ANALYSIS OF THE FAILURE OF A SYNTHESIS GAS ROTOR COUPLING AND TURBINE SHAFT

J.C.Guild: Southern African Institute of Welding, Johannesburg, South Africa

Key Words: Failure Investigation, Rotor, Turbine Shaft, Fatigue, Corrosion Fatigue, Pitting, Coupling.

ABSTRACT

The investigation into the failure of a synthesis gas rotor coupling and turbine shaft is described. The failure was by way of corrosion fatigue, originating at corrosion pits on the inside surface of the coupling.

INTRODUCTION

The failure occurred on a synthesis gas compressor shaft. The failure included the shaft, shroud of the coupling and the spacer section. A previous failure of the turbine coupling had occurred about one and a half years before. The failure resulted in a plant shutdown of approximately three weeks. The broken components are shown in Figure 1.

INVESTIGATION

The turbine shaft had fractured in a position where the shaft increases in section. The fracture of the shaft, which is shown in Figure 2, exhibited two major areas. The bulk of the fracture surface had a slightly torsional, fine, fibrous appearance and the remainder had a smooth well rubbed finish. The failure of the coupling spacer had occurred at the compressor end and as shown in Figure 3, there was positive evidence of a fatigue failure mechanism on the fracture surface. The portions of the fracture faces of the shroud of the coupling which were undamaged had a relatively ductile, fibrous, woody appearance.

Inside the coupling there was a large amount of sludge. This sludge had been noticed on previous occasions. Samples of this deposit were sent for analysis and the following results were obtained:

XRF: Major: Barium

Minor: Iron

Trace: Tin, Cadmium, Zirconium, Zinc, Copper, Manganese, Calcium,

Silicon, Aluminium.

XRD: Mostly amorphous but phases were iron oxides and barium oxides.

Examination of the inside surface of the coupling spacer revealed areas of pitting corrosion at the compressor end which had small, fine cracks associated with them. These cracks were present in a position corresponding to the areas of fatigue but at another point on the circumference. Generally the inner surface of the coupling was covered with a light, black deposit but probably much of this was a result of the subsequent fire. It was also noticed that in the middle of the spacer a groove (approximately 2mm deep) appeared to have been machined around the full circumference of the inside surface. The function of this groove could not be established, however, it did not appear to have any significance in the failure.

Samples of the spacer and the turbine shaft were sent for metal analysis. The results are given in Table 1.

Component	C	Si	Mn	Ni	Cr	Mo	V
Coupling	0.40	0.28	0.86	0.09	0.91	0.21	0.014
Shaft	0.25	0.25	0.43	3.0	0.38	0.41	0.080

Table 1: Chemical analysis of failed components.

Microsections were taken through the fatigue failure areas of the spacer and through the associated fine cracking. These showed pitting corrosion to be clearly associated with the failure. Sections through the fine cracks indicate that these are also attributable to corrosion fatigue. This is shown in Figure 4.

Microsections also revealed that the steel cleanliness of the coupling spacer was poor with numerous elongated sulphides and oxide trailers being present. The microstructure consisted of a fine grained tempered martensite. Hardness tests on these sections gave the hardness as being between 318 and 320 Hv. (30kg) This is equivalent to approximately 66 tsi tensile strength.

Microsections were also taken through the rubbed fracture surface of the shaft. These showed that the shaft had been chrome plated, but there was no evidence of cracking emanating from the plating or elsewhere. The steel cleanliness was satisfactory and the microstructure consisted of tempered martensite. Hardness tests gave the hardness as being 317 Hv. (30kg)

CONCLUSSIONS AND RECOMMENDATIONS

This failure appears to have been a repeat of the previous failure. The coupling spacer showed clear evidence of a fatigue failure mechanism which has originated at corrosion pitting of the inside surface. No evidence has been found to suggest that the shaft failed prior to the coupling.

The spacer has been designed with a relatively thin section for this duty and this has necessitated the use of a high strength material. Unfortunately corrosion due to sludge contamination and ingress of moisture has caused a serious reduction of the fatigue strength resulting in failure.

The replacement coupling has been protected by means of epoxy resin. This should improve matters, but it is felt that further steps, such as reducing the applied stress to the coupling, by using a heavier section, and the removal of water from the tube system are necessary to ensure similar failures are avoided.