Predicting the Grid

Author:
Edward H. Kennedy, Tollgrade Communications

Date
12/13/2013

 PDF
Using the next-generation of sensors and analytics to manage the Smart Grid

From the North East blackout of 2003 to last year’s Super Storm Sandy, we’ve learned the hard way that our 100-year-old-plus electricity grid is lacking the intelligence to effectively respond to power outages and disturbances. Here are a couple of key points to consider that call out for the need for a paradigm shift in the way we think about the grid to successfully improve reliability.

Unmonitored distribution

First, there are millions of miles of unmonitored electricity distribution and subtransmission lines that are responsible for delivering power to businesses and residential customers. The unmonitored distribution network is also where most of the world’s outages occur and they are on the rise. For example, electric disturbances in the U.S. have seen a dramatic 265% increase in the raw number of major outages occurring since 1984 with an annual cost as high as $188 billion to the U.S. alone. In North America, our distribution grid is the largest part of a utility’s network, spanning over six million miles and is estimated to be eight times larger than transmission.  And, according to the Edison Electric Institute (EEI), it’s also where 90% of the outages occur.

Aging infrastructure

Second, our grid is an old and aging infrastructure. In the U.S. its estimated 70% of transformers are over 25 years old and quickly approaching their end of life, while in extreme heat and weather, many are being stressed well beyond their productive limits. This aging, stressed infrastructure also contributes to more outages as failing equipment causes an estimated 25% of outages. With the DOE projecting summer peak demand to increase by almost 20% during the next 10 years and demand expected to double by 2050, many believe we have an opportunity now to make investments to modernize our grid to avoid some of the worst case scenarios being prophesized now.

Grid dependency

Finally, it is a fact that our modern economies, and GDP, are becoming more “grid dependent.” When a factory has to shut down because of a power quality issue, it costs businesses in the U.S. as much as $25 billion dollars a year, according to the Electric Power Research Institute (EPRI). With better, real-time monitoring, utilities will be able to deliver the power quality required for today’s modern economy.

All of this requires monitoring of the medium voltage grid. To monitor this entire distribution network, utilities need more than smart meters.  They need smart-grid sensors that are inductively powered, can provide visibility to low amperages, and can integrate with wireless systems to transmit real-time grid health data back to the utility back-office for additional analysis, trending and alarming. Now, with more affordable and flexible communication options including Wi-Fi, WiMAX and cellular, this vision is a reality. These Wi-Fi or Cellular sensors can be installed in minutes and can span the millions of miles of unmonitored power lines from the substation to sectionalize feeders and communicate real-time grid health data back to analytics software, which alerts utilities to problems for faster resolution (see Figure 1).

Click image to enlarge

Figure 1: Sensors installed on power lines communicate real-time grid health data.

Tollgrade Communications is the first to market with this type of an integrated offering. With a global footprint and over 25 years of experience in providing cutting-edge network assurance solutions, Tollgrade has built a reputation for improving the reliability and operational efficiency at the world’s largest utilities and telecommunications providers, helping them to reduce customer down time, modernize their networks, reduce their carbon foot print, take on new sources of renewable energy, and recover lost revenue (see Figure 2).

Click image to enlarge

Figure 2: Improved oversight enhances reliability and operational efficiency

Tollgrade provides the only solution that takes their smart grid sensor data through four layers of analysis to build a foundation for Predictive Grid℠ analytics. This is a critical requirement to improve overall accuracy and effectiveness of grid data. For example, in head-to-head field trials, older generation Fault Current Indicators (FCIs) created as many as four false alarms to every true outage event confirmed by Tollgrade. False positives are detrimental to utilities, especially as they integrate the data from smart grid sensors and FCIs with other back-end systems such as Data Historians, Outage Management Systems, and Distribution Management Systems. Tollgrade offers the advantage that as the LightHouse SMS software “hands-off” this data, it can be trusted (see Figure 3).

Click image to enlarge

Figure 3: Multiple layers of analysis improve overall accuracy and effectiveness of grid data

Our rules-based, predictive analytics database takes advantage of waveform signatures that hold the clues to how outages and power quality events could be better handled in the future. This signature model has worked in other industries like the computer security industry, but has not been attempted before to improve grid reliability. By providing four layers of analytical processing, the software accurately detects and classifies faults that cause outages in real-time for example, permanent faults on main feeders and blown fuses on laterals. They can also detect troublesome power quality events like momentary outages. This enables utilities to perform outage restoration faster because they know more about the fault and how to repair it. For example, in two very large deployments in North America, the solution is detecting blown fuses on laterals 20 – 30 minutes before customers call to report the outage. All events are linked to map-view displays that make it easy for utilities to pinpoint the locations of problems for quick response and resolution (see Figure 4). 

Click image to enlarge

Figure 4: The ability to pinpoint the locations of problems enables quick response and resolution

By having a library of power quality and outage events combined with load planning data in one software package, utilities have the situational awareness they need to react to grid changes and faster. We believe one day they’ll even be able to predict future problems before they occur.  The software is also customizable and over-the-air upgradable – allowing the solution to grow and evolve as utilities face 21st century challenges and applications.

While this may sound futuristic, think back to the cell phone you were using just 5 or 10 years ago compared to the type of powerful technology you now have at your fingertips. The types of technologies utilities are beginning to embrace are putting them on a similar path from their current-state toward a future, predictive grid. 

A number of the world’s leading utilities, including Duke Energy and CenterPoint in the U.S., Toronto Hydro in Canada and Western Power Distribution in the United Kingdom are leading the charge with the deployment of smart grid Sensors. For example, Toronto Hydro, the largest municipal electricity distribution company in Canada, eliminated 550,000 customer outage minutes with sensors operating on two feeder lines alone.

Ivano Laboricca, their vice president of asset management said, “From an asset manager’s perspective, this technology will allow us to know immediately if there is a problem that is easy to fix or if it is a serious problem that requires capital investment.” For utilities, data from smart grid sensors is unlocking new possibilities from a network that was once hidden to them.

Tollgrade Communications 

RELATED