T&D World Magazine

BG&E Makes a Smart Grid Case

Placement of fault circuit indicators is optimized by using internally developed methodology.

Fault circuit indicators (FCIs) have been part of utility distribution networks for more than two decades. After a fault occurrence, utility personnel would patrol and inspect the FCIs to see if the devices had detected the fault. Following visual identification, they would inspect the feeder either downstream or upstream from the FCI, depending on the FCI state.

The majority of FCIs installed during the past two decades did not have any communications device associated with them, so the information detected by those FCIs was not available to the system operator. In addition, those devices did not differentiate between permanent and temporary faults, and sometimes they would reset before utility personnel had the chance to inspect them following the fault.

Advances in technology and communications have resulted in the development of FCIs with advanced features and communications capabilities. Today, FCIs are designed as low-cost sensing devices, which are easily deployed on the utility's power system network. These devices can be installed on the power lines using a standard hot stick, which locks them into place. This is a desirable feature because the sensors can be easily removed for maintenance or for placement at another position.

Communication Is Key

During the last decade, many utilities have incorporated different distribution automation (DA) methods and applications to automate their distribution system operations. FCIs help utility personnel to identify a fault location faster, thus improving system reliability. In 2011, Baltimore Gas and Electric (BGE) developed a smart grid distribution system pilot project on six test feeders with three main objectives:

  • Implement volt/volt-ampere-reactive control to reduce energy consumption.

  • Install capacitor bank controllers with two-way communications to improve the reliability of field capacitor.

  • Reduce the customer average interruption duration index (CAIDI) by installing FCIs with remote notification.

Currently, BGE uses different types of FCIs on its distribution network. Some FCIs are an integral part of equipment installed on the network (padmounted switchgear), while some FCIs are placed in strategic locations to help locate faults faster (underground cable). However, the vast majority of these FCIs only have local fault indication. To improve CAIDI, BGE added FCIs with communications modules and remote notification as part of a smart grid project to investigate whether using these devices with advanced capabilities could improve the reliability of the distribution system.

To implement the installation and commissioning of FCIs on six test feeders successfully, BGE developed a three-step process that consisted of defining which FCIs to install on the system, developing new methodology for the optimum number and optimum placement of FCIs, and performing cost-benefit and discounted cash-flow analysis.

Choosing the FCIs

Several types of FCI devices are available on the market today, and they all differ based on the amount and type of information that can be sent back to the utility office, and the communications networks used to transmit and receive the data.

During the initial process, BGE evaluated several FCI features. Two of the most important features were power requirements for the reliable operation of FCIs and the optimum amount of information from FCIs needed for successful fault location.

The BGE system has more than 600 feeders with implemented fault detection, isolation and restoration (FDIR) schemes, and five out of six test feeders are equipped with FDIR. Depending on the design, FCIs can require primary currents in excess of 30 A for normal operation, so placement of these devices becomes challenging on the feeders with FDIR scheme because optimum locations for the placement of FCIs do not carry such load throughout the whole day. The optimum amount of information assessment was needed so system operators would not be overwhelmed with the unnecessary information FCIs are capable of reporting.

For the purposes of the smart grid pilot project, BGE used FCIs manufactured by two different companies: Cooper Power Systems for overhead feeders and GridSense for both overhead and underground feeders. BGE had been using Cooper Power Systems' OutageAdvisor solution for several years, except the units installed in the field were capable of local indication only.

Leveraging the existing Yukon platform used for capacitor bank control, the OutageAdvisor solution with remote notification capability was used, communicating on the Verizon network. These FCIs detect the high rate of change of current followed by the loss of current, which triggers the event. After the protection devices have a sufficient time to operate, the sensor determines the nature of the fault (temporary versus permanent) and transmits this data to the Yukon server. In addition, there is a local fault indicator that provides an indication to the field crew that the sensors have seen the fault. The notification about the event occurrence is transmitted via short-message-service text message and e-mail.

GridSense provided solutions for the monitoring of both overhead and underground feeders. On overhead feeder, BGE used a solution called LineIQ along with cell communications. LineIQ sensors are easy to install, and offer current and voltage measurement, fault direction, fault waveform, load profile, power factor, line status and condition, ambient and conductor temperature, and time-stamped event recordings. This information proved to be extremely helpful when, during a fault occurrence on one of the feeders that had several LineIQ sensors, analysis of the data helped to locate the fault location within a few hundred feet of the true fault location.

On underground feeders, TransformerIQ was used along with On-Ramp Wireless communications. The advantage of the TransformerIQ solution was the current transformer supply draws power from insulated cables without any need for a battery as a power source. TransformerIQ features include the reporting-by-exception of faults and outages, three-phase load monitoring with additional feeder points, 12-minute average or root-mean-square load recording, voltage and power-factor alarming on surge/swell, magnitude and phase indication on faults >2 cycle, and cable temperature and transformer oil temperature monitoring.

BGE placed the units inside padmounted switchgear and padmounted transformers. The transformer oil temperature monitoring was an excellent additional feature. During the installation and commissioning process, the oil temperature sensor indicated higher-than-normal readings, which helped to identify the problem with the transformer, which was repaired before any damage was done to it, and the industrial customer did not experience any unwanted interruption.

Optimum Placement of FCIs

To determine the optimum number and locations for the placement of FCIs on the six test feeders, BGE developed a new methodology along with existing practices and knowledge. When the fault happens on the distribution feeder, protection and control devices operate to isolate the fault. Next, DA reclosers that are part of the FDIR scheme operate, thus minimizing the number of customers affected by the outage. The optimum placement methodology will be explained on the 7639 feeder from the Dover Substation. This feeder has two DA in-line, normally closed reclosers and three normally open tie reclosers.

The first step in the process was to gather historical CAIDI data for each of the feeders, along with overhead and underground cable failure data. The average of the last five years of CAIDI data was used as the baseline for the customer average outage time. Then the feeder was divided into zones based on the number of in-line DA reclosers. Failure data was collected on this feeder and each of the zones on the feeder, separated into sections defined by nodes.

Each section was characterized by the number of miles, percentage of overhead and underground miles, and number of customers in each section. Finally, a CAIDI-equivalent formula for average outage time per customer per feeder was developed as a function of the number of zones, the customer outage time in a particular zone, the probability of a fault occurrence in each zone as well as the total number of customers on the feeder.

The customer outage time in each zone was calculated based on time to initiate the ticket after the fault occurrence, line crew response time, feeder inspection time and time to fix the fault. Initially, it was expected FCIs would have a major effect on the improvement of feeder inspection time, which they did, but throughout the pilot project, it was observed a few minutes also were saved for the time to initiate the ticket.

Based on historical reliability indices, number of faults, causes of faults and number of protective devices in the zone, analysis was done to obtain the probabilities of fault occurrence in each zone. Normally, historical data for the location of each fault is not readily available, and even if it were available, it would not provide enough information for this type of analysis. But that would not be the case anymore. For any desired number of FCIs, analysis was done using the Monte Carlo analysis and random mutation hill climbing approach to calculate the expected average outage time, predict CAIDI and perform the cost-benefit analysis and discounted cash-flow analysis.

First, analysis was done without any FCIs deployed on the system to calculate the projected average outage time (AOT) and compare it to the five-year average CAIDI value. This was necessary because the analysis was done by using the average values for a certain number of parameters, such as failure rates for the overhead or underground cable, or response and repair times), rather than probability distribution. The difference between the two values of 0.139 p.u. was added in subsequent analysis for predicting the expected CAIDI for any desired number of FCIs. Subsequently, Monte Carlo analysis was done using probability distributions rather than average values for the same parameters, and the analysis resulted in smaller deviation between the projected AOT and CAIDI.

Cost-Benefit Analysis

From the results of the analysis, the greatest benefit of adding the FCIs on the system was reduced CAIDI and increased customer satisfaction realized because of CAIDI reduction. Even though it is extremely complex to put a dollar price on customer satisfaction, utilities can only benefit from increased customer satisfaction and lower CAIDI. The analysis also showed, for the 7639 Dover feeder, the optimum number of locations for FCIs is four. The analysis predicted a reduction in CAIDI of 0.324 p.u. Installing FCIs in the fifth location resulted in minimal CAIDI improvement (0.011 p.u.), which could not be economically justified. In any case, the optimum number of FCIs on the feeder can be calculated by using the cost-benefit analysis and principle of diminishing returns.

For any investment in a new technology or application to make sense, it is necessary to estimate the potential benefits of placing a particular number of FCIs on the distribution system. Besides the increased customer satisfaction and CAIDI reduction, there are generally two other benefits for the utility business: additional revenue realized from reduced CAIDI reduction and cost reduction realized from nonutilization of the distribution line crew, as a result of the reduced patrol time during the fault location process.

Revenue realized from reduced CAIDI depends on the number of customers per feeder, customer average load, CAIDI reduction, the difference in price and cost of electricity, feeder SAIFI and number of feeders. The reduction in patrol time associated with the fault location automatically reduces the costs associated with the distribution line crew. Total avoided cost depends on the CAIDI reduction, feeder SAIFI, cost of the truck and line crew, and number of feeders. The total benefit is equal to the sum of the revenue and avoided cost. The cost associated with FCI implementation consists of the cost of FCIs, installation cost, yearly maintenance cost and useful life of FCIs. Net system savings can be calculated as the difference between the total benefit and total cost.

Cost-benefit analysis results in four optimum FCI locations using the law of diminishing returns (a marginal benefit of installing the FCIs at a fourth location of 0.373 is greater than its marginal cost of 0.281). Implementing FCIs at a fifth-optimum location results in a marginal benefit of 0.086 compared with marginal costs of 0.281. In addition, using the value for total capital investment in FCIs and cost-benefits analysis, it is possible to obtain a net present value of FCI implementation for any desired implementation scenario. [Note that not all North American utilities have the same cost-recovery methods — BGE is not a vertically integrated utility and cannot benefit from additional revenue realized from reduced CAIDI — so parts of this analysis might not pertain to every utility.]

Future Implementation

Going forward, BGE will explore the possibility of integrating FCIs into its Oracle outage management system, so based on the reported field data during the faults such as FCI status and current and voltage fault levels, additional analysis such as impedance-based fault location or load-flow analysis can be performed within the outage management system to identify potential fault locations.

Paul J. Frey ([email protected]) is the manager of Baltimore Gas and Electric's smart grid distribution system pilot project. Frey started with BGE as a test engineer and has held positions in gas and electric operations and planning, strategic customer planning, asset management, system planning and equipment engineering. He received a BES degree from Johns Hopkins University and a master's degree in engineering administration from George Washington University.

Aleksandar Vukojevic ([email protected]) joined Baltimore Gas and Electric as a senior engineer in the smart grid distribution system pilot project department. Previously, Vukojevic worked as a lead power systems engineer for smart grid technologies at GE, a field test engineer and transmission planning engineer at Georgia Power, and smart grid engineer and system protection and controls engineer at BGE. He received a bachelor's degree in applied mathematics from Kennesaw State University, BSEE and MSEE degrees from the Georgia Institute of Technology, and a MBA degree from Robinson College of Business at Georgia State University. He is EIT certified.

Michael S. Smith ([email protected]) is a lead engineer/work leader with Baltimore Gas and Electric's automation and technology unit. Smith has 30 years experience in the utility business specializing in system protection, substation integration and supervisory control and data acquisition.

Table 1. Projected optimum average outage time and CAIDI as a function of deployed fault circuit iIndicators.

Number of FCI locations 0 1 2 3 4 5
Projected average outage time (p.u.) 0.861 0.745 0.655 0.582 0.537 0.526
CAIDI (p.u.) 1.000 0.885 0.794 0.721 0.676 0.665
FCI 10-12 10-12 2-3 2-3 2-3
20-21 12-14 12-14 12-14
20-21 20-21 20-21
22-23 22-23

Table 2. The following results can be obtained for cost-benefit analysis for a certain set of assumptions regarding the values in Table 1.

Number of FCI sets Revenue Avoided cost Total benefit Total cost Net system savings
1 0.433 0.516 0.948 0.437 0.511
2 0.774 0.922 1.696 0.719 0.977
3 1.050 1.250 2.300 1.000 1.300
4 1.221 1.453 2.673 1.281 1.392
5 1.260 1.500 2.760 1.562 1.197

Companies mentioned:

Baltimore Gas and Electric | www.bge.com

Cooper Power Systems | www.cooperindustries.com

GridSense | www.gridsense.com

On-Ramp Wireless | onrampwireless.com

Verizon | www.verizon.com

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.