Share Share this | Facebook Twitter YouTube LinkedIn Flipboard

IEEE: The expertise to make smart grid a reality

Advanced Metering Infrastructure Using Wireless Broadband Networks

Advanced metering with two-way communications has the potential to make meters a core element of an integrated system to better manage utility services. But what kind of communications are appropriate? Smart meter traffic is characterized by small session duration, limited mobility and large number of devices, and as such is not handled efficiently by existing wireless broadband access networks run the usual way.

Broadband wireless networks provide ubiquitous wide area coverage, high availability and strong security and are, therefore, a strong candidate for handling smart meter communications. Wireless operators naturally see an enticing business opportunity in advanced metering infrastructure (AMI), in that they stand to obtain additional revenue streams from existing cellular networks. Government agencies have encouraged such network sharing to reduce AMI's energy footprint. Broadband wireless networks were not designed, however, to efficiently meet the traffic requirements of AMI. So, we need to analyze the network requirements of AMI traffic and spell out the implications for choice of communications systems.

Existing wireless broadband networks presuppose traffic that is typically modeled as consisting of individual sessions, where the session duration or time scale exhibits a heavily tailed distribution and is usually orders of magnitude larger than the packet timescale. That is, the length of sessions varies widely and a typical session requires a great many packets to communicate digitally. This allows for each session to be treated as an independent connection, subject to admission control mechanisms, with associated signaling procedures for setup of radio and network resources. The signaling associated with connection setup represents minimal overhead compared to the total data transferred over the session duration.

In contrast, most AMI traffic is expected to originate from stationary devices (or devices with very limited mobility) and will consist of just a few payload packets between the meter and the meter data management system. Furthermore, it is expected that in normal operations, most meter traffic will be regular as opposed to being ad-hoc. That is, meters will periodically report data on the uplink, and downlink data from the management system may follow. After that, there will be a long period of inactivity until the next time meters report data.

This deterministic behavior, coupled with potentially very long sleep durations between communication attempts with the network, allows for optimizing the operation of the meter so that it is scheduled to connect (or re-connect) to the wireless broadband network only at specific time instances and only for a limited period of time; during that connected interval, the meter and management system can exchange information as needed. We refer to this kind of system as time controlled scheduling.

The advantage of supporting time controlled operation is that the meter is connected to the wireless network only for short intervals of time, as needed, allowing networks' resources to be more efficiently managed and for a very large number of devices to be multiplexed to a common base station. To contact a meter outside of its scheduled connection window, the protocol can be enhanced so that the network alternatively sends a notification indication to one or more neighboring meters that are in a connected state at that time, and the meters in turn relay the request to the meter in question over a secondary wireless channel that uses an unlicensed spectrum, like ZigBee or Wi-Fi.

Congestion control is recognized as another challenge for AMI. The very large numbers of smart meters give rise to potential "traffic burst" scenarios, which can arise when large numbers of devices are simultaneously reacting to a common event, such as a power outage. To minimize the impact on the wireless broadband interface, an application layer congestion control protocol can detect the common event and stagger — that is to say, buffer or queue — transmissions from meters.

Each meter is assigned a probability (p) to transmit an alarm upon detection of a shared event. The meter will either queue (with probability 1-p) or transmit (with probability p) this event. If the message is queued, then the meter will continue to monitor the air interface for an event notification from the network. This notification can be in the form of an explicit message sent from the base station, or, alternatively, the base station may update the transmission probability p to 0. Upon receipt of such notification (which is sent only if another meter was able to successfully transmit the shared event notification to the station), the meter will discard the queued message.

If no such notification message is received after a random period of time, the meter will again attempt to see if the message should stay queued or be transmitted. The process is repeated until either the message is transmitted or an event notification is received from the base station. The algorithm can be generalized to allow for different event transmission probability values for different categories of shared events, with high priority given to more critical events. The delay handling can be different for different categories of shared events. The backoff delay can be made shorter for more critical events.

The standardization of the aforementioned traffic scheduling and congestion control enhancements is underway in several Standards Development Organizations (SDO) including IEEE 802.16, 3GPP and ETSI under the broader umbrella of machine to machine communications. When implemented, these enhancements will allow for efficient operation of AMI over a broadband wireless network.

Contributor

  • Vibhor JulkaVibhor Julka received his M.S. and Ph.D. degrees from the University of Massachusetts, Amherst, and has over 17 years of experience in the telecom industry in system design and 3G/4G wireless standards development. His research interests include wireless networks, machine-to-machine communications, smart grid and sensor networks. He is a member of IEEE and Tau Beta Pi. He has worked as a consultant for Huawei Technologies USA, Inc.

    Read more

  • Ronald MaoRonald Mao is a member of IEEE and a contributor to 3G/4G wireless standards. His research interests include communication protocols and wireless networking. He has over twenty years of experience in system design and product development. He received an M.S. in computer science from DePaul University in Chicago, and a B.S. in aeronautical engineering from Beijing Institute of Aeronautics and Astronautics in Beijing, China. He is a senior manager at Huawei Technologies USA, Inc.

    Read more

About the Smart Grid Newsletter

A monthly publication, the IEEE Smart Grid Newsletter features practical and timely technical information and forward-looking commentary on smart grid developments and deployments around the world. Designed to foster greater understanding and collaboration between diverse stakeholders, the newsletter brings together experts, thought-leaders, and decision-makers to exchange information and discuss issues affecting the evolution of the smart grid.

Contributors

Iñaki LaresgoitiIñaki Laresgoiti, a member of the Engineering Association of Bizkaia, has headed support systems for network operation at LABEIN and later on...
Read More

 

Ronald MaoRonald Mao is a member of IEEE and a contributor to 3G/4G wireless standards. His research interests include communication protocols and...
Read More

 

Vibhor JulkaVibhor Julka received his M.S. and Ph.D. degrees from the University of Massachusetts, Amherst, and has over 17 years of...
Read More

 

Kwok W. CheungKwok W. Cheung is the R&D Director, Market Management Systems at Alstom Grid in Redmond, Washington...
Read More

 

Antonello MontiAntonello Monti, a senior member of IEEE, received M.S. and Ph.D. degrees in electrical engineering...
Read More

 

Ferdinanda PonciFerdinanda Ponci, a senior member of IEEE, received her M.S. and Ph.D. degrees in electrical engineering...
Read More