Symposium case studies

Compressor and Pump Case Studies Showcased at Turbomachinery Symposium
Dozens of detailed case studies on compressor and pump maintenance were a feature of last week’s Turbomachinery Symposium in Houston, Texas.
Mathew Moll, a Rotating machinery Consultant for the DuPont’s Rotating Machinery Group of Wilmington, Delaware, spoke about the correction of chronic thrust bearing failures on a refrigeration compressor. His case study centered upon a 2250 HP 1678 kW 4-stage centrifugal compressor running at 7100 rpm which was equipped with proximity probes. He said that it did not achieve contract performance when operating at high head. A new rotor had been added to the compressor in the 1990s. “The new rotor had improved performance but not over the entire range,” said Moll.” High head conditions led to excessive discharge.”
Due to a couple of thrust bearing failures a root cause analysis (RCA) was begun. Every possible area was investigated including analysis of the lube oil system, as well as spectro-chemical and physical analysis. The RCA team also checked the thrust and radial probes as well as the casing, the temperature of the oil, and the temperature of the thrust bearing.
“The compressor rotor’s running position would suddenly change towards the axial thrust position, about 2.5 mm per month,” said Moll. “This happened after about six months to two years of stable operation.’
According to Moll, lube oil supply and bearing temperature provided no indication that the bearing was in distress. When maintenance personnel removed the thrust bearing, however, it showed heavy scoring. The company brought in the OEM and reviewed the original thrust bearing sizing calculations. Loads of 170 psi load were expected, but loads of 360 psi were isolated.  
“Our conclusion was that we had an overloaded thrust bearing in the compressor,” said Moll. “One recommendation was to increase the balance piston sizing.”
That recommendation was not followed due to cost considerations. A Houston-based company known as Turbo Components and Engineering (TCE) was brought in to find an alternative Julia Postill, a mechanical engineer at TCE, explained that a redesign of the thrust bearing was done to increase its load carrying capacity.
“We decided to do it in such a way as to avoid modification of the bearing housing as that would add cost,” she said.
TCE designed it with a larger overall area in order to reduce unit load from 360 psi to 242 psi – 41% bigger. This was accomplished by switching from four pads to seven pads. The mean velocity was raised slightly as a result. In addition, copper backing material was included for faster heat dissipation and the pivot was offset to further increase load carrying capacity.
What of the rotor motion mentioned earlier? Moll reports that it continued to cycle about 2 mm up or down, although the mean position of the compressor rotor didn’t change on average by more than half a millimeter.
“We were happy with that,” said Moll. “A check of the lube oil no longer showed excessive wear of bearing material.”
Further, the average oil temp change across the bearing increased from 8.5°C to 10.3°C as a result of the retrofit, but this was anticipated in the redesign.
“Our approach cost 10% of what it would have cost by resizing the balance piston,” said Moll.
He ended with two lessons learned from the project: oil analysis is an excellent indication of vibration and thrust position or bearing faults; and bearing discharge oil temperature is not a useful indication in bearing condition monitoring.
Qatar Compressor Case Study
Another case study at the symposium focused upon trip reduction in active magnetic bearings (AMB) in a turboexpander compressor being used by RasGas in Qatar. Two of the units in question were utilized for LNG and another for gas processing.  
Jim Cencula, a Chief Engineer with GE Oil & Gas explained that AMBs are a relatively new technology in this application.
Initial failures were one offs and related to electronic components such as sensor rings, detector boards and batteries. Once Unit 1 experienced 32 trips in a few months, RCA efforts began in earnest. Investigators discovered poor soldering, some bolting problems, and other issues. In addition, excessive axial vibration showed up which was causing trips.
The automatic thrust balance system was not operating properly,” said Cencula. “We needed to reset it and change the logic of the system.”
Under high rate operations, this created low back-wheel pressure and high axial thrust load. During the RCA, an unknown frequency vibration was found at low pressure/high-flow and high-speed settings.
Cencula said that the thrust bias current of the AMBs was upped from 12 amps to 15 amps, and the machine center section changed out. This led to better thrust balance
“The source of the high frequency vibration is still to be determined,” said Cencula. “But there have been no trips for three years now.”