The influence the moon has on life on earth is disputed. However, during an assignment in Southeast Asia, vibration experts from Sulzer were able to clearly confirm that the moon can even affect pumps. With its broad expertise, the Sulzer team was able to solve an unusual case of vibration.
Southeast Asia is a key exporter of LNG (liquefied natural gas) to global markets, and it is increasingly an LNG importer as well. Several LNG terminals have been installed in recent years, in which the shipped LNG is returned to a gaseous state. Usually, seawater is used as a heat source to vaporize the liquefied gas. In one of these terminals, trouble occurred with the vertical pumps that deliver the seawater to the plant. In total, there were four pumps, type TR 900/8/900, each with a capacity of 8250 m3/h, a pumping head of 40 m at a speed of 740 rpm, and a power of 1.07 MW. Because of their construction, vertical pumps—which have a long, unsupported vertical tube, a small foundation, and a heavy motor mass at the top of the pump—are very susceptible to vibration problems. As it turned out, the operators of the terminal also were faced with increased vibration levels.
Condensate extraction and cooling water pumps are often built as vertical pumps. That way, they can lift a fluid from a tank or sump located at a lower level. A typical implementation consists of a vertical pipe with the shaft inside; it has impellers at the bottom, which pump the fluid upwards. The vertical pipe, which is not stabilized, is located beneath the base where the pump is mounted. The motor is installed above the base. This type of construction is susceptible to vibration problems.
After the system was put into operation in April 2013, high vibration amplitudes occurred, with oscillation velocity of up to 17 mm/s (RMS, root mean square). Because of the typical design of vertical pumps, the highest vibration amplitudes are usually observed at the non-drive end of the motor. This often results in the misleading conclusion that the motor is causing the vibrations. Following this misconception, local engineering consultants and service providers took many measurements, performed a great deal of maintenance work, and carried out many tests—all of which focused mainly on the motor. The fact that the vibration amplitudes were not constant also led to this false assumption by the local consultants. All in all, more than a dozen measures were implemented and investigations carried out locally before Sulzer headquarters was contacted in March 2014.As a first step, the Sulzer Core Technology team in Switzerland systematically evaluated the reports of the investigations that had already been performed. Their analysis led to the following hypothesis: the residual imbalance of the rotor, which is within specification, results in excitation forces, which oscillate with the frequency of the rotor speed. The vibration response at the top of the pump is high because the natural frequency of the pump has insufficient separation from the rotation frequency. The engineers quickly realized that the tides caused the fluctuating amplitudes. Depending on the depth of the water, the water mass coupled to the vertical pipe changed and thereby also changed the natural frequency of the pump. The resonance curve shifted (while the rotation speed remained constant), and the vibration amplitude, therefore, also changed with the tidal level. The tides depend on the moon and the sun, which have different periodicities. Both periodicities can be found in the vibration amplitudes recorded in the control room.
RMS vibration values from the plant control room show the correlation between the vibration and the tides.
There are two approaches to solving vibration problems: either the excitation can be reduced or the transmission behavior of the structure, which may include mechanical resonance, can be modified. In this case, the engineers followed the second approach because a certain residual imbalance is inevitable. However, changing the natural frequency of an existing machine of this size is difficult. For that reason, they decided to use a passive dynamic vibration absorber as a countermeasure. Based on this hypothesis and this decision, they prepared field measurements and a prototype absorber. Three of these absorbers were manufactured and shipped to Southeast Asia to verify the efficacy of the absorber during field measurements.
The field measurements in July 2014 confirmed the hypothesis, and they successfully demonstrated the effectiveness of the absorber. The measurement of the vibrations at the motor without an absorber quickly confirmed that a natural frequency lay just below the rotation frequency of 12.4 Hz. This led to high vibration amplitudes when the pump was in operation. Furthermore, it was clear that several natural frequencies changed with the tide. This interrelationship also explains the fluctuating amplitudes.
A dynamic vibration absorber is, essentially, an additional single-mass oscillator that is added to the vibrating structure. This extra oscillator adds a further degree of freedom to the system and, thereby, also an additional natural frequency. The resulting pair of natural vibration modes is characterized by the structure and by the absorber vibrating in phase at the lower natural frequency and out of phase at the upper natural frequency. The coupled system also features an antiresonance between the two natural frequencies. If the absorber frequency is correctly tuned, the antiresonance exactly matches the objectionable rotation frequency. Compared with the response of the structure without the absorber, the antiresonance very effectively reduces the vibration amplitudes.
Performance of a passive vibration absorber. The phase response is shown left and the amplitude response right. The light-blue curve shows the dynamic behavior of the structure alone, and the dark-blue curve indicates the behavior of the coupled system of structure and absorber. The effectiveness of the absorber can be easily seen from the amplitude response.
After the engineers had mounted the three absorbers and tuned them to the rotation frequency of the pump, the vibration amplitudes reduced considerably. Even with different tide levels, the absorbers are always very effective. After the successful demonstration, the absorber design was improved based on the prototype, and a stress analysis was provided. Three of these absorbers were then installed on each of the four pumps in April 2015, and they were tuned to the rotation speed. With these absorbers, the vibration velocities measured in the control room remained well under 2 mm/s (RMS) for all operating conditions.
This case clearly shows that Sulzer has a wide range of experience and the best equipment for solving vibration problems. In addition to vertical pumps, Sulzer can also solve vibration problems on all other kinds of machines. Further examples will follow in the Sulzer Technical Review.