Load Bank Testing - Mining and Exploration
When using a load bank to test a 2000 kW diesel generator, the AGM was set to automatically capture transient events. During de-energization of the load bank, a peak transient of 964 V, or 246% of rated voltage was captured as shown in the figure below. After evaluation and analysis, this was determined to be caused by opening the main generator breaker to the load bank before all the cooling fans were shut off. This is thus a transient recovery voltage event, where the generator breaker interrupted the inductive current to the fans. This sudden change caused the inductive reactance of the motors to try and maintain the inductive current flow. This event shows that a damaging voltage is created that can fail motors and other equipment on the load bank. This led to an operational procedure change: first remove all fans from service before opening the main generator breaker.
Figure 1. Transient recovery voltage from generator breaker switching.
Static Var Compensator Commissioning - Theme Park Industry
Theme parks are constantly in search of new and intense thrill rides, and with this often comes much greater power requirements. In this case, a ride’s reactive power requirement was so great, that it necessitated a static var compensator (i.e. real-time var compensation) to keep the ride from creating excessive voltage flicker on the power distribution system. A thorough study by Acacia proved the need by simulating the impacting ride load as well as the static var to mitigate the issue. After installation of the static var, Acacia monitored the equipment to ensure it was providing the specified vars within the specified time response. Due to the fast load and unload profile of the ride, the static var needed to operate in an open-loop control mode, matching var for var to create a net unity power factor load. This then minimized voltage drop on the grid to nearly zero.
During the commissioning effort, we discovered the response was inadequate and the source inverter modules were producing excessive harmonics. Subsequent investigation by the vendor discovered a need for significant updates in the controls and logic. Unfortunately, that initial update caused a failure in the static var which blew all fuses on the input. Captured waveforms of the fault aided the vendor’s engineer so that a quick determination and correction could be made. Eventually, as shown below, the static var was corrected to yield the desired response. This type of commissioning shows that in many cases with very sophisticated power equipment, supplied equipment is often not set to “run as is”. Often extensive commissioning and troubleshooting is needed to get the equipment operating as specified.
Figure 2. Static var compensator response (Q SVC) to load var demand (Q Lift).
Wind Farm Transformer Failure - Renewable Energy
Detailed investigation of pad mount transformer (PMT) failures led to discovering wind turbine inverter induced high frequency noise being applied to the low voltage line-to-neutral windings. This was not expected, and significant effort was made to ensure this noise was truly present and not a sympathetic response of the data acquisition hardware to radiated inverter switching noise. Measurements were moved from inside the turbine tower to outside on the direct secondary of the PMT. The key issue with this noise is that these transformers were not of any special design to accommodate the high frequency content, and its creation of high peak voltage stress. As shown in the figure below, peak voltage stress due to the noise exceeded 1.5 per unit of the rated secondary line to neutral voltage (398.37 VLN).
Figure 3. High peak voltage stress from inverter high frequency switching noise.
Filter Tuning and Rectifier Misoperation - Petroleum & Chemicals
Acacia was brought in to evaluate a 5th harmonic filter that was thought to have been tuned incorrectly and as a result failed catastrophically. The plant’s initial conclusion was they added capacitors per drawings but incorrectly tapped the reactor. After investigation, the facts of the situation, which were significantly different, became apparent. First, measurements of the filter tuning reactor at various tap positions revealed that the manufacturer’s drawings of tap vs. inductance were grossly in error. They did not present the tuning reactor correctly nor did they construct it for its intended purpose of handling two plant operating conditions with two different arrangements of capacitor units. From this knowledge, the filter reactor was then properly set correctly per measured inductance, desired power factor correction needs and desired tuning.
To be thorough, we also measured the currents produced by both sides of their 12-pulse rectifier system. The delta side was operating correctly with symmetrical and balanced currents as shown below. The wye side however was misoperating and generating an unhealthy amount of non-characteristic harmonics as shown below. These harmonics would have been the major contributing factor that lead to the harmonic filter failure. The filter was then left off-line until the rectifier system was repaired. The waveform imbalance was also causing excessive overheating of the transformer core (measured to be 185o C) due to saturation. Recommendations were made to curtail the process load until the rectifier was repaired.
Figure 4. Delta side balanced current waveforms and expected 6-pulse characteristic harmonics.
Figure 5. Wye side un-balanced current waveforms and unhealthy harmonic spectrum.
Mysterious Feeder Faults - Utility and Coops / Petroleum and Chemical
An oil pumping station utilized a long overhead feeder to transfer prime power generation from diesel generators to a pumping station. This line had a series of mysterious faults and subsequent trips with no apparent cause. To investigate this issue, instrumentation was set to capture faults at one end of the line. After capturing the first fault (see figure below), data was analyzed to determine the fault location on the feeder. That location was inspected and again yielded no conclusion of the cause. After capturing 4 to 5 faults, evaluating the fault location and performing multiple inspections, still no root cause had surfaced.
Figure 6. Voltage and current waveforms for captured line-to-ground fault on overhead line.
This led to recommending hiring locals to literally stand in several locations along the line and visually monitor the line until another fault occurred. From this action, one of the acute locals noticed insect activity on the line where wasps were building mud nests on the insulators. As shown in the figure below, this mud contamination covered significant insulator creep. It would have been just a matter of time until this mud nest would flashover on the insulator. After any flashover and fault for this issue, it is clear that the mud nest would have been vaporized leaving no indication of a problem. There is no solution that can control the nest building activities of these insects. Any material added to the insulator would compromise its insulating capability. The best and only option was replacing the overhead line with an underground cable.
Figure 7. Insulator contaminated with mud wasp nest.
Harmonics Solution with 519 Violation - Utilities & Coop
A distribution system was noted as having a serious problem with telephone interference. During hours when a local dredge operation was running, the local phone system would shut down. Without any on-site measurements, a network model was constructed to simulate the distribution network, complete with all overhead charging and pole mount capacitor banks. Simulations showed that the dredge (utilizing a 6-pulse drive system) created significant, but not necessarily severe voltage distortion on the system. Further analysis evaluating TIF and IT Product showed clearly that the harmonic currents flowing in the system were however producing very high levels of telephone interference. To mitigate the issue, existing pole mount capacitor banks were relocated at an optimal location to absorb the most relevant frequencies (i.e. frequencies in the audible range) produced by the dredge. Upon relocation of the bank, the telephone system worked fine with no interference issues. One side effect however was that the feeder voltage distortion increase up above 5% THD, the IEEE 519 limit for voltage distortion on medium voltage feeders. The utility decided that the excess level of distortion was completely acceptable given the phone system was now fully functional. In this case, exceeding IEEE distortion limits mitigated a more serious issue.
Prime Power Gen Failure Investigation - Petroleum & Chemical
A prime power generator station utilizing up to five parallel connected generators had sustained repeat generator failures. To keep the system running, replacements were sought out and installed quickly. After several replacements, failures were still occurring. Instruments were connected at the sight to several of the generators and revealed a sudden increase in 3rd harmonic neutral current when one particular generator was connected. Immediately, operators were instructed to remove that unit from service to avoid damage. The third harmonic current measured on the problem is shown below.
Figure 8. Neutral current in generator with 13/18 winding pitch.
Further investigation revealed that the generator being put on line when the high 3rd harmonic appeared had a different winding pitch (13/18) than all of the other operating generators (2/3). This is an application no-no when considering multiple parallel connected generators when operating with a solidly grounded system, and the system at this location had multiple solidly grounded paths for such current to flow in. One final plot clearly shows the third harmonic circulating in one of the other generators going away as soon as the odd pitched generator was removed from service. The solution: only add generators in parallel with the same winding pitch.
Figure 9. Neutral current in one of 4 gens with 2/3 winding pitch as 13/18 pitch unit is removed