Imagine you had to predict all your future needs for a resource based on a single datapoint; a day, a week, or even a month. Would that give you enough scope to predict how much of that resource you would need in a year, two, or even a decade? Probably not, but this is essentially the dilemma faced by utility providers today.
In some parts of Spain it is possible to fix your peak electricity usage to just 10 A instantaneous. It means you pay a much lower tariff but it does not allow for the use of several power-hungry devices at once. Now compare that to the more usual 60 A supply most homes have; it may provide ample power today but what about in five, ten, or even 20 years’ time when we are all expected to have many more ‘connected’ devices in our homes?
Managing this scenario has become known interchangeably as the Smart Grid, Smart Metering, and Smart Energy, putting in place an infrastructure that allows utility providers to better predict and manage the generation and delivery of resources including, but not restricted to, electricity.
AMR & AMI
Simply generating more energy is not the answer to meeting the future demands for electricity; the grid as it stands today is not capable of distributing significantly higher levels of power and homes are not equipped to accept it. The answer is to become smarter about how energy is generated, distributed, and consumed.
The key to the feasibility of better energy distribution through a Smart Grid is intelligence, both in the infrastructure and the meters. Known respectively as Advanced Metering Infrastructure (AMI) and Automated Meter Reading (AMR), an important aspect of adding the level of intelligence needed will be how these smart meters communicate, both with other smart devices and the overall infrastructure (see Prime Alliance).
Many options exist today for communication and, of course, the Internet is likely to play a crucial role. However, with the number of Smart Meters likely to far exceed the number of ‘regular’ meters already deployed, managing this new infrastructure will take more than adding a modem. Typically, this will create demand for microcontrollers and microprocessors that are adept at managing multiple forms of communication while also providing the performance needed to carry out complex metering tasks, all with the level of security necessary for the collection and transfer of sensitive (billing) data.
Another consideration is the interoperability of devices connected to and implicit in an advanced metering infrastructure; standards bodies around the world are collaborating with trade associations in order to reach an acceptable level of standardization across the industry even while the deployment of AMR/AMI solutions ramps up. In order to avoid massive and costly redeployment of equipment once industry-wide standards have been adopted, again, there is a need for a programmable solution in the form of today’s microprocessors.
The flexibility offered by microprocessors and microcontrollers, coupled with the ability to reprogram them in the field using the infrastructure’s connectivity (if necessary), is allowing OEMs to forge ahead with the design and deployment of next-generation Smart Meters.
Security
One of the biggest challenges of any network is providing a level of security appropriate to the perceived risk. For AMI the risk is quite high; not only will connectivity span large areas and installations, but it will also be largely unmanaged. Once the infrastructure is in place consumers and service providers will have ‘access’.
For this reason, building a high level of security into the infrastructure is paramount, and at the Smart Meter level this will include identification and authentication. Supporting these at the silicon level requires a device that offers support for cryptology, secure key storage, random number generation, and tamper detection. In addition, a platform that can provide trusted execution and secure debugging adds an extra level of security.
Freescale Semiconductor defines a trusted architecture as one that relies on a combination of trusted hardware and software to support a wide range of OEM-defined security policies. Freescale has implemented the trust architecture on the QorIQ P1010. Figure 1 shows the device’s trust features, which implement the NIST 7268 system trust model and provides mitigation against threats such as Theft of Functionality, Theft of Third-Party Data, and Theft of Uniqueness.
The QorIQ Data Path Acceleration Architecture prevents Theft of Functionality attacks caused by Denial of Service and Distributed Denial of Service, which can crash a system by flooding it with connection setup requests, by supporting fast-flow classification and policing. The defense of Third Party Data is supported through strong access control and encryption, while Theft of Uniqueness (often achieved through reverse-engineering to create counterfeits and/or clones) is prevented by the inclusion of a Freescale unique ID, which means even other QorIQs would be unable to boot the OEM’s code.
Figure 1: Freescale Semiconductor’s QorIQ P1010 provides a range of security features applicable to Smart Meters.
Concentrators
A significant aspect of the advanced metering infrastructure is (near) real-time reporting of energy consumption, providing a level of visibility into energy usage that has never before been available. However, aggregating this deluge of data will be a challenging task, and even though it may be technically possible for every Smart Meter to communicate back to the energy supplier individually, it would be impractical if not impossible for all meters to do this simultaneously.
The solution is to aggregate the data using concentrators; elements that currently have no counterpart in the energy grid. This new layer of hierarchy will be able to gather data from locally deployed meters, in the region of a couple of hundred to maybe a couple of thousand Smart Meters could be serviced by a single concentrator, which in turn interprets and relays the data back to the supplier.
Concentrators will also be capable of providing distributed control, to provide local energy management. Appliances typically present a variable load to the main voltage supply; by analyzing the patterns of variation in the data gathered using a suitable algorithm running on a powerful ARM® core, means the concentrator could build up a picture of where and how energy is being used in its own sub-net. This data could be used to influence the overall power generated by the supplier across multiple areas.
This is the kind of application Texas Instruments is targeting with their Sitara AM335x ARM Cortex™-A8-based processor shown in Figure 2. The company has developed a data concentrator evaluation module based on the part to give equipment manufacturers a scalable platform on which to develop solutions, as well as adding wired/wireless connectivity through a wide range of daughter cards.
Figure 2: Texas Instruments’ Sitara AMS335x range of ARM Cortex-A8-based microprocessors provide all the integration needed to develop a Smart Energy Concentrator.
Sub-metering
While the focus of a smart infrastructure is predominantly on electricity, the distribution and consumption of other utilities, specifically gas and water, will also benefit from closer control using Smart Meters. In addition, the use of energy metrology at the point of consumption will enable an era of ‘smarter appliances’ without the need to replace them. This multitude of intelligence will likely feed into a local aggregator or ‘gateway’, which will provide another level of analysis and control.
If consumers are expected to invest in or contribute to the cost of their own in-house infrastructure in the form of ‘sub-meters’, then it is imperative that the solutions are low cost. Fortunately, the performance available from low-cost microcontrollers today has reached a level that can easily cope with the processing requirements for this emerging application space.
STMicroelectronics has a clear strategy towards smart energy metering, which includes water/gas meters, as shown in Figure 3. In this example, sensor outputs are fed, via signal conditioning, to a control unit powered by either a 32-bit, such as the STM32L1, or 8-bit microcontroller, including its STM8L Energy Lite family.
Figure 3: Smart Meters will be extended to include gas and water meters, which must be developed using low-power technology like the STM8L Energy Lite family from STMicroelectronics.
The STM8L series is an ultra-low-power platform fabricated on a proprietary 130 nm ultra-low-leakage process technology and includes a family that has been optimized to meet the low power requirements of metering, including gas, water, electricity, and heat meters. The STM8L05 range has been further optimized to offer the best price/performance ratio targeting lower-end metering solutions, delivering 195 µA/MHz and 16 MIPS at 16 MHz.
As shown in Figure 4, TI is also focusing on this area and has produced an integrated solution (microcontroller with analog front-end, or AFE) that specifically targets metrology - the MSP430AFE.
Figure 4: TI is just one of the semiconductor vendors to identify the need for integrated metrology solutions.
Measuring electrical power in a smart or sub-meter, such as a ‘smart plug’, requires both the current and voltage to be measured accurately. Typically, the ‘sensors’ used in this application are simply resistors; high-ohmic resistors are arranged in a simple ladder network to divide-down the voltage, while a much lower-ohmic resistor is placed in series with the live conductor allowing the AFE to measure the voltage drop developed across it to calculate the current drawn. The MSP430AFE integrates up to three 24-bit Sigma-Delta analog-to-digital converters with programmable gain amplifiers alongside a 16-bit microcontroller in a single device (Figure 5). The ADC inputs support a differential voltage of ±500 mV with a negative voltage of up to −1 V, removing the need for level shifting. With a utility-grade accuracy of <0.1% over a wide range of load currents, the ADC also supports simultaneous sampling, eliminating delays between current and voltage sensing, and, subsequently, the need to compensate for any delays in software, as would be the case with solutions providing sequential sampling.
Figure 5: The highly-integrated MSP430AFE offers a 16-bit microcontroller and multiple Sigma-Delta ADCs to create a single-chip metrology platform.
TI has also created an Energy Library for the MSP430 family, which provides algorithms for energy measurement calculation. In addition, as a programmable platform, it is also able to accommodate standard or bespoke communications protocols, while the LCD interface and GPIOs provide a wide range of user-interface possibilities.
Conclusion
Building a smarter metering infrastructure is no longer seen as an option; the threat of rolling power outages in a bid to meet peak demand is just one high-profile aspect of energy management today. Generating more power is not necessarily a long-term solution, as the existing infrastructure will ultimately reach a point where it simply cannot carry any more power.
A sustainable solution is to manage the way power is used, in order to make best use of the capacity available. The drive for ultra-low-power solutions in the mobile domain is now bleeding over to the consumer and industrial sectors in general, as it is only by becoming more aware of power consumption that continuity of supply can be guaranteed.
The electronics industry will play an important role in smarter energy, by providing the solutions needed to monitor and control the distribution and consumption of power, with finer granularity.
References:
- http://www.prime-alliance.org
- http://www.nist.gov/smartgrid/
