Power Generation
Steam Turbine Generator
Root cause of failure identified, preventive action provide. Apportion responsibility and liability
Steam Turbine Failure Investigation
Identify the root cause of catastrophic steam turbine failure and detail action required to prevent re-occurrence.
Our client asked us to carry out an investigation on a steam turbine blade failure onboard their FPSO (Floating Production Storage and Offloading) vessel, which had suffered previous blade related failures since it had been commissioned. Our brief was to identify the root cause of the failure and to recommend solutions to prevent any future problems identified.
During the investigation two independent failures were identified, the second and final event occurred as a direct result of the first failure.
ROTOR 10TH STAGE BLADE FAILURE – ROOT CAUSE ANALYSIS (Using the 5 Why’s principle)
Problem: All of the 10th stage blades fatigue fractured at the root of the blades at the same time, Figure 1 below shows the diaphragm to blade configuration.
The 10th row of blades rubbed against the bottom half diaphragm blade shield in two locations, evidence of this can be seen in attached (Figure 2) and macro of the smaller rub (Figure 3).
There is a significant number of built up layers of blade contact deposit in the micrograph consistent with the high cycle low stress fatigue seen on the blade fracture face.
The rub on the bottom half diaphragm blade shield would be caused by insufficient clearance (Assembly drawing specifies 0.035″).
A – The blades were tipped to a larger dimension. (May have had clearance during normal operation but rubbed on start-up and slow down).
B – The rotor shaft was quenched in water from the gland steam condenser causing rotor hogging (shaft bending causing the blades to touch the diaphragm blade shield).
C – Combination of A & B above.
A – The rotor in the turbine was not the original item, but a refurbished rotor originally supplied in another FPSO. (Our client maintained a spare fleet rotor, once the spare had been used the removed rotor would then be refurbished and become the next spare). It is not known when this particular rotor became the fleet spare. The rotor was subsequently fitted into this machine in 2013
The blade to diaphragm shield clearance for this turbine stage (row 10) is 0.035”, this is achieved by measuring the blade shield diameter during assembly, and grinding the blade tip diameter to achieve the required clearance. The blade shield diameter has a machining tolerance of +/- 0.030”.
The balance of probability is that the 10th stage blades were tipped to a larger diameter, and that this was not identified during the reassembly.
B – The Gland steam condenser is located 4.5 meters above the turbine gland drain connection without a steam trap and no drain was found, despite one being shown on the P&ID drawings. The gland steam condenser fan is rated at 0.2 barg. The system does not work as intended and is likely to be permanently filled with water, during a shutdown the water will drain back and quench the rotor.
A – Turbines fitted with diaphragm blade shields have clearance tolerances that are not ideally suited for interchangeable rotors.
B – The Gland condenser is installed and commissioned incorrectly, evidence of condensate flashing off on the exhaust casing was seen during the FPSO survey
The turbine tripped on high vibration during this event, the client carried out the standard procedure of slow rolling (barring speed) a turbine rotor for 24 hours if a restart is prevented due to high vibration, (it’s assumed the rotor has a bend in it). After 24 hours the turbine was restarted successfully with normal vibration characteristics. The client had no knowledge of the 10th row blade failure at this time.
Over an unknown period time the 10th row blade debris was rumbled around in the casing space (Figure 4).
Nozzle fragments impacted and became “welded” to the casing fasteners (Figure 5). The rumbling continued until the debris was small enough to become lodged in the 11th stage diaphragm nozzle throat areas (Figure 6).
This caused the upstream pressure on the 11th stage diaphragm to exceed the diaphragm design capability, causing it to collapse onto the 11th stage rotor wheel resulting in catastrophic failure of the wheel head and loss of the blades (Figure 7).
The 11th Stage catastrophic failure happened as a direct consequence of the 10th stage blades fatigue failure.
Producing a report that represents the facts of the case, that is supported by objective evidence in the form of data, images, and where required external expert supplementary reports.
We keep in continuous communication with our client’s leadership and operations teams to ensure they are up to date as the inspection progresses.
The Gland steam condenser was not operational (fans stopped) once started the delta T across the cooling water supply and discharge proved that the system is not working due to the fan having insufficient suction capacity for the 4.5m head.
A drain was marked on the P&ID, but was not found, there was no steam trap in the system, and therefore it is almost certain that the pipework is flooded with water.
There was evidence on the exhaust end of the starboard turbine that condensate from the glands was leaking onto the casing, it was recommended that a steam trap be fitted on these lines.
The root cause of this failure was due to the 10th stage blades rubbing against the diaphragm blade shield, on multiple occasions over an undetermined period of time. This was likely caused by insufficient blade to diaphragm shield clearance and / or rotor displacement (hogging).
