Innovative Energy & Research
Open Access

The scope of this report is to analyze and investigate the frequent failure of UZ/ACPT Stainless Steel SIDUM Ejector Condenser. Recommendation is also drawn off for solving the problem based on technical and experienced data.

Steam circuit and process flow

Steam produced in the boiler passes through Boiler Valve 90- 11 before passing to the Hogging Ejector 45-01. The hogging ejector removes the air in Cell 2 and discharges to atmosphere. The ejector is only used during start-up of the evaporator. After start-up, the vacuum is maintained by two ejectors, whose motive force is steam produced by the boiler via Regulating Valve 90-01. The downstream ejector removes non-condensables to the first stage of the ejector condenser. The exhaust steam is condensed in the second stage ejector condenser and the noncondensables are vented to atmosphere. The upstream ejector removes air from Cell 2 of the evaporator. The exhaust steam is condensed in the ejector condenser. The condensed water from the second stage flows to the first stage of the ejector condenser via Isolation Valve 40-08 and a Trap 40-04. This trap can be bypassed if necessary. The common distilled water passes through Trap 40-07 and Isolating Valve 40-12. The trap can be bypassed to the heating bundle of Cell 1 of the evaporator. The thermo compressor steam is supplied via Isolation Valve 90-02. The thermo compressor removes the remaining vapour from the top of Cell 2, compresses it and discharges to the heating bundle, where the vapour is condensed releasing sufficient heat for the seawater to be vaporized. A temperature switch set at 100°C monitors the temperature at the outlet of the thermo compressor. If the temperature exceeds 100°C Steam Valve 90-01 closes and a High Temperature Alarm TAH-60-10 is generated. Figure 1 shows the Desalination Unit and steam path to the hogging ejector and ejector condenser.

Figure 1: ACPT Desalination Unit.

Ejector condenser

The ejectors are operated to maintain the vacuum in Cell 2. The ejector condenser consists of two stages. The downstream ejector removes non-condensables to the first stage of the ejector condenser. The exhaust steam is condensed in the second stage ejector condenser and the non-condensable are vented to atmosphere. Refer to Appendix 1 and 2 which shows the mechanical as built drawing of the ejector condenser (Figure 2).

Figure 2: ACPT SIDEM Ejector Condenser.

Ejector data

Ejector data is presented in the following Table 1.

Table 1: Ejector data.

Ejector condenser data

Ejector Condenser data is presented in the following Table 2.

Table 2: Ejector condenser data.

Frequent failure are taken place for the SS ejector condenser, upon visual examination, Crack like damages are commonly taken place on the 3” 2^nd stage Nossel, and near weld areas of the condenser associated nozzles. The Table 3 shows the Inspection history findings and recommendations. Inspection history shows that almost all the ejector condensers in SIDUM desalination Plant in ACPT has failed, by frequent leak reported, and claimed that chloride SCC is the damage mechanism of shell as detailed below.

Table 3: Inspection history findings and recommendations.

Chemical Composition analysis of 354B04C-Sidem Ejector Condenser as it is last inspected on 21/05/2013 (Table 4).

Table 4: Chemical Composition analysis of 354B04C.

Barometric type (direct connections) are usually employed. The components vulnerable to corrosion attack are: body, nozzles and piping. Body of the ejector condenser (CS cladded with SS, Ti, Cu- Ni) sometimes showing metal loss has now been replaced by Fiber Reinforced Plastic (FRP) in new plants. Nozzles (SS 304 or 316) and condenser piping (316L or Cupro-nickel) are most affected by pitting, the attack is more prominent at the welded seams. On the after condenser (steam/vapor inlet pipe) severe Cl- induced S.C.C. was noted which was attributed to high operating temperature and salt deposition. The 316 L pipe was replaced by Incoloy 825. Presence of 0.1-0.2 ppm chloride in the uncondensed gases is sufficient to produce corrosion. Cupro-nickel and SS are replaced by 254 SMO as an ejector condenser material in some of the plants in Qatar and Abu Dhabi due to their poor performance [1].

Welds in the 300-series austenitic stainless steels, contain a small amount of β-ferrite to prevent hot cracking during weld solidification. In hot, aqueous chloride environment (the same environment as for ejector condenser), these duplex weldments generally shows a marked resistance to cracking, while their counterparts crack readily. The generally accepted explanations for this behavior is that the ferrite phase is resistant to chloride SCC and impedes crack propagation through the austenite phase. Electrochemical effects may also play a part however, under sufficient tensile stress, temperature and chloride concentration, these duplex weldment will readily crack (Figure 3).

Figure 3: Chloride SCC of type 304 stainless steel base metal and type 308 weld metal in an aqueous chloride environment at 95°C (200 F). Cracks are branching and transgranular.

Methods for controlling chloride SCC

Chloride SCC to occur it requires a chloride concentration and tensile stress focused together to cause both intergranular and transgranular cracking. This produce weakening of the metal and eventual failure. Experience indicates that the concentration of chloride in the water contacting the stainless steel is not the critical factor [2]. The main factor in the existence of conditions that allow chloride concentration cells to develop in the absence of concentration cells or stress, chloride levels in excess of 1000 mg/L have not caused stainless steel to crack. Some desalination plants where chloride concentrations exceed 30000 mg/L have not experienced failures, when properly annealed stainless-steel construction was used. The key to preventing stress corrosion cracking is eliminating deposits and designing and fabricating stress-relieved 16 equipment that does not allow concentration cells to occur. 2 out of five methods of controlling CLSCC is addressed here in which are 1. Changing material 2. Changing design, the rest of five are not applicable in the case of Ejector condenser due to cost, operational constrain and ease of apply which are Change the Environment, Coating and electrochemical Techniques.

Change in the material: A common fix or solution of a stress corrosion cracking problem is to subtiture a more resistant alloy for the one that failed. N06600 (Alloy 600), a chromium-bearing nickel alloy, has greater resistance to cracking in high temperature chloride environments than does S30400 (AISI 304) Austenitic stainless steel, and is often used to replace stainless steel components that fail by chloride stress corrosion cracking [3].

It is important to identify the critical species causing the stress corrosion cracking. For example, N08800 (Alloy 800) is a suitable replacement for S30400 (AISI304 when chloride is the causative agent, but will fail as readily as the latter in hot caustic is the cracking agent, although it too can crack in that servise. Alloy 600 expansion joints are commonly used in chloride-and caustic-contamination steam service, despite occasional failures, because it can offer a favorable cost: Life benefit.

Changing in the design: The primary role of design changes for control of tresses ntal cracking is to lower the tensile tresses to below the threshold stress or to a level significantly less conductive to cracking. If cracking is to be alleviated. The total resultant stress from residual stresses. Thermal stresses and stresses from operational stresses, thermal stresses and stresses from operational loads abd oressyre must be considered.

Residual stresses from metal fabrication and unit construction are reduced by two principal methods:

a) A stress relieving heat treatment

b) Shot-peening, sometimes reffered to as mechanical stress relief.

In the first case, a fabricated item is heated to a temperature high enough for relaxation of the residual stresses, and slowly cooled. In the second case, the metal surface is mechanically peened at an intensity sufficient to result in residual compressive stress at the surface.

A second design consideration for control of environmental cracking is to avoid geometries in which solutions accumulate. Dead spaces where steam blanketing or water evaporation can occurs are potential failure sites. Solutions with chloride concentration of 1 ppm can become concentrated to high chloride levels in crevices and other restricted geometries.

A third design consideration is the compatibility of materials throughout the system. Contact of dissimilar metals can polarize one into the potential range for environmental cracking. Chlorides leached from insulation or formed by hydrolysis of organic chlorides in elastomeric seals or plastic devises can cause stress corrosion cracking of austenitic 18-8 type stainless steels. Ammonia introduced to control pH and minimize the corrosion of steel can promote stress corrosion cracking of copper alloys in adjacent equipment 17.

The overall design should be reviewed for materials compatibility if a fix for corrosion in one part of the process is not to cause environmental cracking elsewhere in the system.

Performance of Alloy grouping

The susceptibility of the respective groupings to specific environmental cracking is summarized in Table 5.

Table 5: Performance of Alloy grouping.

Reveling to the historical data, it shows that almost all the ejector condensers in SIDUM Desalination Plant in ACPT has failed, by frequent leak reported. The failure is almost located 18 3” Ejector condenser Nossel/stub inlet of steam. The 3” Nossel ejector condenser original design is CS material, as it is mentioned in inspection reports, whoever failed SS nossels was installed and then recommendations are set to replace it with CS by filler material ER309 due to the frequent failure. The CS showed a better resistance to corrosion as per inspection report , however another corrosion mechanism takes place due to dissimilar metal in contact which caused galvanic attack. The last inspected CS condenser stub showed sever oxidation deposits which caused perforation, Refer Photo of UZ/OSFIP/292/13, 354B04C-Sidem Ejector Condenser. Figure and photos shows that almost all condensers had failed on the left side 3” stub condenser rather than the right side which is still intact. Wrong design may be attributed or Material problem in which CS is failed. Crack Flaw indications was observed by DP inspection which may be attributed by Chloride SCC.

In Summary, It found that the Ejector Condenser in SIDEM Desalination Plant in ACPT had experienced frequent failure, in which it may contributed to deferent Corrosion Mechanisms. Chloride SCC may one of reasons on failure, however we still have lack of information to assess for CLSCC. Information needed are: Temperature, Chloride content pH and Oxygen level, Material is stress relived or not. Thus boiler sample is to be collected and sent to FDS Lab for Water Analysis.

Assessment the susceptibility of CLSCC

Susceptibility to CLSCC is usually assessed from knowledge of environmental variables, i.e. temperature, chloride content, temperature, pH and oxygen level. Condensed water samples collected from Ejector Condenser body. Sample sent to FDS Lab for analysis, Appendix 3 and 4 shows water analysis results collected from ACPT1 (ejector Condenser A/B) and ACPT2 (ejector Condenser C). Samples could not be collected from ACPT2 ejector condenser A/B due to operational constrains. The following Table 6 shows the analysis results. In analysing the susceptibility of CLSCC we use the assessment by API RP 581 ‘Risk- Based Inspection Technology Part 2 [4] which is shown in Table 7 bellow.

Table 6: Assessment the susceptibility of CLSCC.

Table 7: Proposed scheme for CLSCC susceptibility categories in aqueous media where pH ≤ 10, and in conditions where chloride deposits may form by drying out or by deposition.

Table 7 proposed scheme for CLSCC susceptibility categories in aqueous media where pH ≤ 10, and in conditions where chloride deposits may form by drying out or by deposition.

From Table 8 it shows that within Chloride Concentration 1 to 10 ppm and temperature range >66 to 149 the CLSCC category is Medium. Then the susceptibility is modified for additional factors in which the factors in yellow marks are factors attributed for the case of SIDEM Ejector Condenser. In conclusion and reveling to the assessment of susceptibility of CLSCC by API RP 581 ‘Risk-Based Inspection Technology Part 2. It can be concluded that the susceptibility 20 Category to Chloride Stress Corrosion Cracking of All Ejector Condenser in ACPT 1/2 are High.

Table 8: Change in susceptibility category for additional factors.

The result from the water analysis of ejector condenser shows that Chloride content varies from 2 to 5 ppm. Susceptibility assessment is studied b API RP581(2008), and it is shown that the ejector condenser susceptibility category to ClSCC is medium, then additional factors is considered to change the susceptibility category, and it is shows that considering the additional factors, the susceptibility of ClSCC for the ejector condenser is High. Another Damage mechanism is attributed to the Ejector Condenser is galvanic corrosion in which CS Nossel/Stub is attached to SS condenser shell (Design malfunctioning) [5,6].

Avoiding ClSCC by changing the environment is not practical, because deposition of Chloride is possible either on the outer shell of condenser (CUI), nor inside the shell by vapor drying out. In addition deploying Cl scavenger chemical may affect boilers performance. 2 control methods shall be applied to control the failure of ejector condenser.

1. Changing the material to more cracking risistant alloy such as Duplex Stainless steels S32760 (Zeron 100) S31254 (254 SMO), or N06625 (625 Alloy). Alloys Chemical Composition is shown in Table 9. Referring to approximate relative pipe cost assuming that it is applicable to exchangers also, it is noticed that Alloy 625 has much higher relative cost (22.8) compared with 254SMO alloy (9.2) and Ziron 100 (6.8). Alloy 254SMO has better pitting resistance than Ziron 100, and thus it is recommended to use Alloy 254SMO as a condenser material.

Table 9: Alloys Chemical Composition.

2. Residual stresses from metal facricatrion and during constructin is to be reduced by stress releiving heat treatment and shot peening the metal surface is mechanically peened at an intensity sufficient to result in residual compressive stresses at the surface. A good insulation application is to be applied on the outer shell of condeser as per ZADCO Spec for Hot Insulation, to avoid chloride penetration and deposition on the outer shell. It is recommended to monitor the corrosion of the condenser by conducting regular inspection (6 monthly is preferable).