Email This Author Comment On Article About The Author More Articles By This Author

The evolution of smart grid initiatives is ushering a new era of energy generation, transmission, distribution, and customer interaction. It is also presenting a paradigm shift in the way utilities will need to develop and operate the communications infrastructure, which will be a critical foundational element for enabling smart grid functionality. We believe the evolution to tomorrow's smart grid will present a profound transformation in the way utility companies architect their communications infrastructure.

A Changing Power Grid Landscape

The challenge of upgrading aging energy infrastructure combined with the need to balance environmental concerns, peaks in energy demand, and related government mandates has utilities pursuing new energy initiatives, such as smart grid. We define the smart grid as one that uses information and communications infrastructure technology to manage grid information flows to make the power grid observable, controllable, automated, and integrated. Over the next five to ten years utilities are planning to rapidly overhaul grid infrastructure to adopt many smart grid applications. This growth will be driven by the need to meet government-stipulated deadlines (in the UK, Canada, and Australia, for example), and may be supported by the availability of government stimulus funds (as in the case of the United States).

The communications infrastructure component of the smart grid is critical, and the environment will be very different from what utilities have been used to for decades. Based on our experiences in working with utilities on smart grid designs and implementations, we see at least four trends emerging in the way the smart grid will be built, operated, and controlled. These trends include:

• Shift from centralized to peer-to-peer control.
  Utility communications infrastructure traditionally has been developed as star networks with centralized command and control. Smart grid communications will require more decentralization as command and control are distributed across the grid. This shift will require communications infrastructure that supports dynamic peer-to-peer communications.
• Shift from centralized generation to distributed energy resources.
  The growth of distributed energy sources, such as rooftop solar panels and wind turbines, is changing the traditional utility model. A mode is emerging where many much smaller energy sources are connected to distribution grids -- on a neighborhood or premise basis. As utilities revitalize their communications infrastructure, they need to incorporate management of large numbers of discrete, distributed energy sources.
• Shift from few end points with little intelligence to many end points with large amounts of intelligence.
  The penetration of smart end points into the power grid has raised new issues for utility communications infrastructure. In the new smart grid environment, utility telecommunications must reach out to new grid devices, and even into customer premises. Vast new volumes of data can be generated by new intelligent end point devices, presenting the utility with a trade-off decision in terms of bandwidth and latency versus cost. This trade-off must be thought of in the context of how much processing is done at the end points versus how much raw data must be carried by the telecommunications network. The trade-off between distributed intelligence and telecommunication network is just one of many issues for utilities to consider in the design of a smart grid.
• Shift from low data volumes, often slow response time requirements to high data volumes, low latency requirements.
  Smart grid communications systems must support several different data classes that have different behavioral characteristics. Operational data tends to be constant in volume and timing, and so it is fairly easy to understand in terms of bandwidth and latency requirements. Non-operational or telemetry-type data is similarly easy to understand, although the data volumes for a smart grid are much larger than for a traditional grid, something that is also true for smart grid operational data. A third class of data, asynchronous event messages, is generated by smart grid devices in reaction to grid physical events. As a result, these messages come in unpredictable bursts and floods that need to be transmitted and processed with very low latency. This data class has rather different communication performance requirements that need to be accounted for in the design of the smart grid communication system.

Communications Infrastructure Options

These emerging trends will drive new requirements to make smart grids and smart grid applications a reality. The move toward a smart grid will entail an in-depth evaluation of various deployment options related to communications infrastructure. Options include: 1) building and owning a network, 2) leasing a network from a telecoms carrier and 3) building and sharing a network with other utilities. In some cases, the utility will use a combination of leased, shared, and privately owned infrastructure. A high-level view of smart grid is shown in Figure 1.

Figure 1: High-level view of smart grid

In deciding on whether to build or lease a network, utilities will need to build economic models that detail the associated capital and operational expenditures for each option. In addition, utilities will need to assess the level of effort to operate and manage a network of this magnitude. The utility must also analyze the technologies and standards available to enable this critical layer of the end-to-end smart grid solution. Careful and simultaneous evaluation of these variables is needed to decide the direction the utility may want to take in its deployment strategy.

Building a Network: Building a new network that supports a smart grid is obviously a capital-intensive, complex, and multiyear project. Many utilities, nevertheless, are at a point where they need to refresh their networks or are looking to apply government stimulus money to launch smart grid investments.

A significant benefit of building one's own network is designing to meet one's specific operational and security requirements. Utilities can also manage the deployment so as to meet particular coverage requirements and have complete control over the traffic, especially during emergency situations. In addition, the network can also provide economies of scale by providing connectivity to the utility's mobile workforce, the technicians who are servicing the grid network, and the maintenance staff. The network also potentially can be used to support applications for services other than power, such as home area networks. The downsides of building a proprietary network include the high upfront capital costs and the need for experienced telecom people to run large data networks with extremely high reliability. In addition, the organization will need to keep up with rapid technology changes and provide frequent network upgrades.

Key considerations in building a network will be the current state of the utility's infrastructure, its reusability, and the availability of communications infrastructure assets. Technology decisions will need to be made, including what spectrum to use -- licensed or unlicensed -- and how much is needed. Figuring out an effective spectrum strategy is a good first step in determining a deployment strategy. We believe the degree of reliability, control, and security that a smart grid requires can be best achieved with licensed spectrum. A common view is that because licensed spectrum is owned, is private, and can be used for a single purpose, its users will not face interference, performance, or quality issues. However, not all radio frequency (RF) bands in the available RF spectrum are equal. For example, a band within the 700MHz range has far better coverage than, say, one in the 2GHz range. We believe that due to the associated costs and availability of spectrum, design strategies going forward will likely entail some hybrid combination of licensed and unlicensed spectrum.

After determining the spectrum strategy, a utility will need to analyze the available technologies in these bands. RF mesh in the 900 MHz band to date has been a common technology in the unlicensed space, but licensed spectrum opens several new options. For example, some utilities are considering deploying WiMAX for advanced meter initiatives (AMI); that is, a WiMAX modem is integrated within the smart meter. The network technology strategy will most likely involve a hybrid of technologies that includes multiple local and wide area wireless technologies, fixed line, and even microwave. The strategy may also include existing utility assets, leveraging right-of-way fiber lines a utility already owns, and potentially self-healing networks to support the infrastructure.

Design strategies should be based on careful analyses of cost advantages, technology road maps, and operational requirements including maintenance and support. They should also have at least a five-year perspective that includes enablement of smart grid capabilities at the appropriate time.

Leasing Network Services: While some utilities will build their own networks, others will not want to venture into the business of running large communications networks. These utilities may choose the option of leasing network services from a telecom carrier or a service provider. Many major carriers operate wireless broadband networks in addition to fixed-line networks. Some of these telecom carriers are also deploying "4G" wireless networks and aggressively building fiber coverage to the home. Depending on the carrier and the markets they operate in, some appear to have ample capacity as well as the capability to provide performance and quality-of-service guarantees to support smart grid applications. A utility would likely negotiate customized pricing schedules and service level agreements in this scenario.

The obvious advantages of leasing are bypassing the need to make a large, up-front capital investment, taking advantage of the excess capacity carriers have on their networks, and relying on their experienced operational staff. Telecom providers are also accustomed to adopting new waves of technology -- which change rapidly in the telecom space. In addition, with stiff competition among wireless carriers around the world and customer penetration reaching 90 percent and above in most regions, the carriers are keen on capitalizing new opportunities that can help grow their businesses. Machine-to-machine and smart-grid applications are attracting the attention of carriers. This environment bodes well for utilities as it enables access to new subscription pricing that is much more attractive and customized than pricing for traditional voice and data services.

The drawbacks of leasing from a carrier are around lack of control, perceived limitations on security, and sharing of infrastructure with other users with limited provisions for priority and quality of service. From a technology standpoint, connecting with a public carrier will entail developing integration links between the smart grid and a carrier's network. The utility will also be using a network that is shared by perhaps millions of public subscribers; issues such as priority and scale become critical requirements in this environment. Commercial cellular networks do not prioritize data traffic, which is a requirement for many smart grid use cases. Many utilities prefer direct control over all their communications networks.

In addition, utilities want assurance of general reliability and quality of communication service. In leasing from a carrier, utilities will be in less of a direct position to control and manage the quality of service or predictability and reliability of the network. To respond to these concerns, utilities will need to set up service level agreements (SLAs) with the carriers for prioritization of traffic, quality of service, and security in order to meet operational and regulatory requirements. To manage the risk that the carrier may lose coverage or connectivity, the solution may need redundancy or other measures to ensure minimal impact on the electric utility customer.

Building and Sharing a Network: Under this model, two or more utilities would partner to share a communications infrastructure with a carrier or another organization. This model could be an attractive option for utilities in second-tier cities and potentially for gas and electric providers, specifically for the last mile. Each utility would still be responsible for its customer premise network components such as the meters, but would share communications for the backhaul, the core network where traffic is aggregated, and the network operations center.

The advantages of a shared model are the economics -- spreading costs among two or more utilities and gaining economies of scale as more organizations join the network. Sharing could also be a faster-to-market route. Network management could be structured as a shared services center, enabling scarce or experienced talent to be shared. Building and running this network is a large undertaking that will require partnerships of varying scales. In the case of a greenfield, bottom-up build, the utilities can take a shared approach for scarce skill sets and resources that can spread the efforts and the risks across multiple parties in both the build and run stages.

The disadvantages that may arise include issues around control and priority of service, the handling of proprietary information and the governance and management challenges among multiple parties. In addition, as is the case in the build model, utilities will need to place a heavy emphasis on and commit resources to the maintenance and operations of the network. The learning curve will likely be very steep. In addition to service priority, there is the priority of network operations. In the case of a natural disaster or major outage, resources will have to be managed in the most effective manner possible to ensure an acceptable balance of response time for the sharing parties. Security also is greatly affected because utilities will have critical infrastructure at both the front and back end open to multiple parties.

The build and share option, while likely the most financially viable, will require the greatest level of collaboration on not only technologies but on operations. Utilities may have to make significant changes in operational models and business processes to use this approach.

Network Operations and Management

Regardless of the decision to build, share, or lease the network communications infrastructure in the smart grid, there is the business of operating and managing one's networks' applications and end devices. A utility will likely deploy millions of smart grid devices that will require specific levels of service assurance and telecommunications or service provider availability metrics, such as 99.999 percent availability. Network operation centers will either need to be built from the ground up or existing ones expanded and upgraded to handle the new and large number of devices and applications.

Such network operations centers will include engineering and technical support for all of the utility's sites (e.g., substations) as well as smart meters or any other connected devices. A framework will be needed to enable efficient operational support focused on service assurance for the millions of elements in the field. Frameworks such as the Enhanced Telecom Operations Map (eTOM) or Information Technology Infrastructure Library (ITIL) models used by telecommunications carriers could serve as the foundation for a smart grid service assurance framework. Figure 2 illustrates a model that can be customized for utilities, based on what telecommunications companies are doing today to guarantee service to their customers.

Figure 2. High-level communications layer service assurance framework

Whether the network is built, shared, or leased, the utility will have the responsibility of managing and maintaining its nodes and elements. Key to this process is the network operations center, or NOC. The NOC's primary role is to monitor the performance of various smart grid elements and supporting telecommunication network elements. Through an integrated toolset and processes, NOC staff can have an end-to-end view of the services carried over the network, and can quickly assess the impact of faults on customers, infrastructure, and operations. Besides the service assurance of the communications network solution, the NOC can also support deployment/commissioning activities by ensuring the newly commissioned elements and their status are visible in the network management dashboard applications.

Conclusion

For many utilities, the path to high performance in smart grid development will entail deploying new or revitalizing existing communications infrastructure. Utilities will need to plan their communications system by taking a holistic approach that considers an array of variables, including smart grid functional and performance requirements, network ownership and technology options, and network management approach. As the smart grid concept continues to develop, we advocate creating a comprehensive smart grid road map and requirements definition early on in the planning process with the full end in mind, not just meters. Unfortunately, the transformation to a smart grid in no way allows for a cookie-cutter approach to building the communication core. There will likely never be a shrink-wrapped, out-of-the-box smart grid communication solution because utilities vary greatly in both requirements and constraints, based on internal structure, geography, population demographics, local regulation, and economics. If the eventual goal is a smart grid, the key is to plan holistically today and with the big picture of tomorrow in mind.