15.1 Introduction

The previous chapter described at a broad level the concept of the Energy Internet of Things (EIoT); this chapter takes the discussion in the direction of practical considerations in the process of implementing the EIoT concepts covered in that previous chapter.

Fundamentally, the energy process entails generation, distribution, monitoring, control, and consumption. Each of these areas is currently experiencing innovation. In the generation arena, the promise of renewable sources of electricity, including the emergence of distributed energy resources (DERs) is significant. Renewable generation technologies are clean, abundant, and now widespread. Some of these technologies have low ongoing operational cost, although infrastructure investments are required to build the systems necessary to manage them. At this time, the electricity industry operates on the assumption that humans can control every detail of the production and distribution. However, renewable sources of power generate an amount of electricity that typically fluctuates from moment to moment. Furthermore, they generate electricity at the “wrong times” of day, creating availability when there is low demand and shortage when demand is highest. Thus, while renewable resources have been added to the grid in large quantities, wind, solar and microgrids intrinsically create management challenges due to unpredictable and variable energy outputs.

Distribution can also be a challenge since the solar or wind farms may not be in optimal proximity with either the consumer or the existing distribution infrastructure. There is a need to monitor all the energy-supporting elements of the grid. Additionally, it is desirable to manage usage and/or incentivize conservation; it is desirable to deploy home (and/or commercial establishment) devices, such as programmable thermostats, that manage electricity use while homeowners are absent.

As the power industry searches for a new paradigm of distribution and management that can incorporate renewables into the grid, the use of Internet of Things (IoT) devices can provide solutions by flexibly managing demand. IoT concepts specifically optimized to the energy industry are ideal. EIoT can alleviate the cited challenges by coordinating electric demand with electric supply; this can speed up the adoption of renewable and sustainable electricity technologies. EIoT in the home has produced devices such as programmable thermostats that manage electricity use while homeowners are absent. There is a broad industry support for EIoT technologies; Brundu et al. (2017); Elhebeary et al. (2017); Kaur and Sood (2015); Li et al. (2017); Minoli et al. (2017a, 2017b); Papaioannou et al. (2017); Saber and Khandelwal (2017); and Singh et al. (2017) provide an introductory, if limited, set of recent resources describing various aspects of the applicability and usefulness of IoT in the energy context.

This chapter describes a number of key developments that are required to facilitate the broad-scale deployment of a dynamic Smart Grid (SG), managed under EIoT mechanisms. Issues related to control of renewable generation sources utilizing IoT principles are discussed, as are industry initiatives in the context of required Information and Communication Technologies (ICT)/IoT standardization. Furthermore, the absolutely critical requirements for grid security are highlighted.

15.2 Managing the Impact of Sustainable Energy

The classical power grid relies mainly on electricity generation by stable, dispatchable, controlled sources, such as large centralized fossil fuel plants. Distributed renewables are sources of renewable energy produced by local communities and private homeowners and circulated through the common grid. Several examples of distributed renewables are listed in Table 15.1.

The proliferation of new energy technologies presents challenges and opportunities for operators. To examine these issues, it is important to understand the interconnection of energy generation and storage in a grid system. The local generation of energy by many renewable sources presents two challenges: nondispatchability and highly variable output. A nondispatchable energy source cannot be controlled based on the desired energy flow. A variable energy source fluctuates in the amount of energy it produces minute by minute.Therefore, adding large quantities of renewable energy to the power grid also creates large amounts of unpredictability.

Consider the case of distributed solar energy as a variable, nondispatchable renewable energy technology that is undergoing widespread customer adoption and needs to be integrated into the energy grid effectively. One example of this can be observed by examining the solar output fluctuations over 1 day in Oahu, Hawaii, as shown in Figure 15.1. The line is the amount of solar electricity production at 3 s intervals through a typical mostly sunny day. Note the extreme volatility of energy available for electricity production at one site. Complicating this variability is the need for energy to be generated and used almost instantaneously. While some believe current electric grids include the capability to store electricity, this is not the case: there is no significant energy storage connected to electric grids that is capable of retaining electricity for a period of time until it is needed. In the absence of this storage capacity, the amount of electricity produced must match the amount of electricity used at any given time of the day or year. If they do not match, there is a risk of poor power quality (brownouts), the shedding of customer loads from the electricity supply (blackouts), or even damage to the grid. To precisely match supply with demand, currently electricity demand is monitored and production controlled by a live system operator, who schedules what electricity is supplied every few minutes, as well as automatic regulation systems, which can adjust the flow of power to the grid in less than a second.

A typical residential demand curve is shown in Figure 15.2. The first curve shows a typical household's electricity consumption through 1 day. The second curve shows a typical day's solar production, scaled such that total solar production equals total household consumption. The final curve is household demand minus solar production, with negative values indicating power flowing back to the electric grid. Note the timing mismatch of consumption with production and also the increased volatility, taking standard deviation from 1.28 to 2.46 Wh. Specifically, this household consumes about 40 kWh/day. The standard deviation, a measure of the variability of the curve, is shown as 1.28 Wh. The variable solar output curve peaks in the middle of the day when sunlight is at its highest, while the household demand curve has peaks in the morning, as residents get ready to go to work, and in the evening, when residents return home to complete daily household activities such as preparing dinner, watching TV, or doing laundry (IEA Energy Conservation, n.d.).¹ Note that when solar production is subtracted from a household's load, the total load drops drastically. Excess power flows back to the grid during midday hours (hence, the negative values for net load), and power is demanded from the grid in the morning and evening hours. In this illustration, we matched daily solar output to daily demand, such that the daily net load would total to zero. This means that on net the solar electricity produced by the home is being used by the homeowners. However, notice that the variation of the net load has doubled to 2.46 Wh.

As new renewable generation technologies have matured and become economically competitive, the lack of storage capacity within our electric grids has presented challenges. Examples of the energy storage technologies are given in Table 15.2. Because electricity must be used as soon as it is produced, and until storage capacity can be added to the grid, the demand curve (electricity demanded by consumers) must match the supply curve (electricity provided by the available power sources). In the case of two prevalent forms of renewable energy, wind and solar, it is challenging to make the supply fit the demand without adding energy storage to the grid. EIoT is one way to do this.

In the residential solar example, if it is possible to store the energy produced, it would be possible even with a simple strategy to smooth out both the rates of production and consumption of energy. This hypothetical but possible scenario can be seen in Figure 15.3, which demonstrates a simulation that incorporates storage. The first curve shows one way to utilize storage over a day by storing solar electricity in the middle of the day and releasing it in the morning and evening hours. By subtracting this storage curve from the net load curve from Figure 15.2, we end up with net load 2, the resulting load after incorporating solar and storage to a typical household. Note how this net load 2 stays near zero during the whole day, but is still quite volatile, dropping the standard deviation from 2.46 to 0.73 Wh. Note that the battery charges during the day and discharges in the morning and evening hours. In this example, the battery releases 5.9 kWh of energy in the morning, absorbs 24.5 kWh of energy during the day, and releases 17.6 kWh of energy in the evening. The application of this crude storage strategy also drastically reduces variation in the net demand, reducing the standard deviation to 0.73 Wh. Storage is the most effective method to manage the nondispatchability of renewables, while drastically reducing variability of demand.

In the residential solar example, pairing dynamic, short-term reductions from EIoT devices with our storage strategy can provide the best example of EIoT devices managing the complexities of renewable energy. This can be seen in Figure 15.4, which shows a dynamic reduction simulation. The first curve shows how a fast-responding EIoT device could reduce loads quickly in response to solar output fluctuations. The second curve is obtained from subtracting the dynamic reductions from the net load2 curve from Figure 15.3. Note that the use of dynamically-varying loads drops the standard deviation from 0.73 to 0.61 Wh from Figure 15.3. By addressing the sharp peaks of the load and using devices to manage short-lived peaks, variation in net load drops further (down to 0.61 Wh), while maintaining low net energy flows to and from the grid, and balancing renewable generation and consumption. Fortunately, solar output has a good correlation with midday loads; however, solar does not always produce electricity during peak consumption times. Electricity prices are also high when loads are high. It is the case that the production price of wind and solar energy is very low compared to other energy sources. However, incorporating large amounts of non-dispatchable energy sources into the power supply may or may not contribute generation at peak consumption times. This fact can exaggerate the high prices of electricity at those times and will not reduce the capital costs related to generating peak loads.

In the residential solar example, EIoT gives us the ability to shift loads from times when renewables are not available to times that they are. Figure 15.5 shows an example of shifting a load, by changing the operating schedule of a single 2 kW appliance. The first curve shows the impact of shifting a 2 kW load from evening hours to midday when solar production is high. The second curve shows the load shift added to the net load from Figure 15.2. Note that even a very simple strategy like this can move loads in line with solar production and drop standard deviation from 2.46 to 1.94 Wh. The net result is similar to what happens with storage, and also reduces standard deviation, but is not as flexible and granular as storage can be. Still, load shifting is a cheaper alternative to storage, and is another tool to help adapt loads to renewable generation. Although it is difficult to closely match the demand curve to the renewable energy supply curve due to necessities such as use of indoor lighting during evening hours, minimizing disparities between renewable supply and electricity demand is key. Storage capacity that allows electricity to be saved for later use can be combined with EIoT technologies that allow part of the consumer load to be shifted to hours when renewable electricity production is at its highest. This is important because battery storage is currently too expensive to make it economically feasible, requiring technical innovation or mass production of units to bring down its cost in the near future. While the need for cost-effective energy storage is desirable to manage renewable energy sources, the use of EIoT to redistribute consumption loads can reduce the need for both additional energy storage and generation capacity at a low cost today.

15.3 EIoT Deployment

The benefits of using EIoT to manage and incorporate renewables into energy grids and to manage the consumption of energy and environmental impact far outweigh any of the drawbacks of these technologies. In a traditional grid operation model, dispatchable generators are directed to increase or decrease their outputs to ensure that supply matches demand. In an EIoT scenario, demand would be modified to match the supply produced by renewable and conventional energy resources. Internet-connected appliances would use sensors to monitor and alter consumer energy use. This would create dispatchability in consumer demand that would offset nondispatchability in renewable electricity supplies. EIoT devices could measure moment-to-moment variability in electricity generation and adjust their operations accordingly. Overarching patterns of renewable electricity supply could also be considered and accommodated. In combination with energy storage, this process could significantly reduce the need for fossil fuel peakers and allow consumers to avoid paying peak electricity prices.

EIoT devices can be used to enable the integration of renewable energy generation into the electric grid. The variability, unpredictability, and nondispatchability of renewable energy generation sources such as solar and wind make their integration into the grid a challenge. Much of the existing equipment on the grid for managing variations in voltage, load, or frequency is old, bulky, slow, manually operated, and/or not fast acting. Due to the high costs of implementing a modern smart grid, most utilities have avoided such expenditures. But with the connection of more and more renewables, the need for equipment to balance out such fluctuations is slowly mounting.

EIoT devices offer a streamlined solution to this problem. Instead of implementing utility-scale solutions to manage the grid, EIoT devices can respond to the previously discussed challenges of energy management, including the following:

• Variability. Figure 15.1, which shows one day of solar output in Oahu, Hawaii, demonstrates that renewable power sources can have extremely high variability of output, even on one-minute timescales or lower. In a traditional grid operation model, dispatchable generators (typically fossil fuel generators) would be directed to increase or decrease their outputs to ensure that demand was met with the necessary supply. In an EIoT scenario, renewable generation's variability could be directly measured by EIoT devices on the grid, sensors on the grid, or a central control. These EIoT devices could then respond quickly by increasing or lowering electricity loads, mitigating the negative effects of renewable generation variability on the grid.
• Unpredictability. The output of solar and wind tends to have some prediction error, especially in the short term. Predicted output is used to schedule other generator assets to meet expected loads on the grid. In a traditional model, when renewables do not provide the generation that was predicted, other dispatchable generation must be used to balance supply and demand. This dispatchable generation tends to be fossil fuel-fired generation, largely defeating the purpose of operating renewables. In an EIoT scenario, any surplus or deficit between renewable power output and forecast can be quickly utilized by EIoT devices. EIoT devices would receive communications that renewable output was not matching forecast and modify their operations appropriately.
• Nondispatchability. Renewable generation is generally nondispatchable, meaning that one cannot turn it on and off whenever one wants. Further complicating matters, due to the zero-variable cost of renewables, the grid always accepts renewable energy when it is produced. This causes a plethora of problems, such as the potential for negative prices, required larger generation reserves, and high ramp rates (like the “duck curve”²) for dispatchable generators. EIoT devices provide a unique, readily available, and fast-responding resource that can provide changes in load that respond quickly to changes in supply. This aspect of controlling demand, and not just dispatchable supply, is new to the industry and opens up a world of possibilities to managing the future grid.

The widescale adoption of EIoT requires the deployment of EIoT elements in many, if not all, of the SG components, including generation, transmission, and consumption. This supports the “I” and the “T” of the ICT enhancements required to make EIoT a reality. Additionally, a cost-effective but ubiquitous network infrastructure is needed to tie these elements together. This supports the “C” and the “T” of the ICT enhancements also required to make EIoT a reality. The subsections that follow discuss these two realms.

15.3.2 Network Functionality

Another basic component of the EIoT ecosystem is the network. In order for EIoT to be successfully implemented, electronic devices must be able to easily connect to the utility's extranet, to the “cloud,” and/or to the Internet (Paek et al., 2017). Specifically, for the SG benefits to be realized, all of the new EIoT devices discussed in the previous subsection need to be connected through various communications technologies to the control and computing infrastructure, often realized via the Internet, where information about device operations can be shared and optimizations performed. There are three strategies conceived for managing EIoT coordination: a centrally planned and controlled energy management platform, a web of independent control devices making their own decisions about optimization, or a compromise between the two that utilizes gateway devices. Cost-effective connectivity requires the creation of industry standards that allow devices to interconnect and the adoption of safety measures to protect the grid.

Central Control

Under the central arrangement in Figure 15.6, all EIoT devices would communicate via the Internet with one central energy management platform, likely hosted in the cloud. Examples of EIoT devices include smart thermostats, smart outlets, connected batteries, electrical vehicle chargers, and smart water heaters. This platform would take in sensor measurements, device operation schedules, price signals, and renewable energy generation levels. It would then perform an optimization of device operations to derive the best outcome based on price, carbon emissions, grid stability, and/or reliability. The platform would then send out operation instructions to devices in the field, enabling them to operate according to the platform's instructions.

Note that this central platform concept does not need to be all-encompassing. It can be used in small cases to manage a single utility feeder, or in large cases to manage price and stability in a Regional Transmission Operator (RTO) or Independent System Operator (ISO). The main point of this concept is that the device sends its information to a central platform that performs the optimization. This helps to focus the computing power in one location, rather than having large amounts of computing power deployed in smart devices, which would sit idle for a large percentage of time. Having one central computing platform could also enable more datasets to be utilized in optimization and more complex optimization methods to be performed, possibly requiring greater computational resources.

Web of EIoT

Under the web arrangement in Figure 15.7, each smart EIoT device communicates over the Internet directly with other EIoT devices. This illustration shows EIoT devices from two different homes and one business communicate with each other, not a central control, to optimize their operations. The solid lines represent communications flows (e.g., parametric data), while the dashed lines represent control flows. A cloud-based data source provides third-party data like weather and electricity prices. Management of grid irregularities are negotiated between EIoT devices as they collect grid data and communicate with each other about the optimal energy use for each device. This arrangement requires no central control to operate, but requires more communications bandwidth between a large number of EIoT devices. It also requires extra data storage and more powerful CPUs, as each device must now collect data and make optimization calculations itself.

Gateway

There is a continuum of options between the two above. Under the gateways arrangement in Figure 15.8, smaller local hubs, or gateways, can act as “localized” central control devices and manage EIoT devices in that building. The illustration shows two homes and one business with EIoT devices that communicate with gateway devices in each building. These gateways optimize EIoT operations, without the complexity of a full web control strategy. This arrangement reduces the need for each EIoT device to manage multiple connections and negotiations, as was the case in the web of devices strategy. Therefore, these devices can do with less powerful CPUs and memory. The gateway can house high power CPUs and extensive memory, and optimize each building's EIoT operations, while negotiating optimal strategies with other buildings. This strategy reduces communications volume, reduces EIoT complexity, but limits flexibility as each EIoT device depends on the gateway for operations scheduling.

15.4 Industry Standards for EIoT

In order to implement EIoT paradigms, it is clear that industry standards will need to be in place. Many different devices, produced by different companies and even different industries, will need to be able to communicate with each other, share information, and respond to control signals. While some generic IoT standards are now emerging, energy-specific standards are also needed (JaeSeung, 2016; Meddeb, 2016). Equipment and/or products from electric utilities, electrical equipment manufacturers, building controls companies, home appliance manufacturers, and indeed any Wi-Fi-enabled device manufacturer will all be interested in adhering to such standards. The reward for ensuring that all IoT devices adhere to standards and interface with the larger EIoT world is that a new stream of value will be unlocked by incorporating these devices into the electric grid. There have been attempts at producing EIoT data standards, which have yielded a range of success. The first and most successful has been the Open Automated Demand Response (OpenADR), a standard for communicating with smart thermostats; others include the Building Energy Data Exchange Specification (BEDES) and IEEE 2030. These are briefly described in the subsections that follow.

15.5 Security Considerations in EIoT and Clean Energy Environments

Security considerations should be a necessity when it comes to managing EIoT applications (e.g., see (Laszka et al., 2017; Liang et al., 2017)). As long as proper grid operation relies on the action of thousands or millions of individual devices working in harmony over an open Internet, there exists a risk that nefarious (or naive actors) can cause harm to the electric grid. Some of the main security concerns are as follows (among others):

• Cyber Intrusions. A harmful actor could target EIoT devices in specific areas and disrupt their operations to cause damage to grid equipment or cause a blackout in a specific area. The most well-known examples of harmful hacking activities involve industrial equipment and processes. Those devices tend to be relatively secure compared to EIoT devices, which essentially sit on the Internet exposed. This means that EIoT devices need to prioritize security as if the operation of the entire electric grid relies on them.
• Software Malfunctions. Utility planners will design (and have already begun designing) the electric grid to rely on EIoT devices to help manage peak loads. Utility planners are beginning to account for peak load response programs when planning the peak grid loads. This means that utilities are not installing equipment to serve load that is necessary, and instead relying on load-reducing EIoT devices that will respond to peak load signals. If EIoT devices do not deliver on peak load reductions estimated by utility planners, blackouts could result. It is important that EIoT devices act in a repeatable and predictable fashion so that utility planners can act in a prudent manner, reduce capital investments in the grid, and realize the value that EIoT can provide. An interruption at critical peak times can have damaging impacts on the grid.
• Protocol Flaws. EIoT devices will also rely on a plethora of communications pathways, technologies, and providers. Wi-Fi, ZigBee, GPS, fiber, cable, and 3G/4G/5G cell services will be utilized by EIoT devices as well as many other technologies. Any disruption in these services and core technologies can create problems for the reliable delivery of electricity to consumers.
• Operational Complexities. The control logic of EIoT devices needs to be carefully designed to avoid unintended consequences. Nonlinear effects can emerge in a complex web arrangement of devices. Specifically, adverse feedback loops could cause EIoT devices to operate in unintended ways. Devices might wait until grid capacity is available and inexpensive, increasing their load. This will in turn utilize available capacity, increasing price, and causing devices to decrease their load. The cycle then repeats. If enough devices act this way, unstable oscillations of load can occur on the grid. These oscillations are surely not the way consumers expect their devices to operate.
• External Dependencies. All of the EIoT devices installed in the electric grids will rely on many outside data sources, including weather data, weather forecasts, grid data, and more. Any disruption in these data streams could cause unintended operations of EIoT equipment, and possible interruption of operations. These data streams will also need to be safeguarded to ensure their reliability.
• Unintended Consequences. It is possible for EIoT devices to have an intentional or unintentional impact on the economics of the grid. If large numbers of the same device become attached to an electric grid, it would be easy for these devices to work in concert to exercise market power on the grid. In this scenario, market power is the ability to influence or control electricity prices.

The electric grid has historically been a very conservative endeavor. Great care is taken to ensure committed resources are available at all hours of the year to meet the demands of electric customers. It carries a critical role in the health of the economy and the well-being of the country's citizens. Ensuring that new technologies do not disrupt this reliable system is of the utmost importance and should not be taken lightly. The safety and security of grid operations should be carefully considered before implementing any EIoT device.

15.6 Conclusion

The increased deployment of green energy has a beneficial effect not only in the context of stewardship of the planet, but also for corporate or national economic reasons. There is strong global interest in deploying renewable energy sources, but also making sure that grid reliability is maintained. However, because green sources often are of the “variable energy source” type, very granular control of the resources is needed. EIoT principles and technologies are well suited to this task. This chapter described a number of key developments that are required to effectuate the deployment of EIoT devices, managed by the SG, managed under EIoT mechanisms.

Because naturally-produced energy can be unpredictable and highly variable, grid operators have two potential responses. They can supplement renewables on the supply side with fossil fuel sources that idle production to offset the variability of renewable energy production. Better yet, they could adjust load on the demand side using EIoT devices that can shift some consumer electricity load to times when renewable production is strong. The first option could result in high emissions, defeating the original purpose of incorporating renewables. Only the second option, the use of EIoT to adjust demand to match renewable supply, ensures that the benefits of renewables are realized. Equipping electrical devices with monitoring sensors, smart controls, and connecting them to the Internet will not allow EIoT appliances to manage their own energy use and be controlled remotely, but will in fact support the goal of energy flow reliability. Thus, EIoT devices will aid in managing the grid and incorporating renewable sources of energy into the electricity supply. They can provide a measure of demand-side control that will help offset the unpredictability of renewable energy sources as these are incorporated into the grid. This would reduce the need for peakers that represent costly infrastructure investments as well as additional sources of carbon dioxide emissions. Ideally, renewables combined with EIoT will transform the current electrical grid into the cleaner and more dynamic grid that we all deserve.

Standardization and security are two key requirements to move the process along. No matter what configuration is chosen, it will be important to standardize EIoT data so that all devices can “talk” to one another successfully. It will also be important to develop appropriate security protocols and apply these to all EIoT devices. There is a risk of harmful actors using EIoT devices to break into the electricity grid and disrupt the electricity supply. If the functionality of the electric grid depends on individual appliances, each of these devices needs to be appropriately secured.

In summary, as it was discussed earlier, injecting large quantities of renewable energy to the power grid is desirable, but it can create global unpredictability unless the process is properly managed: the variability, unpredictability, and nondispatchability of renewable energy generation sources make the integration of these sources into the grid a prima facie challenge, but one that is addressable. Related to this predicament it is desirable to manage usage and foster conservation by deploying devices, such as programmable thermostats, that manage electricity use while occupants are absent. Additionally, physical distribution can also be a challenge because the generation elements (solar and/or wind farms) are typically not close to the existing distribution infrastructure or to the consumer. EIoT mechanisms afford management capabilities that can “bridge the gap.” Regarding the variability challenge, with EIoT mechanisms the renewable generation's variability can be measured by EIoT sensing entities embedded in the grid, also in conjunction with a central control, enabling the EIoT-based devices to respond quickly by increasing or lowering electricity loads, thus mitigating the negative effects of renewable generation variability on the grid. Regarding the unpredictability challenge, a surplus or deficit between renewable power output and forecast can be quickly analyzed by EIoT sensing entities and EIoT devices will receive communications that renewable output is not matching forecast and that they should modify their operations appropriately. Regarding the nondispatchability challenge, EIoT devices provide fast responding control mechanisms that can provide changes in load to respond rapidly to changes in supply.

The expectation is that the challenges alluded to in the discussion above will be overcome in the next few years and the EIoT-based smart grid will become a practical reality.

References

1. Brundu, F.G., Patti, E., Osello, A., Del Giudice, M., Rapetti, N., and Krylovskiy, A. (2017) IoT software infrastructure for energy management and simulation in smart cities. IEEE Transactions on Industrial Informatics, 13(2), 832–840. doi: 10.1109/TII.2016.2627479.
2. Conti, F., Schilling, R., and Schiavone, P.D. (2017) An IoT endpoint system-on-chip for secure and energy-efficient near-sensor analytics. IEEE Transactions on Circuits and Systems I: Regular Papers, 64(9) 2481–2494. doi: 10.1109/TCSI.2017.2698019.
3. Elhebeary, M., Ibrahim, M., Aboudina, M., and Mohieldin, A. (2017) Dual-source self-start high-efficiency micro-scale smart energy harvesting system for IoT applications. IEEE Transactions on Industrial Electronics, PP(99). doi: 10.1109/TIE.2017.2714119.
4. IEA Energy Conservation in Buildings & Community Systems. (2015), Annex 42: The Simulation of Building-Integrated Fuel Cell and Other Cogeneration Systems (COGEN-SIM). Three-Second Daily Demand Data, Part 1: (IEA 3 Second Daily Demand 2_4_05_to_2_14_05 part 1.zip). Three-Second Daily Demand Data, Part 2: (IEA 3 Second Daily Demand 2_4_05_to_2_14_05 part 2.zip). Available at http://www.ecbcs.org/annexes/annex42.htm.
5. IEEE (2011) Standard 2030-2011 - IEEE Guide for Smart Grid Interoperability of Energy Technology and Information Technology Operation with the Electric Power System (EPS), End-Use Applications, and Loads.
6. JaeSeung, S. (2016) IoT standards toward its next stage. IEEE Communications Magazine, 54(7), 14–16. doi: 10.1109/MCOM.2016.7514158.
7. Kaur, N. and Sood, S.K. (2015) An energy-efficient architecture for the Internet of Things (IoT). IEEE Systems Journal, 11(2), 796–805. doi: 10.1109/JSYST.2015.2469676.
8. Laszka, A., Dubey, A., Walker, M., and Schmidt, D. (2017) providing privacy, safety, and security in IoT-based transactive energy systems using distributed ledgers. Proceedings of the Seventh International Conference on the Internet of Things, ACM, Linz, Austria. ISBN 978-1-4503-5318-2/17/10. doi: https://doi.org/10.1145/3131542.3131562.
9. Li, Y., Orgerie, A.-C., Rodero, I., Parashar, M., and Menaud, J.M. (2017) Leveraging renewable energy in edge clouds for data stream analysis in IoT. CCGrid '17 Proceedings of the 17th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing, Madrid, Spain — May 14–17, IEEE, pp. 186–195. ISBN: 978-1-5090-6610-0 doi: 10.1109/CCGRID.2017.92.
10. Liang, L., Zheng, K., Sheng, Q., Wang, W., Fu, R., and Huang, X. (2017) A denial of service attack method for IoT system in photovoltaic energy system. Network and System Security, International Conference on Network and System Security, NSS Springer, pp 613–622. (Part of the Lecture Notes in Computer Science book series (LNCS, volume 10394).)
11. Meddeb, A. (2016) Internet of Things standards: who stands out from the crowd? IEEE Communications Magazine, 54(7), 40–47. doi: 10.1109/MCOM.2016.7514162.
12. Minoli, D., Occhiogrosso, B., and Sohraby, K. (2017a) A review of wireless and satellite-based M2M services in support of smart grids. 1st EAI International Conference on Smart Grid Assisted Internet of Things (SGIoT 2017), July 11–13, Sault Ste. Marie, Ontario, Canada (With B. Occhiogrosso et al).
13. Minoli, D., Occhiogrosso, B., and Sohraby, K. (2017b) IoT considerations, requirements, and architectures for smart buildings – energy optimization and next generation building management systems. IEEE Internet of Things Journal, 4(1), 269–283. doi: 10.1109/JIOT.2017.2647881.
14. Paek, J., Gnawali, O., Vieira, M., and Hao, S. (2017) Embedded IoT systems: network, platform, and software. Mobile Information Systems, 2017. Article ID 5921523, doi: https://doi.org/10.1155/2017/5921523.
15. Papaioannou, T.G., Kotsopoulos, D., Bardaki, C., Lounis, S., and Dimitriou, N. (2017) IoT-enabled gamification for energy conservation in public buildings. Global Internet of Things Summit (GIoTS), Geneva, Switzerland, June 6–9. doi: 10.1109/GIOTS.2017.8016269.
16. Roberts, D. (2016) Why the “Duck Curve” Created by Solar Power is a Problem for Utilities. Available at https://www.vox.com/2016/2/10/10960848/solar-energy-duck-curve.
17. Saber, A.Y. and Khandelwal, T. (2017) IoT based online load forecasting. Green Technologies Conference (GreenTech), Ninth Annual IEEE, March 29–31, 2017. doi: 10.1109/GreenTech.2017.34.
18. Singh, N., Dayama, P., and Randhawa, S. (2017) Photonic energy harvesting: boosting energy yield of commodity solar photovoltaic systems via software defined IoT controls. e-Energy '17 Proceedings of the Eighth International Conference on Future Energy Systems, Shatin, Hong Kong — May 16-19, Proceedings Pages, ACM, 56–66. ISBN: 978-1-4503-5036-5. doi: 10.1145/3077839.3077857.
19. U.S. Department of Energy (2016) Building energy data exchange specification scoping report. Stakeholder & Technology Review, August. Available at https://energy.gov/sites/prod/files/2013/12/f5/bedes_scoping_080113.pdf.