Comment

1 August 2006

When Power Takes a Powder

North Carolina power plant reduces lightning damage with dense, low resistance ground grid installation

By Thomas M. Toms

In the next few years, manufacturers will spend precious time and money designing the perfect automation system, but some will fail unless they consider power and its consequences as part of the whole. Designing an adequate power supply system for a control system is only a small part of properly designing the entire electric distribution system for a facility. It’s critical to consider the potential for outages, surges, sags, and high voltage transients.  

A North Carolina power plant learned these lessons the hard way after repeated damage from lightning strikes to the instrumentation systems on a Richmond County combined cycle power plant. The strikes triggered Progress Energy Carolinas, Inc., to investigate the plant grounding system. The goal was to meet personnel safety requirements and prevent damage to instrumentation systems from future lightning strikes.

A survey of plant underground utilities revealed a fragmented ground grid wasn’t installed as designed. To guide remediation efforts, computer models reproduced the as-designed, as-found, and as-remediated grounding system. Looking for clues to instrument failures, engineers analyzed graphic representations of the step-and-touch potentials for compliance with personnel safety standards and models of the ground potential rise (GPR) resulting from simulated lightning strikes to selected plant structures.

Grounding system functions

The grounding system for a generating station provides three important functions: personnel safety from electrical fault conditions, lightning protection, and a termination point for the shielding of instrument cables. The grounding system normally consists of bare copper cables, buried 12-18 in. deep, interconnected in a rectangular grid pattern. Ground rods and ground wells often improve the grounding system. Stingers, or pig tails, connect above-ground equipment to the ground-ing system.

Located in the Sandhills region of North Carolina, the Richmond County Energy Complex has high soil resistivity due to the sand content of the soil. Plant engineers originally planned the Richmond County Energy Complex as a four-phase construction process. The ultimate plant would be four 2-on-1 combined cycle power blocks, each with about 500 megawatt (MW) capacity. They commissioned Phase I in 2001 as four simple cycle turbines that could be converted to combined cycle operation in the future. The Phase I grounding system used a 20’ x 20’ ground grid and two ground wells. The proposed Phase II design was a 50’ x 50’ grid with 20 ground rods. Phase II added a 2-on-1 combined cycle power block plus one more combustion turbine and began commercial operation in June 2002. Construction of Phases III & IV has not yet begun.

See the “Phase I & II, Initial Grounding System” figure for the final Phase I & Phase II ground grid design. 

The crew anticipated grounding problems since substations and other plants in this region have a history of lightning damage. In the summer of 2002, prompted by recent lightning damage to equipment at several sites, primarily in the Carolinas, the combustion turbines operations requested an assessment of the grounding system at all combustion turbine sites. This included measuring the soil resistivity and ground grid resistivity as well as point-to point resistance, a model of the ground grid and simulated step-and-touch potentials, GPR, and recommendations for improving the grounding system.

Grounding system remediation

Richmond plant personnel reviewed all correspondence related to the grounding system and decided to install wet ground wells to lower the resistivity of the grounding system. This recommendation appeared in the original assessment of the proposed Phase II grounding system but never saw implementation. While installing the wet ground wells, plant personnel were unable to locate ground grid cables at the locations indicated on plant drawings. They continued to trench well beyond the area where the cables should be located with no success. This occurred at several locations while trying to locate buried ground cables.

Since plant personnel were not able to locate several ground grid cables and had identified nothing else to explain the lightning damage, they formed a project team to assess the grounding sys-tem. The team decided to determine, to the extent possible, the location of buried ground grid cables and to verify the quality of the electrical connection for the equipment grounds.

They contracted a company specializing in identifying underground utilities to produce a map of the buried ground grid. The survey was limited to open areas since the surveying equipment would not work reliably in areas of strong magnetic fields or adjacent to large metal structures. This restriction meant they could not survey the area between the units or areas within 10 feet of metal buildings. The contractors induced a radio frequency signal on a known ground cable and used a receiver to trace the signal to produce data for the map. In instances when they doubted the identified signal was a ground cable or some other buried cable, they excavated a small hole using a truck mounted vacuum machine to remove the dirt. Using this approach helped establish a high confidence level for the buried ground grid map. 

Based on the results of the underground survey, they developed a plan to remediate the buried ground grid, adding new cables where voids existed. Where they noticed cut ground grid cables, they planned splices. After identifying the remediation points, they again employed the survey company to identify and mark cable endpoints. Trenching and hand digging exposed cable connection points. And they installed nearly 2,500 feet of new cable and exothermically connected it to the existing ground grid.

Ground grid modeling

Because of lightning damage at several plants, including the Richmond Phase I units, Progress Energy elected to test the grounding system at all combustion turbine sites. Since Phase II at the Richmond plant was under construction at that time, they hired a consultant to model the proposed Phase I and Phase II ground grids. The modeling was to assess grounding sys-tem performance, recommend improvements to the grounding system, and identify any personnel safety issues. A model of the as-designed ground grid already existed from the initial assessment.  

Comparison of the GPR graphics shows reduction of the voltage rise to an acceptable level for the majority of the plant site. They added additional ground cables in selected areas to further reduce the GPR around the administration building and the fuel forwarding area. Ground resistivity measurements from October 2000 and September 2002 showed values of 1.24 Ohm and 1.44 Ohm, respectively. Soil resistivity was 3113 Ohm-meters to a depth of 98 feet and 698 Ohm-meters below 98 feet. These values saw use in the computer models.

Damage to instruments

Damage to electronic devices occurs when the voltage difference between circuits exceeds tolerable limits. If the input terminals of a 1-5 volt analog card are subjected to a 50 volt transient (relative to the instrument ground), the circuits will probably be destroyed. However, if the instrument ground is also subjected to the same 50 volt transient, the electronic circuits will survive the transient. It is the voltage difference that destroys electronic circuits, not the absolute magnitude.

An equipment fault or lightning strike will produce substantial current flows throughout the grounding system, resulting in a voltage gradient (the GPR) across the plant site. The peak voltage depends on the ground mat resistivity, the ground mat-to-earth resistivity, and the current intensity. The GPR will represent a mountain peak at the point of entry.

The voltage difference an electronic circuit sees will depend on the physical location of the instrument versus the location of the electronic controller. As part of Progress Energy’s assessment of grounding systems, the consultant modeled lightning strikes to four structures at the Richmond County plant.

Verifying equipment grounds

In addition to remediating the ground grid, you need to make sure equipment grounds connect to the grounding system by a low resistance connection. To verify equipment grounds over 850, we tested connections to measure the resistance to the ground grid. Since the generating units could be online while we did the testing, and since leakage  current in a ground lead could be  dangerous to testing personnel, we developed a procedure to test for high-leakage current, using a clamp-on current/Ohm-meter to measure leakage current prior to disconnecting any ground cables. We recorded resistance and leakage current for each cable. Leakage current had to be less than 100 mA before we disconnected a ground lead. Where possible, we disconnected bolted ground leads and measured the resistance to the ground grid with a digital low resistance ohmmeter (DLRO). In a few cases, the clamp-on Ohm-meter produced erroneous readings, and the DLRO was the only reliable measurement. For the Phase II units, the testing identified 83 remediation points with missing ground cables, high resistance connections, or excessive leakage current.

At the end of the project, the team realized it was possible to minimize lightning damage to equipment by installing a dense, low resistance ground grid. The grid resistance ground grid is more important than the grid-to-earth resistance since the flatness of the GPR is more important than the voltage magnitude. While the simulated voltages are higher for a simulated strike to the Unit 7 HRSG using the as-found ground grid, the differences are small. Also, maximum GPR for other simulated strikes is almost equal or higher for the Model B grounding system. It isn’t proven the compromised ground grid is solely responsible for the lightning damage. It is possible a combination of items contributed to lightning damage. Since we enhanced the ground grid and repaired grounding connections, the plant has not experienced any equipment damage due to lightning strikes. 

About the Author

Thomas M. Toms is a lead technical project manager at Progress Energy Carolinas, Inc., in Raleigh, N.C.