Archive for the ‘Post Deployment’ Category

The Smart Grid – more than just smart meters

Posted by Utilimetrics on January 19, 2012

By Jonas N. Olsen, On-Ramp Wireless, Inc.

Search the term ‘smart grid,’ on the Internet and you’ll get a long list of articles about smart meters. But the smart grid goes far beyond the meters. Often located in underground basements or on impossibly high rooftops, devices associated with operating the smart grid can be hard-to-reach, especially in metro or other challenging environments, where there is no hardwire Internet connection. 

Distribution Automation (DA), which has the potential to significantly improve the performance of the smart grid, also struggles to connect these “smart devices.” DA systems strive to improve reliability of the smart grid through situational awareness, outage management, and faster response times when a fault is discovered. A smarter distribution system can also assist in utility capital planning by highlighting changing load conditions over time. Utilities are interested in implementing DA because it improves their bottom line and it is an autonomous project – they don’t need to communicate with their consumers.  

Half the battle of optimizing a DA system is connecting these billions of “smart” devices cost-effectively. Utilities need to be able to deploy a secure and reliable wireless remote monitoring system throughout their distribution network to accomplish asset monitoring, fault indication to improve outage restoration, alert to power quality issues, capture power theft and act as a hub for demand-side load management, which ultimately lowers cost of operation and maintenance costs for utilities.

For electric utilities, an added challenge is the advent of distributed generation, where electricity is generated from many small energy sources, and the introduction of Electric Vehicles (EVs), which bring new pressures to utilities’ distribution grids. However, wireless remote monitoring systems can address this too, especially as they become increasingly prevalent. 

Wireless Systems for Remote Monitoring

While utilities are increasingly using wireless technology for remote monitoring applications, it should be noted that not all wireless technologies are created equal. There are significant differences, which ultimately determine their applicability (cost and performance) to a specific application.

The key characteristics of a wireless system that determine its overall applicability are:

  • Coverage: The system’s ability to transmit a signal over a long distance.
  • Capacity: This can be defined in two terms. First, there is the actual application throughput (good put) from a single end device (such as a pressure sensor) in the network. Second, the overall network capacity must also be considered. This refers to the ability of a concentrator (or Access Point) to process data from nodes in the network. This is what we call the overall throughput capacity.
  • Power consumption: In remote monitoring applications, many end points must rely on batteries as the main source of power. The preference is for lower power consumption to extend the span between battery replacements. In some installations solar or other renewable sources can be used to supplement a main battery.
  • Latency: This term relates to the time it takes for information to move through the system in either direction (from the remote device to a central collection system and the other way around).
  • Communication type: Wireless (or any communication system for that matter) operates as either simplex (communication only one way), half-duplex (communication both ways but not at the same time), or full-duplex (same time, bi-directional communication).

Wireless spectrum allocation is another concern that must be addressed. Wireless systems perform over a wide range of frequencies, from a few kilohertz to high frequency gigahertz systems. Many frequencies are licensed and typically bought by private companies through public auctions. Other frequencies are designated unlicensed and can be used free of charge. The unlicensed frequencies, however, are associated with rules and regulations about how the free spectrum can be used by various different private operators. An example of a rule would be guide use of popular technologies such as Wi-Fi and Bluetooth. The rules and guidelines address the amount of power output and the occupied bandwidth that can be applied in the allocated free spectrum. The rules vary from country to country, and operators need to observe and comply with local restrictions. Many remote monitoring applications operate in the free and unlicensed frequencies. This is mainly due to cost concerns, as many of these applications do not warrant the high cost of dedicated frequencies or the monthly recurring fees incurred when renting this spectrum of a third party operator.

It is important to recognize that different wireless systems mix and match these characteristics in various ways. This also means that there isn’t a “one size fits all” wireless system that is ideal for any application. The unique application requirements of a flow measurement system, for example, are very different from a low latency, factory floor SCADA application, which may require millisecond response times. Some applications will require very high data rates, while others just process a trickle of index data throughout the day. Pick any of the above mentioned system characteristics and the same kind of comparisons could be made.

Most remote monitoring applications fall into a category where range and low power consumption is prioritized. Range, in this sense, should be understood as either great distance (e.g. >10 km), or as the ability to penetrate obstacles, like vegetation, building, etc. Low power is key, as many remote devices will require monitoring without access a continuous power source (i.e. battery operated). Relatively small amounts of data are typically transmitted and capacity therefore tends to be a minor concern. What is important, however, is the aggregate data rate at the collectors/Access Points. If a wireless system has great coverage it is likely to provide coverage for many thousands of devices from a single network infrastructure point. This “Access Point” must provide sufficient throughput capacity to, robustly, receive and process data from all of the covered devices. This is where many narrow-band radio systems fail to meet the requirements of utility customers.

Finally, one needs to consider the communication type. Some applications can survive with simplex communication. This would be the case when all the application is intended to do is to collect data from a remote point. For an application where two-way communication is needed (resetting alarms on remote devices or changing configurations) a duplex system must be deployed.

Backhaul Options

An additional concern is backhaul from the remote site to a central data processing site. Most remote operation is far from the main hubs for IT infrastructure. When a private wireless system is installed (as opposed to using public infrastructure like a carrier based GMS network), it is up to the user to provide all connectivity links in the system. A wireless system that uses unlicensed spectrum will typically terminate in a set of wireless access points or gateways, which then need to be connected to the overall company network. This can be done in various was, but the most commonly used methods are a direct connection to the Local Area Network (if available), backhaul via a public cellular network (again, if available), and finally through satellite links. These options are listed both in terms of preference and cost.

Systems Integration

Integration with a process automation platform has to be considered. For a wireless remote monitoring system to be effective, it has to present the collected data in an industry standard format. An end-to-end wireless remote monitoring application will provide every step in the process, from integration of the wireless module with the remote sensor, wireless networking and networking infrastructure and conversion of the data to a standard format, such as Modbus or OPC. This allows for simple integration, both with on-site process automation systems and backend historical data storage.

Conclusion

As electric, cable, and telecom utilities increasingly work to improve their DA systems while having an eye on their bottom line; they should look at wireless remote monitoring solutions. With the right system, utilities should be able to pinpoint a problem exactly where it occurs so that their work crews can go directly to the affected area to fix it, and don’t have unnecessary downtime. In some cases, preventative maintenance system integration will even avoid failures altogether. A system should also be able to integrate fault indicator alarms with work order systems for simple and automated dispatch of workmen. Beyond the workforce, a connected DA system also leads to low power consumption by limiting peak power requirements, better capital planning, and fewer outages.

Lastly, the network should ultimately allow for a low-cost, fully-automated Distributed Grid, which enables e.g. fault indication (above and below ground), transformer monitoring, substation automation and other applications that were previously thought unfeasible to automate. When these applications come “online”, utilities will see significant enhancements in key performance metrics.

Jonas N Olsen is the VP strategic partnerships for On-Ramp Wireless, Inc., which is currently deployed by a Western utility for its wireless communication system.

Posted in Distribution Automation, Electric Vehicles, Post Deployment, Remote Monitoring, Smart Grid, Smart Meters, Systems Integration | Leave a Comment »

DTE Energy and PECO’s Experiences With Outage Management Systems

Posted by Utilimetrics on October 27, 2011

With AMI deployment comes the benefit of having real-time information. New and advanced outage management systems (OMS) collect automatic messages for alarms and outages. But as utility companies adjust to the advanced levels of maintenance that come with AMI, AMR and OMS, do they have field operations, dispatch teams and call centers ready for all of this data?

This article highlights the Outage Management Systems education session at Autovation 2011 Monday, Sept. 26.

DTE Energy has been working to integrate AMI into its OMS agenda, starting with internal workshops, which explain the benefits of OMS for utilities:

  • Obtain early outage detection.
  • Receive notification of momentaries.
  • Receive improved restoration information.
  • Send the right crew the first time.
  • Reduce okay on arrivals (OKA).
  • Prevent/ reduce customer callbacks.
  • Detect trouble behind trouble.
  • Improved customer satisfaction.
  • Reduce call center volume

Bob Sitkauskas, DTE Energy manager of AMI, reviewed DTE’s implementation of AMI data into their outage systems and the use of their Complex Event Processor (CEP).   Items to be considered in the implementation include:

  1. Collection Engine
  2. AMI/MDM
  3. Enterprise Service Bus
  4. Complex Event Processor (CEP)
  5. Outage Processor Interface (OPI)

The CEP successfully filtered out over 12,000 momentaries incorporating the “brother/ sister” concept in CEP where PONs received after 10 minutes are matched against PRNs (Power Restoration Notification) received on the same transformer in the previous two hours. If found, the late PONs are dropped to avoid creating an outage and an erroneous field visit

The advanced OMS also identifies problem meters in the field and intentional interruptions that were not properly reported by field personnel.

Although the integration has proven successful for DTE Energy, Sitkauskas outlined several challenges that come with interfacing to a legacy OMS. For example, the CEP could not handle the volume of PONs in a timely manner. In addition:

  • Work that was planned and scheduled through DTE’s customer service billing was not processed through the CEP and into OMS resulting in false outages.
  • Electricians were performing work for customers which required them to remove the meter, thus resulting in an outage.
  • Line crews were performing intentional interruptions without following established process of notifying Central Dispatch prior to an outage.
  • The Power Restoration Notification was received after five minutes resulting in an outage. A circuit breaker opened for 30 seconds and then closed resulting in an erroneous truck roll.

Sixty days after the initial installation, AMI was reinstalled in the OMS process flow. The installation consisted of creating additional CEP/ OPI filters, implementing OMS enhancements, reinforcing process with Central Dispatch and Field Operations. This implementation was restarted in phases, from station to station.

After working to re-tie the AMI to OMS, DTE Energy has been able to prevent false outage and erroneous truck runs, perform on demand reads in OMS, utilize AMI data for system outage data and analysis (SODA) reviews, utilize supervisory control and data acquisition (SCADA) data to validate sustained outage, and provide a daily status report.

“Start small,” recommended Sitkauskas. Prior to implementation, it’s important to test the installation.

Outage Management with AMR at PECO

PECO completed integration of its AMR and OMS systems in 2006, and eight years later, the Exelon Corp subsidiary that served the southeastern region of Pennsylvania revisited the journey to integrate and reviewed the benefits.

Kevin Cornish, Enspiria Solutions and Glenn Pritchard, PECO discussed the opportunities that have resulted from advanced OMS:

  • Improved customer satisfaction
  • Power status verification
  • Reliability analysis
  • Future outage prediction

Today, this system provides significant benefits daily, and specifically during storm restorations.

Pritchard explained that “pinging” is a valuable tool in outage verification. Pinging refers to querying the AMR network to determine if a meter has recently communicated. (PECO received roughly 125,000 pings annually). Whether you’re checking to see if a customer is truly out, the validity of a job packaged prior to dispatch or that a job is complete, pinging will save your company a lot of headaches, according to Pritchard.

If an automatic assessment outage lasts longer than 20 minutes, it is automatically pinged. If the ping responds with “Power On,” the outage is cancelled. In the instance that it indicates “Power Off,” a transformer analysis is performed to potentially escalate the event into a larger outage. PECO’s results show that since 2004, 64,205 pings were cancelled, 19,550 were not.

The outage is identified, dispatched and resolved before any customers notify PECO of the event.

As an example, Pritchard described a “Summer Slam” event in July, 2006. Thunderstorms caused nearly 400,000 power outages. Twelve hundred customer outage calls were cancelled without crew dispatch due to the meter pings. Seven hundred fifty customer calls were escalated into primary events via pings to neighboring customers’ meters. Conservative estimates indicate AMR helped save in excess of $200,000 in avoided labor costs during this storm alone.

With the success of the simple meter pinging application, several enhanced tools were developed:

  • Transformer analysis
  • Fuse analysis
  • Circuit analysis
  • Batch pinging

PECO’s AMR and OMS implementation project was a transition from concept to success, and now AMI. The project has created daily benefits well beyond the original estimates. The success of this project has advanced the metering industry as a whole by proving that meter-based outage management benefits are real.

If your utility has an OMS story to tell, please share your experiences (challenges and successes) with your peers. There are several ways you can do this:

  • Submit an abstract for Autovation 2012, Sept. 30-Oct. 3 in Long Beach, Calif. The Call for Speakers will open soon.
  • Provide a byline article for News Link or agree to be interviewed by News Link staff for an article. Or, submit a blog post. Contact Janice Greenberg.
  • Consider hosting a regional learning lab or participating in a webcast. Contact Debby Scheck.
  • Start a discussion on the Utilimetrics LinkedIn Group

We look forward to hearing from you!

Posted in Autovation, Meter Data Management, Post Deployment, Smart Grid, Smart Meters | Leave a Comment »

Beyond the Meter

Posted by Utilimetrics on October 25, 2011

Lessons Learned from Oncor and Portland General Electric

For many years attention has focused on pre-deployment and deployment of advanced metering systems (AMS).  As utilities enter the final stages of deployment they face new challenges as well as tremendous opportunities for integrating technology within the utility and improving operations. 

Autovation 2011 covered the entire utility technology lifecycle. This article highlights the Beyond the Meter education session Tuesday, Sept. 27.

Oncor, the sixth largest utility in the U.S. began deploying fully functional AMS in late 2009. About two million of Oncor’s 3.2 million meters have been deployed with full deployment scheduled for 2012. This fully integrated system provides:

  • 15-minute VEE (validate, edit, estimate) data to customers, REPs and ERCOT (for settlement).
  • 2-way transactions (disconnects/ reconnects, on-demand reads, etc.).
  • Secured connections and services to home area network (HAN) devices via ZigBee SEP 1.0 radio frequency interface.
  • A common Web portal for REP, customers and customer authorized 3rd parties (GUI and APIs)

So how does it all work?

“You need a robust testing environment,” said Mark Carpenter, CIO of Oncor, Texas’ largest regulated transmission and distribution utility that serves 7.5 million people statewide. Carpenter is also a newly-elected Utilimetrics board member.

“In theory,” he said, “it’s nice to specify exactly what you want before you actually start building it.” Carpenter explained that when inventing the system in a dynamic environment, clarification and modification contribute to a continuous and repetitive process.

“Remember, [AMS] is not just a meter reading system,” said Carpenter. “This is a SCADA system.”  And as Carpenter specified, “It’s extremely important to know and understand your market.” According to Carpenter, the Public Utility Commission of Texas-led Advanced Metering Implementation Team process has worked well in Texas.

When designing the system, Oncor adhered to solid design principles, factoring in security from the very beginning. In an effort to make the systems most efficient, Oncor:

  • Included performance monitoring;
  • Designed the system for ease of upgrade/ modification;
  • Planned for evolving CIM interface changes;
  • Considered multiple software/ FW changes in advance; and
  • Provided robust system synchronization

In the testing/ building phase, Carpenter said that the two most important things to consider are:

  • Establishing robust development and test environments will help to maintain strict version control; and
  • Maintaining strict version control

“Managing data is a big deal,” said Carpenter. He explains that utilities must be continuously monitoring these large integrated systems, which require “constant care and feeding.” Oncor generates about one terabyte per month, within the two million meters. “Don’t wait to establish data retention policies.”

And continual performance improvement is imperative: “It’s important to always remember to continually validate the end-to-end production system,” said Carpenter, “especially after modifications.”

As a utility, your main focus on customers and stakeholders is key: “There are stakeholders in this business,” says Carpenter. “This isn’t just about technology—it’s about everybody.”

Revenue Protection with Smart Meters

Eric Spack and Steve Sprague are leading a unique mission at Portland General Electric (PGE). The PGE team is taking revenue protection to the next level and beyond, utilizing new technology to work more efficiently.

A proactive approach to revenue protection utilizes alarms and generates leads based on interval data and primary metering. A major part of the team’s workload consists of confronting marijuana growers whose operations result in huge losses for the utility.

“We had 45 meters, from which we were missing about 1,400 kilowatt-hours,” said Sprague, “and at the end of the month, we had 20,000 kilowatts missing.” Utilities are facing huge losses from thefts like these, and at PGE, in a state where growing medical marijuana is legal, these operations are oftentimes extended beyond legal limits, and utilities are paying the price.

Over the last three years, AMI has dramatically improved energy recovery for PGE, from 32 Mwh in 2007 to 44 Mwh today and 75 percent of leads for the Lost Revenue Protection are generated by readers:

  • Tampers & Diversion;
  • Stopped/ Damaged meters;
  • Multiplier errors;
  • Lost meters;
  • Drug houses; and
  • Safety issues

These smart meters maintain current capability, allowing for real-time usability. What specifically can the meters do?

  • Tamper alarm: If the meter is pulled or removed, an alarm is generated with a date and timestamp.
  • Alarms scored: Leads are automatically prioritized.
  • Lead generator: All the leads are sent through a portal to Energy Recovery where they are reviewed and assigned to ERU Investigators or meter men.
  • Leads filtered: Without filters, alarms are useless and “we are filtering out 68 percent of the alarms and leads coming in.”
  • Filtering against: WMIS, Service Link, Outage, which avoids wasted time on wasted trips.
  • Added benefit: Not only generates leads but allows PGE to use the information on existing cases and leads from other sources.
  • KWH analytics: Low use, high use and zero use, it reads abnormal usage patterns

Once the norm is established, Point of Passage metering installations are screened to prevent from losses. Meter failures and alarms, however, do not cause the largest losses. The problem therein lies with theft and particularly, grow houses.

When marijuana grow operations overload transformers and connectors, it’s at the expense of PGE.  “Houses are not meant to be greenhouses,” said Sprague. The usage thefts are typically in the range of $1,500 to $2,500 per month, according to Sprague and Spack. Ninety percent of power thefts supporting grow operations are done by splicing in ahead of the meter.

“[At PGE] we have a 100 percent success rate in criminal grow diversion cases,” said Sprague. In 80 percent of those cases, money was recovered.

By learning how grow operations work, PGE adjusted to them and hunted them down, and by the time they were done, according to Sprague, they worked about 60 grow sites and billed roughly $620,000.

If your utility is near completion or has already completed deployment, please share your experiences (challenges and successes) with your peers.  There are several ways you can do this:

  • Submit an abstract for Autovation 2012, Sept. 30-Oct. 3 in Long Beach, Calif. The Call for Speakers will open soon.
  • Provide a byline article for News Link or agree to be interviewed by News Link staff for an article. Or, submit a blog post. Contact Janice Greenberg.
  • Consider hosting a regional learning lab or participating in a webcast. Contact Debby Scheck.
  • Start a discussion on the Utilimetrics LinkedIn Group

We look forward to hearing from you!

Posted in Autovation, Post Deployment, Revenue Protection, SCADA, Smart Grid, Smart Meters | Leave a Comment »

 
Follow

Get every new post delivered to your Inbox.