By Martin Baier
This article uses a case study to illustrate the side of power quality that is not often portrayed. With increasing awareness of the power quality issues among the electrical equipment users, it is not uncommon to see that power quality problem is suspected 'a priori', long before serious investigation is undertaken into the root cause of the problem. This approach can often mask the real cause of the problem and delay its correction while incurring unnecessary costs.
An oxygen producing plant in the Midwest US operates a large induction motor driving an air compressor, as well as several other medium-size induction motors. A four-pole 2500 hp compressor motor is supplied from a 4160 V bus along with another 600 hp motor, while the balance of the motors are supplied from a nearby 480 V motor control center (MCC).
Shortly after the plant startup a strong vibration was detected on the blower, the motor and the associated piping. The vibration was so severe that it could be felt in the adjacent electrical room where both medium voltage and low voltage MCCs were located. The blower regime consists of about five-second full load cycle followed by the same length period of no load.
Comprehensive testing and vibration analysis resulted in some modifications to the drive mechanical design. Steel pedestal frames were filled with concrete to modify the natural frequencies, a flexible clutch was installed between the feeder blower and the motor and the shafts were realigned and balanced repeatedly. Because the impact of the modifications was only marginal, the attention was diverted to the electrical system.
Importance of power quality as it relates to the starting and operation of induction motors is widely recognized. General symptoms arising from inadequate power quality in distribution systems directly supplying motors include abnormal motor heating and vibration. The causes of such symptoms are low/high voltage, voltage unbalance and harmonics. An electric motor itself can also be the source of vibration. Mechanically born causes such as the rotor unbalance due to the bent or unbalanced rotor and shaft misalignment with the driven machinery are some examples. Such problems can be easily diagnosed by comparing the vibration signatures during motor operation and its coast down. Uneven air-gap or broken rotor bar and other winding problems, on the other hand, are not as apparent with the motor on line nor, are these problems easily detected by the motor electrical protection.
In the case at hand, the motor was ruled out as a possible culprit after there was no resultant change to vibration levels and its pattern following the motor and blower replacement.
At this point in the investigation, on-site personnel connected the power analyzer to the 4160 V bus via the switchgear potential transformers (PT). The waveform captures revealed an alarming number and severity of voltage waveform notches and the decision was made to employ the services of a power quality expert. The comprehensive analysis of the power supply to the motors, with the objective to locate the source of disturbances, seemed logically justified. The fact that there were no loads in the system that could cause such type of notching seemed to be of secondary importance.
While it is not the intent of this article to guide the reader in power quality survey methods, the author would like to point out a few commonly overlooked details. When a power quality survey is undertaken as a means of identifying the cause of a problem rather than merely quantifying the power quality problem or deficiency, a few important rules should be born in mind:
1. Correlation of symptoms and disturbances: While the importance of accurate and concurrent analysis of the waveform in all phases of the electrical supply is generally recognized, often overlooked is the importance of being able to correlate the time of occurrence of the symptom (vibration) with the suspected cause (disturbance). Investigations can become very time consuming and costly when random occurring disturbances are suspect. On the other hand, vibration is easy to monitor and correlate with the performance of the electrical system. Although power quality analyzers commonly available today do not feature a vibration pickup input, fast multichannel recorders can be used or inexpensively adapted for this purpose.
2. Point of instrumentation connection: It is recommended that the survey initiate at the most remote end of the distribution where the problems are observed or suspected, systematically proceeding toward the source addressing all intermediate voltage levels.
This power quality survey began at the point of motor connection, at the medium voltage MCC. The power quality analyzer was connected to all three phases via the existing potential transformers connected in Òopen WyeÓ configuration. The potential transformers were rated 4,200/120 V. The fourth instrument channel was set to monitor motor current via ÒAÓ phase current transformer rated 400/5 A.
The first set of recordings was taken with the compressor off-line. Two additional recordings were taken with the drive being started and idling respectively. The fourth set of data was taken during the compressor load cycle. All data sets were subjected to the preliminary on-site evaluation.
The analysis of the collected waveform disturbances revealed three types of events that all seemed to be distinctly related to the operation of the blower drive.
The first type is best described as the general corruption of the voltage waveform in one or more potentials, while significant reduction of the waveform RMS value accompanies the event. It was noted that none of the voltage disturbances were accompanied by a similar anomaly in the motor current. Most of the disturbances of this type were found in A-B and C-A potentials. All of the disturbances of this type occurred during the load cycle of the compressor.
The second type of waveform irregularity had a distinct ringing-transient pattern that usually accompanies the exchange of energy between shunt-connected capacitors and the inductances of the system. Here, transients were present and coincident in both voltage and current waveforms, while the RMS value of the waveform was not significantly affected. These disturbances were traced to the inrush phenomenon of the capacitor at the subject motor during starting. This transient phenomenon is quite common and any connection to the objectionable vibration could be ruled out.
The last type of irregularity was observed on one occasion only. Its pattern resembled the previous type, again affecting all potentials and the current, with the distinct ringing pattern on all channels. Because no intentional switching of the motor under test took place at the time of the event, this disturbance was attributed to the starting of the adjacent 600-hp motor. This phenomenon is sometimes called a sympathetic inrush. Again, no connection could be established between this type of disturbance and the vibration due to the transient nature of this disturbance and the frequency of its occurrence.
From all collected data, the first type of disturbance was most significant with regards to the objective of an investigation. The significant depression of the RMS content and the potential of negative sequence voltage in the supply to the motor were believed to have serious implications on the operation of the motor. The quantity and severity of the waveform irregularities was simply overwhelming. It was also interesting that none of the disturbances even remotely resembled the notching first spotted by on-site personnel several weeks earlier. Because the source of such irregularity was not immediately obvious, another set of voltage measurements was taken at the 480 V MCC, supplied from 4,160 V bus via another transformer. None of the objectionable signatures were recorded during the blower operation at this location, which immediately brought the issue of equipment integrity into focus.
The equipment was deenergized and thoroughly inspected. The problem was soon spotted in the auxiliary contact assembly that is part of the circuit between the PT secondaries and the terminals available for monitoring. Both moving and stationary contacts were oxidized and not fully conducting. The presence of vibration during the blower operation set the relative movement between the fixed parts in the cell and the PT carriage causing the intermittent contact that closely tracked the vibration acceleration pattern. The picture became suddenly complete.
The contacts were reconditioned and their replacement recommended at the next earliest opportunity. After the equipment was returned to service, disturbances were no longer present and the power system was given a clean bill of health.
Lesson learned
It has likely happened to everyone who once hooked up a power analyzer to forget to turn the instrument off while moving it from one location to another. Any attempt to analyze such an overwhelming amount of data that inevitably results can be quite a frustrating experience. Also not uncommon is to experience a loose connection with similar results. Perhaps the best defense against ghosts is to remember that disturbances are very seldomly of a local occurrence. Instead, they invariably tend to travel in a power system among various voltage levels with various degrees of attenuation and cancellation. Hence, one of the most useful rules of thumb: One should always remember to compare the measurements from at least two locations preferably using different means of voltage detection (PTs, probes, etc.). If this approach fails to locate the problem, using two different power analyzers or a power analyzer and a scope is the next logical step.
Martin Baier, P.Eng., is with Cutler-Hammer's Engineering Services and Systems. ET