Setting the context for smart meters
The embedded market sees great potential in improving utility supplies but it presents complex challenges, says Alistair Morfey.
The EU2020 Energy Targets have been approved by most European countries to reduce climate change. They aim to cut energy consumption and greenhouse gas emissions from 1990 to 2020 by 20%. Indeed, in the UK the government has also passed an act to reduce greenhouse gas emissions from 1990 to 2050 by 80%.
These are major drivers for the rollout of smart metering across Europe. Other drivers are: concerns for energy security, flexible commercial arrangements for energy supply and a range of initiatives to enable energy to be generated and used more intelligently (support for micro-generation, intelligent smart grid, electric vehicles, smart appliances, etc). This has led to huge programs across Europe to roll out smart gas and electricity meters by 2020.
However, the programs have become much more complicated than originally envisaged, as they require buy-in and agreement from so many parties (governments, companies and consumers). Particular areas of concern have been the attempt to standardize interfaces (to create interoperable devices and systems) and the increased need for more security and personal data privacy. There is also a general acceptance that smart metering alone will not achieve significant energy or carbon reductions, unless it enables a growth in smart appliances and the smart grid afterward. The mass rollout of smart meters across Europe is now a huge activity. This article addresses a few of the issues it raises.
The ownership structure of the energy supply varies considerably from one European country to another. This has a major impact on how smart metering devices and systems should best be designed. For example, France and Italy have large monopoly electricity suppliers, but these companies do not supply gas. By contrast, the UK has six major energy companies who supply electricity and gas, coupled with a market that encourages consumers to switch supplier (and there are about 100,000 switches per week).
In some countries the meters are owned by the grid operator, whereas in others they are owned by the energy suppliers. This changes the rules that must be applied when a consumer switches from one energy supplier to another. This has a big impact on the security and access rights within a meter.
The funding of the smart meter rollout also varies from one country to another. In the UK it will be funded by the energy suppliers.
Equipment manufacturers want to reduce costs by increasing economies of scale and standardizing designs across the whole of Europe wherever possible. In practice this is possible at present for the metrological parts of a meter, but not for the new smart functions for communications, security and payment methods. The metrological parts must comply with the European Measuring Instruments Directive (MID), which sets down that once a meter has been approved in one European country, it can be sold across the whole EU.
There are now many European and national working groups trying to achieve similar levels of standardization and interoperability for the new smart functions. But it seems to be an uphill battle, and there is a general acceptance that many of these functions will continue to vary from one country to another. In addition, people are still learning about the implications of adopting smart metering, so the requirements continue to change month by month in all these countries.
Many manufacturers are designing their equipment in a modular (logical or physical) manner to cope with these regional variations. Such modularity is based on interface definitions (e.g., APIs, communications protocols, RF standards, data objects, hardware modules, full profiles, etc.). Companies normally want to adopt open standards, but sometimes find they have to define their own proprietary interfaces because there is no clearly accepted open standard, or because they want to retain control and confidentiality about a certain aspect of their design.
The energy suppliers also have a powerful influence over these design decisions, as they are normally the entities purchasing the metering equipment. They want to buy equipment from different manufacturers that is interoperable and can be used identically with their head-end computer systems. Such interoperability requires agreement and standardization on many interfaces (e.g., data objects, security, home area networks (HANs), wide area networks (WANs), etc.).
Possible future UK smart metering infrastructure
A question of standards
There is much debate about which data object standard should be used. In Europe, the main options are the Device Language Message Specification (DLMS) and ZigBee Smart Energy (ZSE). In the USA the main options are ANSI C12 and ZSE. This is not only important for interoperability, but also for security.
It is good practice to have end-to-end security between the meters and the operators’ head-end computers. An important security component is for the messages to have signatures at the application layer. There should not be any intermediate translation in the message journey (e.g., from the meter to the comms hub to DCC to the utility) as this would break the message signature (which is why security people prefer the link to be truly end-to-end).
The communication between the head-end—energy supplier or grid operator—and the meter is much more secure if both ends can verify the signatures attached to each message they receive. This means that the message cannot be translated between them, which means that the head-end and meter need to use the same data object formats. This is one of the reasons why it is so important to secure agreement on which data object standard and associated signature format should be adopted.
The communication solution also varies from one country to another. Much of the planning in the UK is being led by Ofgem, the gas and electricity regulator, and the Department of Energy and Climate Change (DECC). All 26 million UK homes have electricity and 22 million of them have gas. Figure 1 shows what UK infrastructure may look like after the smart metering rollout.
UmI module mechanical requirements
Source: Cambridge Consultants
The UK plans to change all gas and electricity meters to smart ones that contain two-way communications to the metering HAN. In addition it plans to install a communications hub and in-home display (IHD) in every household. It is hoped that the IHD will engage the consumer to be more energy-aware, leading to consumption reduction (rather like the fuel efficiency gauge in a car).
The communications hub will link to a national WAN, the metering HAN and hopefully a consumer HAN as well. The consumer HAN will provide smart metering information (e.g., time-of-use pricing) to smart appliances. The consumer could tell a washing machine when to finish and let it choose a low-carbon or low-price time to run.
The metering HAN will probably be wireless and could use ZigBee or Wireless M-Bus technology. There are fewer attempts at present to standardize the consumer HAN, but Wi-Fi is a possible solution. The WAN is a long distance link from all homes to a new organization called the Data Comms Company (DCC), and it will pass messages on to the relevant energy supplier. This means that if a consumer switches from one energy supplier to another, there is no need to change the WAN connection.
The WAN technology has not been chosen yet, and a variety of wireless and wired technologies are being considered. The WAN may end up supporting two logical data pipes: slow (say, 30 minutes) for energy suppliers, and fast (say, 5 seconds) for grid operators to run smart grid functions.
The communications solutions in France and Italy will look very different. The electricity and gas companies are not cooperating, so there will be separate WAN connections to the home for each. This is particularly challenging for the gas meters, which are battery-powered. These countries do not plan to create HANs within the home. There will be no IHD and no connections to smart appliances.
Another area of functional variation is prepayment. Most countries just use credit meters, which only measure how much energy has been used (kilowatt hours of electricity or liters of gas) so that a bill can be calculated. A prepayment meter contains a switch or valve, so that when it has run out of money, it turns off the energy supply. In the UK, the prepayment method used to be coins, but is now based on tokens (typically smart cards) and in the future will be based on token-less transfers from the energy supplier’s head-end to the meter (over the WAN and HAN), analogous to pre-pay cellular phones.
The plan for UK smart meters is that they will all be remotely software configurable to function as either a credit meter or a prepayment meter, without requiring any house visit or change of equipment. The addition of an off-switch to the meters adds considerably to the security risk of the system, so the strength of the security design has to be raised accordingly.
Different utilities, different needs
Electricity meters have been electronic for many years, whereas most gas meters have been completely mechanical, using a bellows sensor and mechanical index. With the advent of smart metering, electronic capabilities are being added to gas meters. Some are moving to electronic metrological sensors, but most are sticking to a mechanical bellows sensor (the main metal meter body). However the index, a plastic display box mounted on the front of the main meter body, is being changed from mechanical to electronic (Figure 3).
A smart meter for gas supply
Countries and states that use gas often want to adopt shared systems for the gas and electricity meters. In such cases, the key design issues are driven by the gas meter because it has so little energy to run from. Modern smart gas meters must last for 15 years from one or two AA batteries—an average consumption of 15uA from 3V—for all functions: metrology, communications, payment, user-interface and security. This is a considerable challenge, especially when the metrology complexity has increased to support time-of-use pricing. The meters must also be able to support energy price changes every 30 minutes.
If the smart metering system must have common interfaces for communications and security—which is what most countries want—then the interface selection must be driven by the gas meter and not by the electricity meter. Furthermore, many countries consume far more domestic energy through gas than through electricity. In the UK the average home consumes over six times as many kilowatt hours through gas as through electricity. So the opportunity for energy and carbon savings is also greater from domestic gas than domestic electricity. This is not true for all countries, but it is the case for many northern states and countries.
It is therefore ironic that most worldwide smart metering developments have been led by electricity meters, rather than gas. The problem is that many of the communications and security solutions chosen for electricity meters can never be supported by gas meters as they require far too much power. By contrast, all the communications and security solutions that can operate on gas meters can easily be supported by electricity meters.
Another interesting area is security and personal data privacy. The significance and requirement levels for security in smart metering systems have increased considerably over the past few years.
In the Netherlands, consumer and political concern over the lack of personal data privacy in the new smart meters was so great that the whole program was stopped and delayed for three years. Other countries are keen to avoid similar problems, so they are accepting that the bar must be set much higher for security and privacy in any smart metering systems they adopt. Nevertheless, California is now also seeing much the same controversy grow over the installations being sought by utility PG&E.
An important distinction is whether data is personal data or machine data. If it is considered to be personal data then it triggers certain aspects of European Human Rights law and must be treated appropriately. The Netherlands, Germany and the UK all now have strong national security groups defining the requirements for data security and privacy in smart metering equipment and systems. Many systems are considering full Public Key Infrastructure (PKI) technology based on asymmetric cryptography with certificate authorities and private/public key pairs for each device.
Most schemes are based on the U.S. cryptography standards developed by the National Institute of Standards and Technology (NIST). The use of Elliptic Curve Cryptography (ECC) 256-bit primes has already been adopted for European ePassports and medical records and has now been adopted by several smart metering standards. The energy consumption for ECC key exchange or signature generation/verification is quite high, so most systems use ECC to generate a temporary session key. This is then used for symmetric encryption (e.g., AES-128) of subsequent packets between the two endpoints in the session.
Good security design is based on end-to-end communications where the receiver can prove the identity of the transmitter and knows that the message has not been tampered with in transit. This is achieved with message signatures and device certificates from a trusted party. The validity of the signature is broken if the message is modified or translated at any intermediate node, so it is best if this can be avoided. The certificates bind a public key to a device ID, so it is important to adopt a standardized format for the device ID. A useful approach is to adopt the IEEE EUI-64 device numbering scheme, where each manufacturer can get its own OUI-24 prefix and then manage their own 40-bit suffix in production.
Uncertainty over R&D
The problem for the equipment manufacturers is that there is a lot of uncertainty and change in this market from region to region and from month to month. They cannot afford to make R&D investments that will quickly become out of date. One way to protect their development investments is to adopt a modular architecture (logical or physical) that allows them to separate the invariant functions from the variant. Stable interface specifications help them to achieve this goal.
Cambridge Consultants is a multi-disciplinary, multi-market engineering design consultancy whose activities include the development of many energy, communications and security devices over 50 years. Current smart meter designs pull together all these activities. Cambridge Consultants has developed three interface specifications to help companies to develop their smart metering systems. The Universal Metering Interface (UMI—Figure 4) specifications cover modules (based on SPI), opto-ports (based on EN62056-21) and security (based on ECC-256 and AES-128). UMI has already been adopted by several organizations, including Elster Gas in Europe. The specifications will be transferred to the UMI Alliance when it is formed, but are already available free to anyone who is interested today.
A ZigBee UMI module
The next few years will be interesting. There will be huge rollouts of equipment to every home in Europe and to many in the USA. This is a massive challenge, but it’s what we need to do, and as engineers, that’s what we like!
Alistair Morfey (Alistair.Morfey@CambridgeConsultants.com) is a Technology Director at Cambridge Consultants. He is a member of the UK DECC/Ofgem Security Technical Experts Group for Smart Metering. Read more about the UMI standard at www.CambridgeConsultants.com/umi.