The Big Easy?
Editor's note: This month we take a left turn, so to speak, into the civil world, examining an application of automated switchgear, not in an industrial or commercial setting, but on a lake-spanning bridge. Those who have visited New Orleans know that Crescent City citizens draw their drinking water from a vast body of water known as Lake Pontchartrain.
Those who have visited New Orleans know that Crescent City citizens draw their drinking water from a vast body of water known as Lake Pontchartrain. And even though New Orleans is renown as "the Big Easy," those living outside the city have the onerous task of traversing this lake across a 24-mi.-long causeway.
In fact, the structure—the longest overwater bridge in the world—has some 30,000 vehicles cross over it every working day. To better accommodate commuters
, numerous cell phone towers and variable message boards have been erected along the bridge's length. And the need for reliable electric service for these communication devices is key to the bridge's owner-operator, the Greater New Orleans Expressway Commission, as the bridge serves as a major hurricane evacuation route.
Recognizing these factors, the commission initiated a $79-million capital improvements master plan in late 1995. One of the largest projects in the rehabilitationore. Each power supply was generated by separate utilitie
The open point between the two feeders was the bascule drawbridge located at the 16-mi. point of the structure. To provide manual sectionalizing in the old scheme, switchgear was located at each of seven traffic crossovers that connect the 80-ft. gap between the spans. These connections also function as emergency pullover areas.
But with the addition of the message boards and particularly the cellular antennas, the power loading levels were far beyond the original capacity. In fact, the voltage drop increased to 40% at the end of the feeders—insufficient to operate the four 100-kW motors that powered the bascule. As a stop-gap measure, a generator was installed to power the motors.
Besides an undersized power supply, the bridge's high-voltage cable itself was nearing the end of its useful life. The aerial cable, strung to a messenger wire under the bridge, began to experience an increasing number of faults due to insulation degradation.
Furthermore, in the original electrical system, there was no mechanism, either automatic or manual, for tying the two feeders together. In the event of a fault along one of the feeders, maintenance personnel could isolate the faulted section using the switchgear located at each of the seven crossovers. Service could then be restored up to the faulted section, but not beyond. If an outage occurred on the utility feeder, the load supplied by that feeder would remain out of service until the utility's repairs were complete.
New and improved
In assessing the situation, four essential components were identified:
Increase the voltage level of the supply feeders.
Replace all high-voltage cable.
Provide the capability to tie between the north and south feeders.
Automate fault isolation and restoration.
For the first part of this undertaking, the voltage supply feeders were increased to 24.9-kV with wye configurations, coming from both the north and southshore stations. One problem in working with two different utilities was that they have different distribution voltage standards. The issue, however, was addressed by means of step-up transformers.
Additionally, at each shore a loadbreak switch with a resettable vacuum fault interrupter (RFI) has been installed to provide fault isolation and sectionalizing between the utilities and the bridge. Along the bridge itself, nine other switchgear units have been installed to provide fault isolation and restoration. These units are also configured with the loadbreak switches along the mains. Depending on the specific loading, one or three 3-phase taps are protected by the RFIs.
Controlling the system, of course, was another major factor. A supervisory control and data acquisition (SCADA) master station, installed at the Causeway Commission's south shore maintenance facility, communicates to each switchgear unit loc
ation via a second fiber-optic loop using a Mirror Bits DNP 3.0 protocol. A full-graphic operator interface with event recording and alarming is provided. Maintenance personnel can put the automated restoration system into a manual mode to temporarily modify the configuration or to return the feeders to normal after some automatic operation (see "SCADA at a Glance," below).
Call in the marines
While the solution certainly breezes along from a reader's perspective, the project was far fro m easy. In fact, having to work 24 miles directly over water made the job unique.
The most significant of these challenges involved the following:
Voltage level. One of the first decisions involved a determination of the most efficient voltage level to supply the bridge. A number of engineering studies and cost-benefit analyses were conducted to decide between 14.8-kV and 24.9-kV supplies. At 14.8 kV, the maximum steady-state voltage drops were calculated to be 7.5%, while at 24.9 kV they were calculated to be 1.5%. Both compare favorably to the previous 40% voltage drop. However, the 7.5% voltage drop proved excessive for a new system that may be required to take on additional loading in the future. The cost increase of moving to 24.9 kV was determined to be reasonable.
Lack of real estate. The question of installing anything on a 24-mi.-long bridge is "Where do you put it?" A number of options were evaluated before designing the final solution. As with any replacement project, the existing facilities must remain in service during the construction, so the location of the old switchgear and aerial cable was not a choic
e. Besides, the existing 5-ft. x 7-ft. x 4-ft. enclosures located at the crossovers were too small. An expansion of the crossovers was considered, but ruled too costly. Another option was the installation of a new platform between the spans. Getting a crane to that location to drive the pilings, however, was not possible. Installation of the switchgear below the bridge was considered as well, but that option was also infeasible because of instances of high water.
The solution lay in installing structures off the side of the bridge itself to hold the new automated switchgear. Complicating the choice, however, was the requirement that these structures be located 30 ft. away from the bridge to allow for future expansion. As a result, four 90-ft. pilings were driven into the lakebed to support a 15-ft. x 20-ft. concrete pad, upon which sits a fiberglass electrical distribution vault.
Installation of new cables. As in placing the new switchgear, the same issue existed with the new high-voltage cable and new fiber-optic control cable. Various options were explored, including installation of conduit or hanging the new cable from a messenger wire. Neither solution, however, facilitated future expansion. In the end, a cable tray was designed that resides under the overhang off the side of the bridge.
Automated restoration. To maximize the availability of the bridge, the project included a performance specification requiring the new high-voltage electrical system to be fully automated, both for fault isolation and for restoration of service (see "Breaking Down the Automated Control System," p. 46).
FAT Tuesday testing
Developing unusual solutions was just the first part of the job. Extensive testing was also required. Fortunately, this was accomplished within a short nine months leading to factory acceptance testing (FAT) in September of 2001. Gulf Engineers and Consultants witnessed the work at the Canada Power Factory—the project's electrical supplier and system integrator—where all switching and fault scenarios in normal conditions, as well as fault scenarios in abnormal conditions, were tested thoroughly, including controllers, software and master station functions. The system performed well in all aspects, providing a high degree of confidence in the long-term success of the project.
Now, whether it's the daily commute or folks evacuating from a tropical storm, causeway officials are confident traffic—and power—will keep rolling.
Breaking Down the Automated Control System
To support the complex control functions of the automation system, it was deemed critical that there be distributed logic and peer-to-peer communications between the controllers.
Programming a central control system with all of the logic, variables and possible combinations necessary would be difficult to say the least. It would need to direct the isolation and restoration of a high-voltage electrical loop with 11 switching devices—such that it would operate reliably and within a reasonable amount of time.
Additionally, future modifications to account for new switchgear, or perhaps only temporarily to account for some abnormal condition, would be costly. Distributed logic, on the other hand, offered an elegant solution involving the following:
Automatic transfer control. This function allows for the restoration of the entire bridge in the event of a feeder loss. At each end of the bridge, the last switchgear unit is designated as a "shore switch." In this mode, upon loss of one utility feeder, the opposite shore switch will open, isolating the entire bridge from the out-of-service feeder. The normally open tie switch will then close, restoring service to the entire bridge. Upon restoration of the utility feeder, the system is designed to return to the normal configuration automatically in an open transfer (break before make). In the case of losing both feeders, the shore switches will automatically open, but the tie switch will not close.
Fault isolation scheme. This function automatically isolates any fault on the bridge, restoring the maximum possible service. The system differentiates between faulted cables (10 schemes), faults within the switchgear (11 schemes), faults between the control transformers and switchgear (11 schemes). The RFI then taps the system and deals with each one accordingly, isolating the fault and restoring the maximium power load.
Fault scheme in transfer mode. One of the technical strengths of the scheme is that it can operate beyond the first contingency. This function provides fault isolation in the event of a fault when the system is abnormal, whether due to the automatic transfer control, the fault isolation scheme or simple manual switching.
Paralleling feeder protection. A requirement set forth by the utilities was that the supply feeders must not be paralleled under any circumstances. The automation system will not allow any automatic or manual switching to occur that causes the feeders to be paralleled. All return-to-normal switching is achieved by using an open transition between the feeders.
Inside the Switchgear
Each switchgear unit is equipped with the following:
Switch controller cabinet
Single-mode fiber-optic modem/transceiver
Resettable fault interrupter (RFI) control box
2-Position motor operators for switch and RFI ways
3-Phase 100:0.5-amp current transformer per RFI way
3-Phase 600:5-amp current transformer for relaying way No. 1
Internal 14,400:120-volt potential transformers
Color-coded pressure gauge (0-15 psig) and 2-psig pressure switch
SCADA at a Glance
The following data and functions are generated by the system's supervisory control and data acquisition master station:
Phase current (amps) and phase voltage (kV) at each switchgear unit.
Status indication per unit for: remote or local mode, voltage restoration mode, switch position, resettable fault interrupter (RFI) position.
Alarm indication per unit for: relay, loadbreak switch pressure, battery power, AC power, mirrored-bits communication and paralleling lockout.
Remote control per unit for: switch open/close control; RFI open/close control; relay front panel target reset.
Alarm log and printer for event logging.
Backup tape drive.