By Ericsson Australia Utilities Specialist, Yochai Glick and Broadband Strategy Manager, Colin Goodwin.
Around the world electricity utilities are updating and re-architecting their power networks. This is largely in response to growth in user demand and the restructuring of generation (including distributed supply from renewable generation such as wind and solar). However, there is also a compelling need to re-think existing architectures to incorporate far more pervasive communications. The resulting “smart grid” is a synthesis of energy, IT and communications infrastructures.
Some power engineers may argue that communications has been part of utility networks for decades. This is true. SCADA networks have been monitoring power transmission lines and equipment in substations since the 1950s. Utilities have been controlling user demand through tone control of off-peak hot water heaters and pool pumps since the 1960s and possibly earlier.
However, these are modest communications components incorporated into a power network architecture that has changed little since the basic model was introduced by Edison and Tesla in the 1880s. And it is still the case that the great bulk of utility distribution infrastructure (located in streets and on poles) is almost entirely offline – such that utilities often have no way of detecting faults other than waiting for their customers to call and complain.
Today, energy utilities are testing, piloting, or rolling out the building blocks of their smart grids. In a smart grid the key enabler is the communications infrastructure that overlays and intertwines with the power distribution infrastructure. It works in conjunction with field devices and office applications to monitor and manage power distribution infrastructure, making automatic adjustments as real world events such as storms, fires and runaway trucks damage the integrity of the physical grid.
The benefits for consumer and business users will be a more robust supply of energy with reduced carbon emissions and tools to help users reduce their own carbon footprint.
This discussion of smart grid goes beyond the deployment of smart meters (also called advanced metering infrastructure or AMI). The introduction of smart meters has predominantly been focused on the introduction of ‘time-of-use’ tariffs and in-home displays, with the expectation that consumers will modify their behaviour to consume energy at times of lower demand. While the concept of smart grid incorporates the introduction of smart meters, it is more far-reaching in that it directly helps utilities better manage their power networks.
The challenge to engineers when implementing a smart grid is to understand the dynamics of monitoring and managing the power network and to map this into communications traffic requirements (throughput, latency, etc.), to ensure that the communications infrastructure can be scaled and deployed to meet realistic future requirements.
As a provider of smart-grid communications networks, Ericsson has a keen interest in facilitating a detailed understanding of this subject and is collaborating in a program of research to model these requirements and validate the models against real power networks.
This modelling is expected to cover the following range of scenarios or use cases.
Grid monitoring and control
Just as SCADA networks have been used to monitor and control the transmission and substation portion of power networks for many years, smart grid aims to extend the monitoring into the distribution network. Examples of devices to be monitored include transformers, fault detectors, pole-top switches, sectionalisers and reclosers. Communications traffic requirements are expected to be modest, but status reporting will be frequent.
When monitoring identifies grid failures, the root cause of the failure will have to be quickly identified and appropriate devices commanded to fix or minimise the consequences of the problem. This is often referred to as FDIR (fault detection, isolation and recovery). Traffic volume requirements are expected to be small, but commands and responses must be dealt with quickly and with high priority.
The volume and nature of communications regarding FDIR will be determined by the extent to which detection and recovery are dealt with centrally or through local distributed mechanisms, or a combination of central and distributed control mechanisms.
The introduction of grid monitoring and control is expected to greatly reduce the frequency and impact of faults in the distribution networks. This will lead to improved reliability figures for the utility.
Pervasive communications will also enable the introduction of new regimes for using and maintaining grid equipment. Dynamic rating allows switching and dispatch decisions to be based on the actual condition of equipment due to their operating environment and operating history rather than their factory specifications. Condition-based maintenance allows grid equipment to be serviced according to actual load history and condition rather than on fixed schedules. This avoids the costs of too-early servicing and prolongs asset life by avoiding too-late servicing.
Advanced metering infrastructure
Where smart meters are introduced the assumption is that they will be read frequently (even several times a day).
Smart meter traffic volumes are expected to be small and infrequent for each meter, but substantial in aggregate across the thousands of meters to be read. While each read may have a low priority, the records are of high importance to the utility as they are the basis for charging customers.
Demand management of in-home devices is likely to be done through the meter. This may be regular (e.g. turn-on/ turn-off hot water systems based on time of day), or for emergencies (e.g. fine-grained disabling of particular devices such as pool pumps or air-conditioners) in emergencies or for supply shortages. Traffic volumes are expected to be small and infrequent but substantial in total.
Service connection/disconnection (i.e. disabling or enabling energy supply to a premise) offers significant benefits. Perhaps the most obvious is being able to deal with a change in ownership of a property. There are also benefits when dealing with disaster recovery, such as the gradual ramping up of supply on the distribution grid after a major failure (this was an issue dealt with recently in Queensland in dealing with the aftermath of Cyclone Yasi and the statewide floods that occurred there). Traffic volumes are very low, but of relatively high priority.
Perhaps the most challenging “use case” regarding smart meter communications is the remote over-the-air (OTA) updating of the meter firmware. In worst-case scenarios this could apply to all meters in a region and may have to be done over a relatively short period (e.g. a day).
While some calculations of this use case suggest it can be a very challenging task, with thousands of meters having to receive megabytes of data, these calculations tend to ignore techniques that allow efficient use of the communications infrastructure (e.g. the use of multicast).
Distributed energy resources
While large sources of renewable supply (windfarms, solar, geothermal, wave/tidal) are increasing in number, they do not of themselves change the topography of the power grid, other than by requiring new transmission capacity to allow connection. However the irregular nature of their supply may lead to the introduction of new elements within the grid, notably distributed storage (e.g. flow batteries). And these elements will themselves have to be monitored and managed.
Co-generation or tri-generation, where business owners generate electricity (typically from gas), introduces another new source of supply, which is somewhat predictable, and somewhat distributed throughout the grid (albeit likely to be centralised in certain areas such as business districts).
More challenging is the introduction of autonomous distributed generation throughout the grid (small-scale solar/photovoltaic, small scale gas turbine or fuel cells), which requires the re-thinking of the distribution network into a bi-directional grid, along with the increased desirability of using embedded storage to smooth supply. Smart meters will typically be used to record energy generated, and this must then be read for cross-charging and billing settlements where incentive tariffs are in place.
Distributed energy resources will be an increasingly important aspect of the smart grid. It will introduce new devices and processes for monitoring and managing the grid. Current understanding suggests that these changes, while challenging for the power grid, will only drive modest communications requirements… but the area is new and much remains to be learnt.
The introduction of electric vehicles must drive a re-architecting of the grid and not just because of the massive increase in nighttime demand that seems likely. What other element on the grid is simultaneously a load, embedded storage, distributed generator, and roams from place to place throughout the day?
The mobility of electric vehicles introduces a need to handle the authentication of devices, prior to the transfer of energy to or from the vehicle. A thoroughly novel concept for the grid, which was designed for supplying energy to premises that stay put.
It is difficult today to be confident of the communications requirements for a smart grid with a large population of electric vehicles as the business models supporting them are still being tested. However, early work suggests that the changes in IT systems and energy infrastructure will overshadow the communications requirements.
Then there is the need to interact with field staff as they go about changing, maintaining and repairing the grid. In addition to field-force automation which supports the optimised scheduling and dispatch of crews, the ability to communicate real-time grid status and outage information to field staff will significantly improve fault-resolution times. This is an unusual smart grid use case in that it requires mobile telecommunications.
There will also be a need to implement communications gateways such that field staff can easily communicate with corporate/office-based staff, and in some cases with other field teams (e.g. emergency services workers) who may be using quite different wireless technologies.
During normal operations communications requirements are expected to be relatively modest, though traffic could be more substantial when crews are in a section handling a major event.
As discussed, there are a wide variety of requirements for smart grid communications, from regular low priority traffic to mission-critical disaster-handling traffic.
Utilities have considerable experience with the communications requirements for the real-time monitoring and managing of the high-voltage transmission portion of the grid, moderate experience with the medium-voltage portion, and least experience with the low-voltage distribution network.
Understanding how the combined grid, communications, and IT systems will interact requires sophisticated modelling, and the testing of the results of modelling by validation against real smart grids as they are deployed.
What is already clear is that it is essential that utilities design for the future. A communications network that is designed merely to handle meter reading cannot deal with the complexities of a smart grid with distributed generation and electric vehicles.
The introduction of smart grid is not a simple bolt-on to the existing power grid. Smart grid enables very different and very efficient processes that will increase the reliability of the grid, optimise demand, and reduce the carbon emissions and operating costs of the grid.
To achieve these important benefits will require investment in communications infrastructure, smarter grid equipment, and new people and IT processes. However, to make it possible for utilities to introduce future-proof communications for their smart grids, it may be necessary in many jurisdictions to change the legislated basis for access to the capital to fund this infrastructure, so that utilities really can build for the future.