Comparison of the corrosion products before and after SG design changes

In the period 1994-1999 we focused our study on the comparison of the phase composition of corrosion products taken from the NPP Bohunice before and after changes in the feed water pipeline system.

Schematic drawings of VVER steam generators (SG) with indicated places of scrapped corrosion specimens are presented in Fig.3.

Serious damages were observed in the region of T-junction (position 4 in Fig.3) as well as of pipe-collector and outlet nozzles on many VVER440 SGs after approximately ten years of operation [9,17]. Therefore, the former feed-water distributing system has been replaced by an advanced feed-water distributing system of EBO design at SGs of NPP Jaslovske Bohunice [18, 19]. The advanced system consists of a V-shaped junction connected to the left — and the right part water distributing chambers both located above the tube bundle and few feed water boxes with water ejectors inserted into the tube bundle and connected to the distributing chamber by distributing pipelines.

After five year’s operation in the SG No. 35 in the NPP outage one feed water box and corresponding distributing pipelines were replaced by new ones with the aim to analyse their overall stage and corrosion products on walls. For comparison, some parts of the former feed-water-distributing system from the SG Number 46 were cut out and analysed.

More than 50 specimens were collected from the NPP Bohunice secondary circuit in 1998­2000. The investigation was focused mainly on the corrosion process going on in steam
generators SG35 with new design and SG46 with old design. Nevertheless, additional measurements performed on the corrosion products from SG31 and SG32 confirmed that corrosion process in all 6 steam generators of one reactor unit is the same and corrosion layers are on the some places altogether identical.

Fig. 3. Cross section of SG46 (Numbers indicate the places, where the specimens were scrapped)

All measured specimens contain iron in magnetic and many times also in paramagnetic phases. Magnetic phases consist in form of nearly stoichiometric magnetite (y-Fe3O4), hematite (a-Fe2O3), and in some case also iron carbides. The paramagnetic fractions are presented in Mossbauer spectra by a doublet and a singlet. Its parameters are close to hydro-oxide (FeOOH) parameters or to parameters of small, so called superparamagnetic particles of iron oxides (hydrooxides) with the mean diameter of about 10 nm (see Table 2 and Table 3).

MS confirmed its excellent ability to identify steel samples phase composition although its sawdust form and relative small amount (~ 100 mg). Our experiences with such measurements applied on different VVER-440 construction materials were published in [15, 20-21]. From other works it is possible to mention [22]. MS confirmed an austenitic structure of STN 17247 steel and ferrite structure of STN 12022 steel. Differences between these two materials are well observable (see Table 2 and Fig. 4 and Fig. 5). According to the in-situ visual inspections performed at SG35 (1998) and SG46 (1999) as well as MS results, significant differences in corrosion layers and material quality were observed. The feed water tubes in SG46 were significantly corroded after 13 years operation.

Sextets

Singlet

Sample

Hi

Arel

H2

Arel

H3

Arel

IS

Arel

Fig.

(sawdust)

(T)

(%)

(T)

(%)

(T)

(%)

(mm/ s)

(%)

SG35

29.3

18.7

26.3

18.5

22.8

27.4

-0.12

35.4

4

SG46

33.3

80.5

31.2

19.5

5

Accuracy

±0,1

±0,5

±0,1

±0,5

±0,1

±0,5

±0,04

±0,04

Table 2. MS parameters of the steam generators base material

Fig. 4. MS spectrum of SG35 base material

Fig. 5. MS spectrum of SG46 base material

Our results confirmed that during operation time a faint oxidation surroundings was in the observed steam generator SG35 after 5 years of operation time and the corrosion samples were fully without base material particles.

Magnetite was identified as dominant component in all studied samples (see Table 3). Mossbauer spectrum of the steam generators (both SG35 and SG46) surface layer is the superposition of two sextets with hyperfine magnetic field HefA = 49,4T and HefB = 45.8T. Sextet HefA corresponds to the Fe3+ ions in tetrahedral (A) sites and sextet HefB corresponds to Fe[10]+ and Fe3+ ions in octahedral (B) sites in magnetite spinel structure (Fe3O4).

Sample

Magnetite

Doublet

Sin

glet

Ha

(T)

Arel

Hb

Arel

rAB

IS

QS

Arel

IS

Arel

(%)

(T)

(%)

(a. u.)

(mm/ s)

(mm/ s)

(%)

(mm/ s)

(%)

L754

49.0

35.0

45.9

65.0

0.538

L755

49.0

35.3

45.9

64.7

0.546

L757

49.0

34.8

45.9

65.2

0.534

L758

49.0

34.1

45.9

62.3

0.547

-0.20

3.6

L789

49.0

34.4

46.0

65.6

0.535

L790

49.0

34.9

46.0

64.3

0.543

-0.18

0.8

L759

49.0

35.4

45.9

63.9

0.534

-0.20

0.7

L777

49.0

35.0

45.9

65.0

0.538

L786

49.0

35.2

46.0

64.8

0.545

L787

49.1

36.5

46.0

56.1

0.651

0.22

0.67

2.0

-0.20

5.4

L760

49.0

34.1

45.9

64.3

0.530

-0.17

1.6

L761

49.0

35.0

45.9

63.9

0.547

-0.23

1.1

L762

49.1

34.8

46.0

56.4

0.617

-0.20

8.8

L779

49.0

33.4

45.9

62.9

0.531

0.10

3.7

Accuracy

±0,1

±0,5

±0,1

±0,5

±0,04

±0,04

±0,5

±0,04

±0,5

Table 3. MS parameters of corrosion products taken from the steam generator SG351

In contrast to magnetite, whose spectrum is characterised by two sextets, the hematite phase present in the powders gives a single sextet. The relatively narrow line width (Г) of the a — Fe2O3 (mainly 0,24 4 0,26 mm/ s) indicates presence of a well-crystallised phase with few, if any, substitutions of other elements for Fe. However, in some spectra (mainly from filter deposits studied later), both the lower hyperfine field and the larger width (about 0.33 — 0.34 mm/ s) could indicate a poorer crystallinity and/or a higher degree of substitution. These findings are in good agreement with those obtained by E. De Grave [23]. Similar inspirative results (focused also on corrosion products from VVER-440 construction materials) were published in [24-26].

For the ideal stoichiometric Fe3O4 the quantity rAB (ratio between A and B sub-component areas) is equal to 0.535. In the case that magnetite is the dominant (sole) phase in the sample, the deviation from the ideal value of rAB is minimal (see Table 3). Significant deviations could be explained by a small degree of oxidation of magnetite, resulting in presence of vacancies or substitution by non/ magnetic irons in the octahedral sub-lattice. Slight substitution of other elements (Mg, Ni, Cu, …) for Fe in the magnetite lattice is not unlikely, and this has a similar effect on the A — to B-site area ratio. Therefore, it is not feasible to conclude anything quantitatively about the degree of oxidation. Qualitatively, it can be inferred that this degree must be very low. [11]

During visual inspection of removed feed water dispersion box (1998), 2 disturbing undefined metallic particles, fixed in one of outlet nozzle, were found. Both were homogenised and analysed by MS. It has been shown that these high-corroded parts ("loose parts" found in outlet nozzle of ejector) originate not from the 17247 steel but high probably from GOST 20K steel (probably some particles from the corrosion deposit from the bottom part of the steam generator moved by flow and ejection effect into the nozzle). 2

Fig. 6. Position of corrosion product scraps from the feed water dispersion box (SG35)

Code

Hematite

Magnetite

Base material

Doublet 1

Doublet 2

H1

Arel

Ha

Arel

Hb

Arel

H4

Arel

H5

Arel

IS1

Arel

IS2

Arel

(T)

%

(T)

%

(T)

%

(T)

%

(T)

%

(mm

/s)

%

(mm

/s)

%

M005

49.0

35.4

45.8

64.6

M006

49.1

36.5

45.9

63.5

M007

50.0

16.9

49.2

25.6

45.8

38.2

33.0

1.6

0.84

17.7

M008

49.0

35.6

45.9

64.1

M009

51.5

13.4

49.1

32.1

45.9

54.5

M010

49.1

36.5

45.8

63.5

M012

51.5

12.5

49.2

31.9

46.0

55.6

M013

48.8

25.3

45.7

40.5

33.0

30.2

30.8

4.0

M014

49.0

9.9

45.8

13.6

33.0

66.6

30.7

9.9

M015

48.5

6.0

45.6

8.6

33.0

73.1

30.6

12.3

Accuracy

+0,1

+0,5

+0,1

+0,5

+0,1

+0,5

+0,1

+0,5

+0,1

+0,5

+0,1

+0,5

+0,1

+0,5

Table 4. MS parameters of corrosion products taken from the steam generator SG462

Mossbauer measurements on the corrosion specimens scrapped from different position of the feed water distributing system show that the outside layer consists exclusively from magnetite but the inside layer contains also hematite. Its amount decreases in successive steps towards the steam generator. The cause of this result is probably in fact that outside the system there is boiling water at the temperature of approximately 260 °C with higher salt concentrations and inside there is the feed water at the temperature up to 225 °C. Changes in the inside temperature in region (158-225 °C) can occur in dependence on the operation regime of high-pressure pumps in NPP secondary circuit.

The most corroded areas of the former feed water distributing system are the welds in the T — junction (see Fig. 7). Due to dynamic effects of the feed water flow with local dynamic overpressures of 20 to 30 kPa or local dynamic forces up to 1000 N (in the water at the pressure of about 4,4 MPa) on the inner pipe wall in the region of T-junction, the content of corrosion products was reduced and moved into whole secondary circuit. Particles of the feed water tube of SG46 base material were identified also in sediments.

Fig. 7. Position of corrosion product scraps from the feed water dispersion tube (SG46)