By Andy Harding and Christos Aslanidis

The European Union is committed to achieve a 20% reduction in greenhouse gas emissions by the year 2020, compared to 1990 levels, with at least 20% coming from renewable sources. They have the added motivation to reduce dependency on “less stable” sources of energy from outside the European Union and are adopting a multi-tiered strategy, not only to improve energy efficiency of appliances in the home, but also to increase consumers’ awareness and their ability to influence how energy is managed within their home.

Comprehensive management of energy demand and consumption across the whole grid, from the energy generator all the way to the consumer, leads to the concept of smart grids.

This article focuses on the technology selection for the connection of the home to the grid, the so-called “last mile” or wide area communications network (WAN). So which technology, or technologies, should be chosen for a smart grid WAN? A good starting point would be to look at established WANs. It is a fact that more than 90% of the established and planned automatic meter reading (AMR), automatic meter management (AMM) and smart grids in Europe have selected powerline communication (PLC) as the WAN technology of choice. Cellular technology (e.g. GSM/GPRS) tends to be used in less densely populated areas, or in those areas where large scale rollout is not part of the plan. DSL is being considered as a possible future solution, for example in the UK, where there is an ambitious government-backed plan to achieve 100% DSL penetration by 2012.

Today’s requirements cover applications such as remote meter reading, remote tariff updates, remote disconnect and tamper detection, whilst the ability to make remote firmware updates gives confidence that the life of meter assets can be extended in the future. Once the smart grid is in place the next step is to use it to allow demand side management, as well as other applications. There is a danger that uncertainties about “unspecified” future needs can lead some people to reject available and proven technologies with the argument that “if it’s already available then it can’t meet my future needs.” The risk is that unproven technology is just that – even if it promises to be “future-proof”.

It is also helpful to understand the different topologies that would apply to the different technologies (Figure 1), especially when evaluating comparative system costs, power consumption, or how reliable access to the network would be.


Figure 1 – Typical network topologies

In the case of PLC the local powerline network is connected to a “concentrator” or “data collector” in the local substation. This is an important point – the data is “concentrated” at this point which makes for very efficient communications over the backbone. Typically a single access to the concentrator in the substation allows the exchange of relevant data from up to a few thousand meters at once. It is also possible in PLC networks to send “broadcast messages” that are sent to all meters. Thus, all meters in a network can be updated with a single message such as for tariff or firmware updates.

In cellular networks a local transmitter mast acts as an intermediate step for the communications (but it would not concentrate data). It is a point to point connection, and a point to point communication is required for every data exchange. This is not ideal for efficient network communications, either from the data capacity or from the energy efficiency point of view. In the DSL case there would be a local switch, but effectively the communications are point to point and therefore less efficient.

The data requirements that were estimated in the 2007 UK Energy Retail Association (ERA) Smart Metering Operating Framework are detailed in Table 1. This suggested, for example, that 100% of meters would be read once per day, while 25% of meters could have a new tariff set once per month. Assuming a network size of 500 meters per concentrator, it can be concluded that for electricity meters a net WAN data rate of 35 bits per second would be required. For gas meters the data requirement is much lower, just 3 bits per second, resulting in a total requirement based on known services today of only 38 bits per second.


Table 1 – Estimated requirements for initial UK smart grid deployment

Available PLC technology already delivers more than 10 times the net data rate required with a roadmap of interoperable higher speed devices that perfectly fit the data requirements of tomorrow’s applications. Cellular solutions would also fit in terms of data capacity, whilst DSL solutions are over-sized and over-specified.

The University of Leuven in Belgium carried out a study into comparative implementation costs of smart metering communications technologies (Figure 2). The real cost of the technology selection includes both the initial modem cost as well as the ongoing cost of ownership such as data costs. These are free for energy providers that own the last mile network, but not for DSL/cellular services where data needs to be paid for. In this study the initial costs for PLC and cellular are fairly similar, whilst the cost of ownership is dramatically higher in the case of cellular.


Figure 2a – Comparative ownership costs over a 15 year lifetime
Figure 2b – Cost of ownership for 25 million meters

The architecture of the network is an important factor in energy consumption. If we think simplistically about communications modems we see that PLC and cellular modules have similar consumption, whereas DSL requires more power.

However, when we look at the backbone the situation becomes far more complex. The data concentrator plays an important role in PLC networks as it “concentrates” data and so communications between the concentrator and head end system are more infrequent. On the other hand, every communication for a cellular or DSL system would be a point to point communication through the local mobile mast or PABX through the backbone to the back office.

A key issue for energy providers will be to make sure that connection with the network is reliable. With a PLC network, interrupting the communication requires tampering with the powerline. And without communications there is no power, so there is little motivation for a consumer to try this. With cellular communications the problems of network coverage are well known, especially in buildings, as is consistent network availability.

With DSL it is likely that an additional dedicated line would be required. Otherwise it would be too easy for a consumer to interfere with their broadband router.

In terms of coverage, a PLC network will use meters as a repeater. This is a common network architecture in suburban areas, and it is perfectly possible to achieve coverage of an area as large as 30 km2 from a single concentrator.

DSL does not provide good coverage today, but if the target of the UK’s Digital Britain initiative is achieved it could provide a solution, albeit with higher power consumption, cost and network reliability issues.

Finally, a PLC network will offer additional benefits that are not possible with other technologies because of the existence of a PLC concentrator in the local substation. For example, in load control the concentrator can analyse which phase meters are connected to allow network engineers to decide which phase to connect new buildings to.

Furthermore, network problems can be identified and located in the early stage and repairs can be scheduled appropriately even before power has failed. Such functionality is currently implemented for instance by Guernsey Electric.


Figure 3 – Overview of smart grid WAN technologies

This article has highlighted a number of considerations when planning to implement a smart grid WAN. These include the cost of ownership, power consumption, reliability of network access, exposure to risk of tampering, data requirements and the need for mandate from government.

Plotting all these on a spider chart (Figure 3) one can see that powerline offers the strongest solution in almost all aspects. The one area of concern is that access to the last mile would need to be regulated, however regulation will be required whatever technology is selected and so should not be a barrier.

Considering that there could be uncertainty about reliable access in rural areas then a mixture of powerline and cellular solutions would appear to be the most appropriate technologies to use. Clearly smart grids can be established without the need for any new wires!