The ability of communication (and therefore coordination) between grid components can be seen as the defining aspect of smart grids in contrast to conventional power grids. Therefore, communications play a major role. As outlined in Section 2, the primary reason why communications are needed is the more decentralised structure of power grids that results from a wide application of renewable energy resources. Conventional communication strategies in the grid, such as power-frequency control, are designed for a low number of active nodes (adjustable generators) in the grid [12]. More distributed systems require a higher level of coordination in order to maintain stability [13]. However, not only power balancing requires communication. Besides today's Supervisory Control and Data Acquisition (SCADA), there are three major application domains for communications in smart grids.
3.1. Active Distribution Grids
To allow the integration of a high density of distributed generation in existing medium-voltage infrastructure, an active control of generation power on the basis of voltage or power flow measurements can be utilised at critical points in the grid. One of the main barriers for connecting new generators to the grid is that power feed-in increases the grid voltage at the feed-in point. This is in particular the case in sparsely populated areas, where the grid is not very strong, but a considerable amount of renewable energy resources is available. The line voltage has to be kept in an allowed band (e.g., +/− 10% of nominal value) by the grid operator in any case. The worst case occurs when there is no load but strong energy generation on the feeder.
In an active distribution grid, the generation of the distributed generators is managed according to the voltage at critical points. If the voltage rises too high, reactive power management is performed. This is done based on voltage measurements. If this is not effective enough, even the active power can be curtailed [6]. The selection of generator to be curtailed can be done technically, but also economically (price balancing). Such an active management of generated power in a medium- or low-voltage feeder is basically a form of multiobjective control. The challenge here is that sensors, controllers, and actuators are very far from each other. Voltage and power information has to be communicated over dozens of miles once every six seconds or so. The automation infrastructure used has to be highly reliable. Often, the protocols are transported over a variety of different media, depending on the available communication links. For a wide application of such approaches, a common information model for generators and other systems that are part of this control is required (see, e.g., [14]).
3.2. Smart Meters
Smart meters can be part of a smart grid, but they are not the same as smart grids. Although the origins of smart metering technology lies in remote meter reading, many other aspects play a role for smart meter deployment than an automated approach to track consumed kWh. Smart meters are primarily though to inform the energy consumer about his/her consumption and the current electricity tariff. According to economic principles, only with this information the end user becomes a rationally acting market participant and the market-driven optimisation of the energy system can work. Besides the need to track down electricity theft in some countries, the implementation of this principle is the key driver of smart meter deployment at many places of the world.
While in some countries smart meters are area-wide deployed, in other countries the debate about their benefits is still underway. On the positive side, these systems simplify the accounting, and consumers can be promptly informed about their energy consumption. More data is available from the grid, and network development planning can be done on the basis of real data instead of worst case models. Failure detection becomes easier and voltage bands can be used more efficiently. On the negative side, the costs are very high, and it is basically assumed that the consumer will pay the price. Further, there is a severe lack of standards. Long-term reliability and data security questions are not yet completely answered.
From a technical point of view, smart metering systems can be enabled to feature more than consumer information and remote meter reading. Smart metering systems can generate snapshots of the consumption state of the whole grid so that grid operators can examine in detail how much power was flowing to where in the moment of the snapshot. Further, it is possible to measure real consumption profiles, perform on-line power quality monitoring and even remote switching of loads.
Smart meters are interconnected by means of communication links, usually narrowband power line communication to data concentrators at the transformer stations. From here, backbone networks (e.g., glass fibre) bring the data to control centres. Existing communication infrastructures are an essential precondition of smart metering systems, but in many grids they are still nonexistent.
3.3. Automated Demand Response
One of the greatest challenges for a broad utilisation of renewables is to maintain the power balance under the presence of highly fluctuating load and generation. Basically, there are three options to solve this problem: conventional backup, energy storage and automated demand response. Today, the most widely used solution is conventional backup, where a drop of generation from renewable is replaced by powering up fast (but inefficient) conventional plants. However, with a rising share of renewables in the system, this option gets more and more inadequate. An alternative in future can be a wider utilisation of storages (even batteries of electric vehicles) in combination with control actions on the demand side [15].
There are several different possible ways of interfering with the operation of electrical loads (see, e.g., [16]). The simplest form is load shedding. The main goal in load shedding (or curtailment) is to cut or reduce loads in critical grid situations without consideration of the user process functionality. Load shedding that is motivated by grid stabilization is usually interfering with customer interests and can only be applied in emergency situations. A minor group of loads, such as lighting in unoccupied rooms, and so forth. can also be curtailed without loss of user comfort. However, in the context of sustainable energy usage, these loads should be switched off in general rather than be used in demand response operations. Modern demand response has to undertake its measures without noticeable changes in the performance of user processes. Therefore, load shifting is the preferred option of intervening with load operations.
Since they are seen as a supporting tool to match supply and demand under the condition of supply from fluctuating renewable energy resources, control measures on the demand side of the power grid play a key role in most "smart grid" visions. Of course, it also makes sense to adjust the generation to the supply by featuring generation technologies where this is possible, for example, for residential combined heat and power (CHP) systems.
Based on their specific processing, properties, and energy storage functionality, there is the possibility to reschedule energy consumption of certain loads. Energy can either be stored in real energy storages, such as thermal storages, or as conceptual energy storages that can be exploited by rescheduling a process to a later point in time (load shift) [17]. Load shifting can be performed in various processes, for example, washing, cleaning, heating, chilling, and pumping. These electricity-consuming processes have, depending on the application, certain degrees of freedom in their time schedule. Each potential flexible load has a certain individual capability that it can commit to the overall system. This capability has usually two dimensions: energy and time. The consumption of certain portion of energy can be pre- or postponed for a certain time. The method can be used for peak reduction, but also for other options such as the provision of short-term balance energy. Load shifting does not aim to reduce energy consumption in long-term, but to reduce peak loads by shifting consumption to off-peak times (see Figure 2) or to adjust it to volatile generation patterns. This can be achieved either with a human-in-the-loop approach (e.g., time-of-use tariffs), where the energy customer gets incentives to shift consumption herself/himself, or it can be done in an automated fashion, where the customer is out of the loop.