coat (Figure 4 LHS).  Generally flanges showed no or limited cracking whereas for the web areas cracking was more prevalent (Figure 4 RHS).  In general the main circular hollow sections (CHS)

used for columns showed no failure.  Cracking was apparent on some of the CHS legs where potentially undue stress had been applied.  For the 100-PAR modules, there was a distinct difference observed in the performance of the lower levels (level B) in comparison with high levels in the pipe racks (levels F and G).  For lower areas there was no visible sign of cracking observed however cracking was observed on the upper levels.

图片:5.png

Figure 4 – Glossy under-layer with mirror image of markings appearing on back side of Chartek LHS.  Cracking observable in the webs of beams RHS
For the top level of the pipe racks and SPP structures cracking tended to be hairline in nature particularly where the beams had a narrow web and it was difficult to control film thickness (Figure 5).

图片:6.png

Figure 5 – Fine hairline cracking on smaller beams / columns with narrow webs

There was recognition that Chartek 1709 system applied for the project had performed well on some of the structures.  This was particularly true of the BOG compressor (Russian scope of work) (Figure 6) as well as for the Siemens modules (Figure 7).

图片:7.png

Figure 6 – BOG compressor, 2 coat scheme, thickness 12mm

图片:8.png

Figure 7 – Siemens modules with passive fire protection thickness between 6-11mm
The general opinion was that the issues observed with the passive fire protection (PFP) are manage-able in comparison with the cold spill protection areas.  By the close of the week, the general view was that the passive fire protection quality alert would be downgraded from the observations during the site visit.  The expectation is that it would be possible to find solutions to enable commissioning dates to be met subject to agreement through the contract chain.


Cold Spill Protection
After joint investigation, Mike Murphy (STL) summarised the findings on the cold spill protection systems as follows and AkzoNobel are in general agreement with the findings:
Defect Type A
Widespread cracking was observed on the external faces of circular columns at all locations.  This was the case for the Main Cryogenic Heat Exchanger (MCHE) and Jetty modules.  Cracks penetrated the full depth with widths up to 2mm.  There were signs of disbondment from tap testing however all the material had remained in place (with the exception of samples taken for analysis) maybe due to the geometry of the structural steel

图片:9.png

Figure 8 - Typical observations for the CSP system applied to the MCHE and Jetty modules

Defect Type B
Widespread cracking was observed on the undersides of the beam flanges.  This is present on the MCHE Modules.  Leaving the material in its current condition could result in material disbonding and falling – This is a potential safety hazard and needs to be considered as a priority.

图片:10.png

Figure 9 - Typical observations for the CSP system applied to the MCHE and Jetty modules

Defect Type C
Some cracking on webs and insides of beam flanges.  This is present on MCHE and Jetty Modules as well as associated SSTs

图片:11.png

Figure 10 - Typical observations for the CSP system applied to the MCHE and Jetty modules
Further Observations
Significant snow and frost coverage as well as the polar winter hampered observations in relation to evaluating the total areas which needed to be repaired (Figure 11).  Estimates are required to be taken from the  MCHE modules where the surfaces were not covered to the same extent

图片:12.png

Figure 11 – Frost and snow cover on jetty module structures
No cracking had been observed during transport reported by STL although a thorough survey by AkzoNobel had not been possible at the dockside.
In general all the cryogenic protection system appeared tight to the substrate and the residual strength of the pieces which were not cracked appeared high even at low temperatures (figure 12).  From the markings on reference beams which were used to monitor the progression of the cracks, most of the cracks had marks associated with them suggesting that there were no new cracks appearing.

图片:13.png

Figure 12 – CSP system tightly “bound” to the steel.  No corrosion observed to date
Cracks had propagated through the primer layer upon temperature cycling.  Removal of samples revealed a cohesive split of the primer layer in all instances (figure 12).  Where pieces had been removed previously there were no signs of corrosion (several months).  The primer thickness remaining was sufficient to cover the peaks of the profile of the blast.  It would not be expected that this would remain the case once the temperature increases in the summer.  It appears possible to repair the cracks without pieces falling away from the substrate.
For the jetty modules, it was suggested that cracking on the supports, upon joint investigation for the modules, were likely to be due to structural loading.

图片:14.png

Figure 13 – CSP cracking on supports for the jetty module
Hairline fine cracks appearing in the SST structures are generally located at the web flange interface only where beam sizes are smaller and application control more difficult.

Further evaluation made by the Joint Investigation Team week commencing the 6th of February 2017
A further site visit was conducted for the remainder of the joint investigation team who could get access to Sabetta in January.
No further failure mechanismswere reported by the joint investigation team.  Samples were taken to evaluate the root cause of the cracking.  The following samples were taken from the structures:

图片:15.png

The following analysis is to be undertaken by the joint investigation team:
§ To carry out visual survey of defective areas
§ To select representative samples for chemical & physical testing
§ 3cm Ø & 8cm Ø core samples for
§ FTIR (Fourier Transformation Infra Red)
§ DSC (Differential Scanning Calorimetry)
§ TGA ( Thermal Gravimetry Analyiss)
§ SEM fractography (Scanning Electron Microscopy)
§ 25cm x 25cm panel sample
§ DBTT (Design Brittle Transition Temperature)
§ To fit strain gauges for monitoring crack development in live cracks
§ To commence root cause analysis
The temperature at Sabetta was lower than for the initial investigation (-35C as oppose to -20C in January). Cracks were reported to still be propagating in some areas whereas for others the crack propagation had ceased.

For the passive fire protection,
The situation for the Defect type A was essentially similar to the observations in January.  Some of the beams exhibited cracking on the flanges as well as the web which were not observable in January.  However the majority of the defects were still apparent on the webs of the beams.
A greater number of fine cracks were observable on small beams (Defect type B) between the flange and web interface.
For the cold spill protection system,
The maximum width of the cracks had increased to 5mm due to further shrinkage upon cooling with respect to the steel work.  It would be expected that when temperatures increased the crack width at the temperatures observed in January would be similar.  The material shows no sign of falling off the steel which consistent with the observations in January.
It was confirmed that there is a requirement to reinstate the PFP and CSP protection for the MCHE.
Repair Proposals
Initial thoughts on repair options for passive fire protection and cold spill protection were sent at an earlier date for comment after the visit to site in January  (Appendix 2)
These were not formally circulated at the wrap up meeting in February by the joint investigation team. Development of the repair procedure will occur week commencing the 13th of February following face to face meetings and feedback from the fabricators responsible for the application of the cold spill protection systems and passive fire protection systems for Yamal LNG.

Appendix 1 – Observations made on the 5th of December 2016
The following observations were made on the 5th of December after an inspection of the modules for (Train 1). According to the Painting and Fireproofing QC department, the module was safely delivered by sea to site and installed successfully. Between the 1st of November and the 14th of November, no visible defects were identified apart from one crack. During this time the ambient temperature range on site was between 0 and -5oC. Further cracking was reported to be found after the temperature dropped, reaching a lowest value of -34C.
AkzoNobel TSRs were at the construction site of the plant reviewing the extent of Chartek 1709 application required on the erection joints of the shelters and supports of the “Train 1” modules.  During inspection of the modules, AkzoNobel technical service representatives found that structural members showed cracks in the coating systems consisting of Intersheild 300, Intertherm 7050, Chartek 1709, Interthane 990.
The estimated quantity and length of the cracking were found to be in excess of 1200 linear meters.
An investigation is underway as to why cracking has occurred in the Intertherm 7050 / Chartek 1709 system (CSP system) on site (note: Chartek only areas have not been reported to show cracks through the system to the substrate – all show poor adhesion between layers of Chartek).  Senior technical field service personnel and technical authorities have applied for site access to further assess the structure prior to comment as to the potential root causes.
Painting system 11C (Intershield300+Chartek1709+Interthane990)

图片:16.png

Spot testing (tapping) has been carried out for areas showing cracking that voids have been observed between the Chartek layers.  Upon removal of the top layer of Chartek, the sub-layer of Chartek appears glossy in general.
CSP system ((Intershield 300+Intertherm7050+Chartek1709+Interthane 990)
Inspection of decks where the steel is protected against cold spill and hydrocarbon fire (CSP+PFP). Taking into account the complexity of the task and limiting factors (darkness, cold, lack of scaffolds) the inspection was carried out directly from the decks of the Module and therefore significant gaps in knowledge exist regarding cracking extent and crack propagation.
The inspection showed that almost all the columns have cracks in length from 0.5 m to 2.5 m. Some of the cracks are open with a width of 2-3 mm, clearly showing cracking down to the steel substrate. Spot tapping reveals the presence of some voids under the coating system. The location of the cracks are of a different character (vertical, horizontal, diagonal, combination), with cracks often passing through or linked to the areas of the coating applied around the supporting elements welded to the columns and beams (known to be highstress points). The total DFT (dry film thickness) of the system from the specification is from 30-44mm. When measuring the DFT of the coating in the area of cracks however a DFT up to 50-60mm could be measured in some areas.
The nature of the crackig on the beams is different when comparing with the columns. There are no “open” cracks and observations show surface cracking. In most cases they were located on the lower I-beam flange and on the webs.
Typical cracks are shown in the following pictures:

图片:17.png

泄密了吧

good

厉害了，有没有后续修补的方案

有没有原文地址