The democratization of energy: How machine learning is empowering both the consumer and the utility

The advent of technological advancements in the power industry is fundamentally changing how we produce and consume electricity, improving grid management as well enabling the creation of new business models: Enel Green Power ('Enel') is at the forefront of this transition, and redefining what it means to be a utility company.

The call for innovation

For more than 100 years, the way we produce and consume electricity has largely remained the same. The need to innovate the power industry is being driven by multiple macroeconomic and social trends. The first is the call to a more sustainable and zero-carbon future, driven by the increased use of renewable energy, an intermittent resource. The second is the need to electrify more than one billion consumers globally who live without electricity [1]. Lastly, an increasingly digital economy will drive energy demand, which is expected to grow by 28% by 2040 [2]. Each of these trends requires energy providers to create a more balanced generation portfolio, ensure reliable power, and improve asset performance.

These challenges also present unique opportunities. The convergence of technology with physical assets gives utilities unprecedented access to big data, creating an interconnected system that is in constant communication. The figure below represents the information flow diagram of the ecosystem described above.

 

 Source: GE Power

By combining digital and physical assets, utilities are able to monitor load profiles in the system in real-time, while simultaneously adjusting grid operations without any human intervention [3].

Machine learning and digitization at Enel 

Machine learning allows for greater predictability, which helps Enel in two broad categories. The first is process improvement, primarily through an improved ability to monitor large industrial assets, which reduces operating costs. While Enel has created innovations in this regard on multiple fronts, some examples include:

  • Big Wind Data Boost: a program that puts sensors in wind turbines, which anticipates predictive maintenance, leading to increased efficiency and immediate cost savings [4]
  • Kaplan Online Optimization System: a proprietary algorithm that increases the efficiency of hydroelectric turbines by decreasing downtime through the automated positioning of turbine blades [5]

The savings from these digitization projects are significant. GE Power estimates that a digital wind farm can deliver up to $100 million in cost savings and increased efficiency [6]. While inward looking factors impact Enel’s bottom line, it is external opportunities that will create lasting value for the company. Due to the digitization of its assets, Enel now has greater visibility into the supply side of energy. In order to better understand the demand side, the company rolled out ‘smart meters’, which push detailed consumption information to the utility. The deployment of 40 million of these smart meters globally gives Enel a large data set, which in turn allows it to forecast demand patterns more accurately [7]. These devices also give the consumer more information on their energy usage, helping them identify potential savings.

Big data changes the utility business model

All data resulting from Enel’s digitization projects are centralized in a ‘data lake’. How then does the company translate all that data into lasting value? The primary benefit is that it is able to integrate more renewable sources into the grid without compromising long-term reliability. By accurately predicting supply and demand, it is more equipped to deal with the intermittent supply of renewable power and can therefore move to a more decentralized system, giving rise to new business models in the process; technologies of this model are outlined below [8]:

 

Source: World Economic Forum 

Indeed Enel has created multiple products in these key technologies. For example, it has commercialized an ‘Energy Intelligence Software’, which uses machine learning to help commercial clients automatically identify actions to benefit from energy-saving initiatives. It offers clients turn-key distributed generation systems, which allow customers to generate their own power locally. This gives the customer control over the energy used to generate the power it consumes, as well as creates a more resilient supply of power in the event of blackouts.

Building on progress 

As decentralized energy model penetration increases, so does the need for seamless interconnection between these systems. The company must therefore create the market for peer-to-peer transactions to allow energy to be traded across systems. To address this Enel is experimenting with blockchain through Enerchain, a software piloted by Ponton with over 24 energy-trading firms [9].

One of the biggest challenges facing machine learning is our increased dependence on computer systems, leaving companies susceptible to cyber security risk. In 2013, 40% of all cyber attacks targeted the energy sector; as a result, 91% of power generation organizations experienced an attack [6]. As the company increases the digitization of its industrial assets, it will increasingly be seen as a potential target, and thus necessitates taking preventive action.

The opportunity for a clean, sustainable future is upon us. Utilities now have the tools to transition toward a zero-carbon economy, with Enel at the forefront of this transition. As the industry increasingly adopts the digital model, how can it ensure the security of energy assets, which help form the foundation of society?

(783 words)

[1] “Energy Access Database”. International Energy Agency 2018. Iea.Org. https://www.iea.org/energyaccess/database/.

[2] “EIA projects 28% increase in world energy use by 2040”. International Energy Agency 2018. Iea.Org https://www.iea.org/energyaccess/database/.

[3]  “Digitization of the Grid”, Transmission & Distribution World, pp. 2–7. Wolf, G. 2016. Available at: http://ezproxy-prod.hbs.edu/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=117401300&site=ehost-live&scope=site.

[4] “The Winds of Innovation Are Blowing In Italy”. Enel Green Power 2018. https://www.enelgreenpower.com/stories/a/2018/06/the-winds-of-innovation-are-blowing-in-italy

[5] “The Algorithm Making Water More Efficient”. Enel Green Power 2017.  https://www.enelgreenpower.com/stories/a/2017/07/the-algorithm-making-water-more-efficient

[6]  “Powering The Future: Leading The Digital Transformation Of The Power Industry”. GE Power 2018. https://www.ge.com/content/dam/gepower-pw/global/en_US/documents/industrial%20internet%20and%20big%20data/powering-the-future-whitepaper.pdf.

[7] “Enel Experience In Smart Grids”. Montone, Alessio. Presentation, 2018.

[8] “The Future Of Electricity: New Technologies Transforming The Grid Edge”. World Economic Forum. 2018. http://www3.weforum.org/docs/WEF_Future_of_Electricity_2017.pdf.

[9] “European utilities: Harnessing the Power of Blockchain”. Laybutt, C. (2018). JPMorgan Chase & Company Equity Research Report. Retrieved from Business Premium Collection http://search.proquest.com.ezp-prod1.hul.harvard.edu/docview/2101195652?accountid=11311

Previous:

A Better Brain: Machine Learning in Guided Meditation

Next:

Siemens Mobility going digital

6 thoughts on “The democratization of energy: How machine learning is empowering both the consumer and the utility

  1. Thanks Tom_Challenger for the very interesting essay – you cover some very important points about the emergence of a smart grid and a decentralized energy model. I agree that the ability to interconnect both the demand and supply sides to allow for information sharing will be key. One of the major challenges utilities will face is dealing with an increasing amount of households adopting distributed generation with battery storage and largely disconnecting from the grid. Utilities will have less customers to socialize the costs of the transmission and distribution system across, and will need to reduce costs in other ways. In this situation, utilities like Enel will need to rely on these technologies mentioned (demand response, energy efficiency, storage etc.) to address potential issues of a more decentralized energy model.

    One area I envision significant opportunity is electric vehicles serving as distributed storage units to balance demand peaks and valleys. This will require significant information sharing between the end customer and the utility, but machine learning can be utilized to most efficiently draw from the vehicle battery at peak power prices and power the battery back up during the lowest cost hours. As a steep increase in the adoption of electric vehicles is likely right around the corner, this is an area that utilities should be following closely.

  2. I like that this article calls out both the benefits and the risks of digitization of energy assets. I am curious to learn more about what is required for Enel to build the market for peer-to-peer (P2P) energy trading. How robust is the blockchain solution that Enel is piloting in a P2P market and what will it take to achieve high levels of adoption and usage? In addition, I would like to know more about policy incentives that will either support or hinder the development of such a P2P market.

  3. Thank you for the very interesting essay.I share your concern in regards to the potential of cyber attacks that could potentially threaten the security of these smart energy grids. However, I would like to share that the US Department of Energy and a group of energy companies are collaborating to manage cyber security risks. Among the current interesting areas of development are [1]:
    a) Designing and embedding cyber security into smart grid at its foundation, with $ 3.4 billion of federal funding allocated to 99 smart grid projects to achieve this goal
    b) Monitoring: Installation of phasor measurement units to bring wide area visibility of grid operations
    c) Outage recovery measures, including advanced metering infrastructure and automated back-up storage

    Sources
    [1] Hawk C., Kaushiva A., Cybersecurity and the Smarter Grid, The Electricity Journal, Volume 27 Issue 8 (pp 84-95), October 2014

  4. I strongly support the trend of energy democratization as it will increase electricity access to more people. I also support clean energy production as it creates less or no pollution to the environment. Renewable energy especially solar PV faces with reliability issue as it relies on nature; however, now that we have more data about energy production and consumption pattern, renewable energy will play a bigger role in the industry to ensure both clean and reliable source of energy to consumers. To deal with the cybersecurity issue, I believe that blockchain technology will help a lot since it decentralizes the control of information to many different computers and make it harder to hack into the database. Most attention in blockchain is concentrated on the financial service sector, so I would encourage more attention to utilizing the blockchain concept in the energy sector.

  5. Dear Tom Challenger,
    Thank you for your informative article on distributed electric grids and how machine learning is benefitting Enel

    I would like to address the final sections of your article: the changing utility business model and building on progress. You did not mention the role which Enel plays in electricity markets – is it an electricity wholesaler, retailer, transmitter, equipment manufacturer, trader or grid controller? Each of these roles has a different application for machine learning, and its ability to use and collect big data. For example, grid controllers (managing and controlling the market for trading electricity, for the benefit of consumers) are in a position to collect data on production, consumption and the impact of price on the market, and make decisions that benefit consumers.

    However, certain agents could apply machine learning to game electricity markets for their benefit. To be able to transmit noticeable amounts of power between small distributed grids require upgraded infrastructure, that is mostly not in place currently. This gives distributors and producers incentives to manipulate their equipment (for example: deciding when to carry out maintenance) and influence the local electricity market. Machine learning will make this easier and smarter. Instead of democratising energy, we could be giving corporations the power to increase our electricity prices when we most need it. How will regulators and respond to this threat?

  6. It is clear that next-generation technologies are disrupting the way power generators, transmission companies, and utilities operate the electric grid. Enel has been quite effective in developing compelling business models (e.g., energy efficiency, distributed generation). However, as the grid becomes more digital the overall system faces a higher grid. However, the main focus of the company strategy is not security. Thus, I believe Enel needs to focus their resources and capital allocation into higher protection of the grid. For example, Consolidated Edison, another global utility, has made cybersecurity a focus, creating a dedicated security team that includes former law enforcement personnel. This will allow Enel to keep up with its competitor and better perform in the long term.

Leave a comment