LOCATING MULTIPLE GROUND FAULTS THE BLUE CHIP CASINO IN MICHIGAN CITY, IN

Transcription

1 LOCATING MULTIPLE GROUND FAULTS THE BLUE CHIP CASINO IN MICHIGAN CITY, IN A SUMMARY REVIEW OF THE FINDINGS AND INVESTIGATIVE TECHNIQUES THAT WERE USED TO RESOLVE THE GROUNDING PROBLEMS Page 1

2 The Blue Chip Casino Indiana law states that a casino operation has to take place on a free floating self propelled pleasure craft. A land based operation as in cities like Las Vegas or Atlantic City is prohibited. Boyd gaming had early on discovered that the law can be used to one s advantage by creating a semi land/water based operation, the so called casino ship. This is however no small undertaking. It requires that a body of water be provided that is capable of supporting a ship with cruiseliner dimensions. Hotel Pleasure Craft The casino is shown on the picture to the right. It illustrates nicely the size of the ship. Amazingly enough, a visitor will enter the hotel and step on board without noticing it. The transition is cleverly made via a floating walkway. To the common observer, hotel and casino have become one unit. Page 2

3 Power is supplied from shore via 4160/480VAC transformers. There are five 480/208/120VAC transformers onboard stepping down the power for further distribution. It is important to note that the secondary side of these transformers is solidly grounded. 4160VAC to 480VAC Shore feed into the pleasure craft A typical /120VAC delta/y 750 kva transformer (marked SP200B) is shown below. A total of four similar transformers provide slot power (SP panels) and lighting power (LP panels) across the ship. From the start, the customer experienced ground fault problems with the electrical distribution system. Certain types of ground faults will cause a current to flow through the ship s and back to the transformer neutral via the grounding electrode conductor. An ammeter type ground fault relay (Manufactured by IPS Systems in Florida) with CT that encircles the grounding electrode conductor measures the current that flows through the ship s. More details will be provided later on in this application note. Worthy of notice were two meters that produced humming noises due to to excessive current flow. 750KVA transformer with open front Five ground fault ammeters Note meter 1 at the top left is pegged Page 3

4 Overview After extensive discussions and a review of the electrical drawings, we came to believe that shipboard power distribution was per the diagram below. The loads such as the slot machines are all from the ship s. Thus, ground faults internal to these devices will not flow through the ship s but must eventually flow LIGHTING/SLOT PANELBOARD SB221 MAIN SWITCHBOARD SP200C SUB-PANELBOARD SP200B L O A D Typical Load Six 150 A circuits 42 circuits isolation LIGHTING/SLOT PANELBOARD SB222 SUB-PANELBOARD neutral One 1000 A, one 600 A, and three 150 A circuits 42 circuits main bonding jumper ground bus Six 150 A circuits Total of 15 Slot/Lighting Panelboards LIGHTING/SLOT PANELBOARD 42 circuits through the main bonding jumper back to the transformer supply. Note that the ground busses and neutral bars are all from their respective panel structures. There is currently no ground current detection at the main jumper location. The main switchboard, sub-panelboards, and lighting/slot panelboards enclosures are all bonded to the. What this means is that ground faults internal to these structures will produce currents that flow through the and will be detected by the CT that encircles the grounding electrode conductor. Also picked up by this CT sensor will be some capacitive leakage currents as a result of the proximity of phase conductors to conductive surfaces bonded to the A clamp-on ammeter wrapped around the grounding electrode conductor displayed a whopping current of 76 A. Even with a notch filter with 100 Hz cutoff frequency, the current was still unacceptably high at a level of 45 A. This set the stage for the investigation to follow. Page 4

5 An Initial Hurdle to Overcome Difficulties soon arose as we began to implement what we felt was a methodical approach for pinpointing the source of the ground fault problem. The idea was to determine the residual current at the entrance to each power distribution panel. Encircling all the phase and neutral conductors with a core-balance or window type CT would provide us with this essential information. It soon become evident, however, that such a task was impossible because of physical constraints and the wiring layout. The idea of using individual CTs wired using the Holmgren method proved to be an attractive alternative. The configuration is illustrated in the sketch below. On a short notice we secured four 1000:5 CTs with a window Holmgreen Method 76 A 76 A Zero Sequence 750 kva 76 A Ground Return opening of 4 to 6 each with a burden capacity of 5 VA. Terminating the Holmgreen connection into the current jacks of a digital multimeter proved satisfactory as we were able to duplicate the current reading from the ground fault relay meter in the grounding electrode path. For this hook-up to work, the ammeter section of the digital multimeter must have a low burden voltage. The Model 83 Fluke Meter meets this requirement when set to the 4 A or 10 A range. In the photo at the right, the CTs are bright orange and the test leads are yellow. Applying this technique to the main switchboard indeed confirmed that 76 A of current failed to return to the transformer via the phase and neutral conductors. Page 5

6 Working Our Way Downstream From the Main Switchboard We now applied the previous procedure to the SP200B sub-panelboard. Again, we were able to measure a residual current with magnitude of 62 A, this being a clear indication that a portion of the fault lay downstream of this panel. We proceeded onward with the tedious task of narrowing our search to one or more of the slot/lighting panelboards that derive their supply from the SP200B panel. To our surprise, each one of these panels got a clean bill of health. We now had to turn our attention again to the SP200B panel and question our earlier assumption that the panel itself was clean and free of any defect. A visual inspection revealed that the neutral bar was in fact solidly bonded to the panel which, as was shown on page 4, is also bonded to the. Removing this bond did indeed clear the fault. Not all was well, however, since the ground fault relay meter in the grounding electrode conductor leg now displayed a current of 47 A. SP200B Sub-Panelboard Line Side Current = 62 A No residual current on the Line Side of the adjacent SP (42 branch) Slot/Lighting panelboard. Note the yellow Hioki clamp-on meter. The picture to the right shows a Blue Chip technician removing the neutral-to-panel bonding jumper. Needless-to-say, safety precautions were exercised since this troubleshooting sequence was conducted with the panels in an energized state. Page 6

7 Back to the Main Switchboard The test equipment was moved back to the Line Side of the Main Switchboard. With the four CTs again connected in the Holmgren configuration we were able to confirm the validity of the 47A reading from the ground fault relay meter. The sketch below illustrates the situation. With a suspicious eye toward discounting the obvious we noted as anticipated the existence of the main bonding jumper. The tip-off, as to the most likely source of the problem, was revealed when we noted that the magnitude of the current flowing through the main bonding jumper was the same 47 A displayed on the ground fault relay MAIN SWITCHBOARD SP200C neutral One 1000 A, one 600 A, and three 150 A circuits main bonding jumper ground bus Page 7

8 Getting Closer The type of behavior we observed seemed to indicate that perhaps, even though a visual inspection seemed to indicate otherwise, the ground bus was not after all. If this is the case, some of the neutral current will be re-directed as shown in the sketch to the right. It remained to answer the question: How do we prove this? Working on a live system reduced our options. The idea was to first connect a heavy gauge wire from the ground bus to the grounding electrode conductor as shown in the sketch below, MAIN SWITCHBOARD SP200C MAIN SWITCHBOARD SP200C neutral One 1000 A, one 600 A, and three 150 A circuits neutral main bonding jumper One 1000 A, one 600 A, and three 150 A circuits ground bus main bonding jumper non- ground bus and then to disconnect the main bonding jumper. This precautionary step was necessary in the event a ground fault occured after the main bonding jumper was removed; the temporary connection provides a path back to the supply for the ground fault current. The ground fault current magnitude dropped to 8 A after removing the main bonding jumper thus confirming our suspicion that the ground bus bar was not as intended by the design engineer. Still, the 8 A figure was unacceptably high. Page 8

9 Success at Last After additional tests we were able to determine that the fault was originating in one of the circuits leaving the SP200B sub-panelboard. This particular circuit feeds a distro box with label SP210 that is located underneath the ceiling as shown in the photo at the right. Several quad receptacle boxes, located in the casino gambling area, are fed from this distro box. Each quad receptacle box supplies power to four slot machines. We quickly observed after opening the quad receptacle outlet box that the neutral was grounded. The ground fault current back at the transformer dropped to 1 A after removing this neutral-to-ground connection. The photo below shows the hardware and test equipment. As indicated earlier on page 3, there are a total of five step-down transformers in the casino. The knowledge gained earlier was applied in earnest to the task of clearing ground faults uncovered in various circuits powered by these transformers.. Page 9

10 Future Recommendations Testing the entire system can be a hazardous task. The continuous gambling operation requires testing to be done under power. Shutting down 2300 slots is not an option. It is highly recommended to install a BENDER RCMS Fault Finding System. This will enable technicians to quickly identify and pinpoint the source of ground fault problems without having to open electrical panels and exposing themselves to shock hazards. In addition to increased shock hazard exposure, an arcing ground fault current that is allowed to persist indefinitely is a potential fire hazard that can cause considerable damage and may result in an extended power outage. An insulation failure is one source for an arcing fault. Causes of insulation breakdown include moisture, dirt contamination, foreign objects, insulation deterioration and physical damage from mechanical stresses and insulation puncture. The fault finding process often brought us within inches of live conductors. Undoubtedly, ground fault events will occur repeatedly throughout the life of the facility. As such, again and again the Blue Chip crew will have to navigate around live distribution circuits to resolve these ground fault issues. It would be of great benefit to give them the right tools to identify these circuits automatically without the fear of litigation and personal injury. The BENDER RCMS460 scanning system is the solution. Contact: Torsten Gruhn (800) info@benderrelay.com T hese symbols are important. ISO9001:2000 is a globally renowned standard with certifi cates awarded in over 150 countries. It implies the application of a leading quality management system to product manufacturing. The CE mark is a declaration by the manufacturer that the product meets all the appropriate provisions of the relevant legislation implementing certain European Directives. In particular, amongst other standards, these products meet the Electromagnetic Compatibility (EMC) requirements of IEC 62020: which governs the emissions of the products and their immunity to interference from other sources. Page 10