Open Access is an initiative that aims to make scientific research freely available to all. To date our community has made over 100 million downloads. It’s based on principles of collaboration, unobstructed discovery, and, most importantly, scientific progression. As PhD students, we found it difficult to access the research we needed, so we decided to create a new Open Access publisher that levels the playing field for scientists across the world. How? By making research easy to access, and puts the academic needs of the researchers before the business interests of publishers.
We are a community of more than 103,000 authors and editors from 3,291 institutions spanning 160 countries, including Nobel Prize winners and some of the world’s most-cited researchers. Publishing on IntechOpen allows authors to earn citations and find new collaborators, meaning more people see your work not only from your own field of study, but from other related fields too.
To purchase hard copies of this book, please contact the representative in India:
CBS Publishers & Distributors Pvt. Ltd.
www.cbspd.com
|
customercare@cbspd.com
This chapter explores the applications of Digital Twin (DT) technology in the Brazilian energy sector and its impact on businesses and society. It highlights how DT applications have contributed to cost reduction, human error mitigation, operational optimization, and technical failure prediction. The chapter also discusses the implementation process and the requirements for developing these systems. Additionally, it explores the potential of leveraging Artificial Intelligence for decision-making support, utilizing Big Data processes to enhance various areas, and employing User Experience (UX) techniques to streamline outdated processes, through the examination of real projects in the wind power monitoring, transmission towers, and data-saving equipment domains, addressing the challenges faced and the benefits derived from its implementation.
Centro de Estudos e Sistemas Avançados do Recife (CESAR), Recife, Brazil
Urbano Chagas
Centro de Estudos e Sistemas Avançados do Recife (CESAR), Recife, Brazil
*Address all correspondence to: eldrey.galindo@cesar.org.br
1. Introduction
To understand the challenges of implementing the Digital Twin (DT) applications in the Brazilian electricity sector, we need some context about the country and it is a complex infrastructure of energy; for this, Table 1 compares the five biggest countries in terms of territory, where Brazil is the fifth with 8.515.770 km2 two times smaller than Russia the biggest country. However, the Brazilian population is third, behind China and the United States, and the ninth economy in the world [1].
Comparison between the five largest countries by territory in the world according to The World Factbook by CIA [1].
Values estimated for 2023.
Values in million kW and estimated for 2020.
Values in trillions quoted in 2017 dollars and estimated for 2021.
The Brazilian electricity sector is composed of four segments—Generation, Transmission, Distribution, and Commercialization—both public and private companies are responsible for maintaining the infrastructure and local operations under the guidance of the National System Operator (ONS). The ONS continuously monitors real-time information about the energy infrastructure, which comprises 23,441 operational power generation units distributed across the country. Figure 1 provides a visual representation of the distribution of these power generation units, revealing the complex structure. Table 2 gives us an overview of the distribution of energy sources used by the five biggest countries. The distribution of electricity generation sources amounts to these countries are similar with one main source upper than 59% of generation, for Brazil and Canada the main source is hydroelectricity, and for others, fossil fuels are the main.
Figure 1.
Print screen of the public data system of ANEEL showing the active generating units distributed in the Brazilian territory [2].
Generation sources
1°
2°
3°
4°
5°
Russia (%)
Canada (%)
United States (%)
China (%)
Brazil (%)
Fossil fuels
59.4
16.5
59.9
66
11.8
Nuclear
21
14.7
19.5
4.8
2.3
Solar
0.2
0.7
3.2
3.5
1.7
Wind
0
5.7
8.3
6.2
9.2
Hydroelectricity
19.1
60.8
7
17.8
65.8
Geothermal
0
0
0.4
0
0
Biomass and Waste
0.3
1.6
1.7
1.6
9.2
Table 2.
Comparative analysis of the electricity generation sources of the top five largest countries in terms of land area worldwide according to The World Factbook by CIA [1] (Estimated data for 2020).
In addition to other emerging countries, Brazil faces significant challenges in its electricity sector. In recent years, the Brazilian government and public agencies have been actively promoting the modernization of the energy sector on two fronts, the search for renewable and clean energies and technological modernization [3]. Comparing the 2020 CIA data with the most recent 2023 ANEEL data, Table 3, we note that solar and wind energy sources have increased during these 3 years. Technological modernization was through the adoption of advanced technologies such as Artificial Intelligence, Big Data, Internet of Things (IoT), and DT.
Comparative analysis of the Brazilian electricity generation sources between 2020 and 2023, where in bold there are solar and wind energy sources that have a considered increase during these 3 years.
These technologies aim to enhance the availability and predictability of the national electrical network. While the first DT projects in Brazil officially began in 2020, initiatives incorporating the concept of DT and utilizing data simulation, virtual monitoring, failure prediction, and power generation forecasting have been underway since 2009, as Brasilian law 13.755/Dez2018 better known as” ROTA2030” which encourages automobile modernization and increases the number of electric cars [4], and various research and development projects that the public data from the National Agency of Electricity (ANEEL) provides into this progressive development. In 2007, the Ministry of Mines and Energy (MME) published the National Energy Plan (PDE) a study and planning of energy evolution until 2030; in this study, MME estimates an annual average investment in the sector of 11.4 billion dollars [5]. According to the research by Wanasinghe et al. [6] about the trends, opportunities, and challenges of the oil and gas industry, among the 199 articles investigated between January 2003 and April 2020, Brazil appears with approximately 15 articles, occupying the fourth position behind the United States (59 articles), Norway (22 articles), and United Kingdom (16 articles).
Next, we will explore three real projects that we worked on and that will apply or have started to apply DT in this sector. We will not focus on the more technical part or the algorithms used but on the challenges encountered at the beginning of each project and mainly on those encountered by them.
Wind power has always played a prominent role in pursuing clean energy. From propelling caravels and mills to driving turbines, humanity has harnessed its power for many years. However, for a long time, controlling or predicting the intensity of the wind was a challenge, making it difficult to forecast the amount of energy that could be generated. Although improvements in weather forecasts and satellite imagery have helped mitigate this issue, a definitive solution remained elusive.
In 2017, a wind farm company approached us with a request to develop a system capable of predicting power generation in their turbines and supporting maintenance decisions by identifying potential failures in advance. We began by asking fundamental questions: How do wind turbines work? And what data do we need to predict their performance?
At a basic level, the wind turbine shown in Figure 2 consists of three main components: the blades, which harness the force of the wind and convert it into circular motion in the turbine; the nacelle, which houses the turbine’s motor and the necessary equipment to convert the circular motion into electrical energy; and the tower, which provides structural support to the entire system. The positioning of the wind turbine plays a crucial role in maximizing power generation, as it constantly adjusts the direction of the nacelle and the angle of the blades to optimize energy capture. Now, we can find the first data necessary to predict their performance, the angle blades, nacelle direction, wind velocity, and the amount of energy generated. Finally, when we study the turbine sensors deeper, we find that about 10 variables were used for turbine control and other variables linked with weather and generation system.
Figure 2.
Simple view of wind turbine showing the three main components (Nacelles, Blades, and Tower) [Authors].
Differently than what we initially thought, after we answered our fundamental questions and identified all variables we need to predict and improve the power generation, the following steps were not developing a machine learning algorithm or a simulation system, the project started with the aggregation of data and creation of a historical dataset. In the process of digital transformation, it is common to find companies at distinct digital levels, and in the electrical sector that is no different, many companies have their equipment with many sensors and exporting data all time, but these data are not processed or saved for a long time and the history is lost. The DT is a big concept formed of three basic parts: monitoring, simulation, and prediction, the first part is commonly implemented in this sector since many laws oblige these processes. The second and third parts are a challenge. Give me leave to do a reflection, if you are an owner of an energy company and you will buy some equipment for a new wind farm, you will search for the most resistant equipment that you can, equipment that does not break; after all, a broken turbine does not generate energy and this is loss money, to help in your quest; your energy supply contracts usually have high fines for non-supply. The reflection here is the equipment is made to not break its lifespan is very long close to 20 years, and for an ideal prediction model or simulation system, you will need to preserve decades of data at the cost of a lot of space. We have not considered on this scene that you save the correct information, the problems that occurred, the maintenance, and the manual optimizations, which are valuable targets.
For this project, we need to create historical data with some months using only the sensor data per minute, and this information was enough for us can predict failures and create alerts on an internal system to help operators in maintaining. In real operation, the application helped to reduce stops to maintaining and the generation continual. Other projects for prediction and optimization and simulation system using weather data were developed by the client.
In this case, the company had the equipment with all the necessary sensors and the know-how to operate and optimize it but had insufficient and unstructured data for a simulation or a prediction of power generated, but when the equipment does not have a sensor, it is impossible to have.
For the energy generated in a wind farm to turn on the light in our house, such energy needs to be transmitted through power lines. At first glance, this process is not complex. A person can use an extension cord to plug in a lawnmower or a lamp in the garden every weekend. New houses are integrated into the energy distribution system all the time, and there are power poles all over the city or along the roads. However, when there is a need to connect cities, big factories, states, and in extreme cases, countries, this infrastructure becomes more complex. The challenge of transmitting energy starts with the generation of energy. The energy generated in the generation unit cannot be stored or is very expensive to do so. In summary, all the energy generated needs to be distributed in real time for consumption, and, if the energy is not used, it is lost. Another obstacle is the distance, except for the small units or home generation, either the main units are built far or their capability to make energy is the biggest that the closed cities need, and the extra energy is distributed to other places. Additionally, energy is transmitted at high voltage to overcome great distances. This is the main distinction between the energy in the city and the energy for the city. Although high voltage is used for power transmission, it is not used for distribution. According to the norms of the International Electrotechnical Commission [7], energy voltage can be divided into three classes. Low voltage is used for consumption and is normally between 110 and 230 V. Medium voltage, between 1000 and 36,000 V, is used for distribution, while high voltage, above 36,000 V, transmission. In Brazil, there are power lines between 13.8 kV (13,800 V) and 800 kV (800,000 V) with the longest power line covering a distance of 2518 km [8].
Due to the high voltages and the high risk of accidents, there are several safety regulations for the equipment involved in the transmission and for the surroundings of the power lines. Some of these lines pass through forest areas, flooded areas, rivers, and mountains, which makes access for maintenance more difficult, and other lines pass through urban and rural areas, close to people, air traffic, and highways. The maintenance of these lines is a challenge, and companies need to use two or three techniques to inspect the lines and identify operational or security risks. One of these techniques involves flying over the lines with a helicopter, which is very specific and requires special training and equipment. This type of flight is expensive and risky due to the high voltage and the risk to the pilot and operators involved in maintenance. Another technique involves a team climbing the towers to inspect and take pictures, but when the lines are live, the people cannot climb up high and have a direct view of the failures.
For the new flow of data and the new interaction with maintenance, it was necessary to use user experience techniques to identify the best interactions. This research was focused not only on the interface and human interactions but also on data and the future implementation of Artificial Intelligence. Some companies still do not give due importance to the usability of systems or procedures. When we talk about new procedures involving data, Artificial Intelligence, and DT, these human interactions need to be even more important. When we talked about this project with drones, we had a completely different process than the one used by the client company, which in itself was a challenge. However, our goal was to go beyond creating a new process and to involve the creation of a digital model of high-voltage transmission lines and towers.
The designed process was divided into four stages:
A drone flew over a few kilometers of a line, taking pictures of towers, cables, and surroundings;
The images were automatically sent to the center, where they were stored and cataloged;
Specialized inspectors tagged the images and identified failures or points of attention, triggering the maintenance teams according to the need;
The images and information were stored, creating a database of photos and transmission line information.
As a result, a database for each tower was created, and for the first time, the company had a history of failures with degradation levels. After some tests, Artificial Intelligence will start to identify and classify the image degradation to assist the specialists, and in the future, the towers will have their digital versions, helping to predict failures or maintenance.
The design process used to map the stakeholders and discover user interactions revealed a new sub-function for the system that was difficult to envision. The system was excellent for training new specialists to find failures, and it is possible to find inspected equipment and request a person in training to reinspect it and compare the results. In addition, some inspections are complex or require more care, and the system was able to allow double inspections, which validate the inspection done and increase the quality.
Energy substations play a fundamental role in the energy system as they are responsible for voltage conversion and the delivery of energy for distribution. It is within the substation that the ONS (National System Operator) controls the flow of energy, defining routes and sequences of actions. For example, during a water crisis when thermoelectric power plants are used to supply the system, the ONS may request certain substations to connect to the power lines of these plants and integrate this energy with other energy sources. Another example is when a substation or a power line experiences issues and requires maintenance, leading to a disconnection of a part of the system. Substations have different equipment and specific layouts tailored to their activities. The structure for transforming high voltage to low voltage differs from the structure used to transmit produced energy. A typical transforming substation consists of main equipment such as circuit breakers, transformers, capacitor banks, and reactors. These kinds of equipment are organized and integrated into a circuit based on their function or load. Security equipment such as fuses and special circuit breakers are also present, and all devices are monitored by sensors that measure temperatures, voltages, amperage, and other data.
In 2020, we initiated a project focused on predicting equipment failures in substations and integrating this information with the internal system. All equipment is monitored, and all substations use a system with software and hardware elements called Supervisory Control and Data Acquisition (SCADA), which allows for the control of industrial processes, data monitoring, gathering and processing, and event logging [9].
Basically, an operation in a substation involves the operator monitoring raw data on SCADA and opening a Service Order (S.O.) if any parameter is non-compliant. Additionally, the operator frequently performs tests using a test device to check the integration of equipment. Early studies revealed a configuration error in the protective system, which was triggered when the temperature reached 51°C. However, this equipment functions normally until 65°C. This information raised curiosity as to why some alarms were activated and yet no action was taken. The answer was simple: human error. The engineer received the request regarding false alarms but found nothing unusual and ignored the request. Subsequently, the operator also started to ignore the alarms. Although this human failure was not serious, it highlighted a problem, which is a classic problem for data scientists. Humans struggle to process large amounts of data simultaneously and make correct decisions based on numbers. Machines, on the other hand, are better at processing such data and supporting human decisions. The algorithms were able to identify configuration failures and provide engineers with the necessary data to make more accurate decisions and precise configurations [10].
An overview of the data flow from SCADA to the control system is shown in Figure 3. However, the most important part, particularly for us, is the predictive modeling step, this is the very point where the DT occurs. To predict failures, the model establishes a normal state for the equipment and compares it with the current state. Based on certain parameters and time series data, the model can determine whether the current state is normal or abnormal. By simulating a future state, the model can precisely estimate how long it will take for the equipment to transit from a normal state to an abnormal state and why [10]. With this information, the engineer can assess the risk level and anticipate potential issues, whether they are serious problems with the equipment or secondary problems such as a sensor malfunction or communication failure.
Figure 3.
General application flow diagram published on the article data mining applied to abnormality prediction in electrical substation transformers [10].
Simulating the equipment and predicting abnormal states has brought many benefits to the company. The project’s results have had an immediate impact on detecting system adjustments and promptly identifying abnormalities, thus preventing equipment failures. Some benefits of the project, however, are not directly related to it. With simulated predictions, it is possible to anticipate more serious failures and optimize human and material resources. For instance, during application tests, the model predicted a failure in temperature-related equipment. When the engineer received this information and investigated, they discovered that a communication component of the sensor connected to the SCADA was starting to fail and would soon cease functioning. This raised an alert since the equipment would no longer be properly monitored, and replacing this component would take several months. By anticipating the order of the component, the time during which the equipment remained unprotected was reduced, as well as the time required for an operator to closely monitor the equipment. This resulted in significant savings in the operation of the substation.
If you try to develop a DT for a company independent of the area, there is a high probability that you understand that things are not easy as they appear to be at first sight. When a client requests a project for a DT, normally the phases “the data lake was done,” “the interface is simple,” or “it is just to integrate with the actual systems” hide great challenges. A system that interacts with others has a clear interface with simple interaction and a correct data flow is not created only by data scientists, data engineers, and developers; in general, Brazilian energy companies does a modest investment in usability and graphical interfaces, preferring cheaper and less user-friendly projects. On the other hand, Camara et al. [11] expose some benefices for the DT when the developed approach focused on humans and understands the real necessities of operators, since the real operation needs to resolve a damaged situation quickly and the reports are not updated in the same velocity, which reveals an asynchrony between the real operation and the digital data. There is a gap in the literature on studies considering this difference between real and theoretical operations and during normal and damaged situations.
Imagine a company that operates in the Brazilian electrical system, this company has some generating units such as a wind farm, some hydroelectric, and a solar farm, but it also has some substations and transmission lines. This company requests a DT for your wind farms, you know the technical challenges that we previously showed, but who is your user? What does the operator expect? In a critical situation, what is the real process? Those questions are not simple to answer, and for this, only a complete team composed of data scientists, data engineers, developers, and designers can find better answers.
Project human-centered is not the only challenge for the Brazilian energy sector, the integration of legacy systems and a data lake not structured are other problems for the DT project, and for this, the projects normally start to create a data lake and considering integrate with the legacy systems, as the examples shown in this chapter and the articles published about the DT on Brazil [12, 13, 14]. According to the research by Wanasinghe et al. [6] about the oil and gas industry in the world, there are 11 main challenges in this area, some of which can be brought to the Brazilian energy sector, as Scope and focus [15, 16, 17, 18]; Cyber security [19, 20, 21, 22]; Data storage and analytics [23, 24, 25, 26]; Business model, people, and policies [17, 20, 21, 27]; Incremental vs. disruptive [21].
Companies in the energy area have been changing their view regarding the importance of designers and the user experience in their projects, but it is still common to find projects that the client does not think this is important. Other challenges are the number of legacy systems and the decoupling of real operations and data.
Brazil is a very large country, with different types of terrain, climates, and ecosystems that bring natural challenges to the electricity infrastructure of the country. With numerous generating units, substations, and extensive transmission lines scattered across the nation, the adoption of technologies like Digital Twin, IoT, Artificial Intelligence, and Big Data becomes crucial for effective asset management. Furthermore, the government’s investments and initiatives in this sector promote the modernization of systems and equipment through the implementation of laws and national plans.
Although some project-specific details were not provided in this chapter, its primary objective was to showcase the challenges faced in developing real projects when companies lack the technical maturity required for a comprehensive Digital Twin implementation. The successful deployment of these technologies involves overcoming hurdles related to technology, user-centric system usability, and data availability. In addition to the articles cited, the examples presented in this chapter serve as real-world illustrations of applications that were developed to address specific challenges in the Brazilian energy sector, such as the absence of data lakes, cyber security concerns, integration with legacy systems, and operational studies.
In conclusion, the Brazilian energy sector is witnessing substantial growth and demands increased technological support to sustain and propel this progress. Energy companies in Brazil are actively pursuing the development of Digital Twins to improve equipment maintenance, enhance energy generation control, and streamline internal processes. However, it is crucial to recognize that certain companies may need to reassess their preparedness for Digital Twin implementation. By addressing any existing gaps in knowledge, infrastructure, or organizational readiness, these companies can effectively leverage the potential of Digital Twins to drive efficiency and innovation in the energy sector.
References
1.CIA. The World Factbook. Available from: https://www.cia.gov/the-world-factbook [Accessed: 09 June 2023]
2.ANEEL National Agency of Electricity. Sistema de geração de Informação ANEEL SIGA (Portuguese). Available from: https://dadosabertos.aneel.gov.br/ [Accessed: 13 March 2023]
3.CCBC – Chamber of Commerce Brazil-Canada. Brazil is living a favorable moment for modernization of the electric sector. Available from: https://l1nq.com/brazil-modernization-of-the-ele [Accessed: 16 June 2023]
4.Brazil. Lei N 13.755, Dez. Brasilia: DF Diário Oficial da União; 2018
5.MME. Plano nacional de energia 2030. Ministério de Minas e Energia; colaboração Empresa de Pesquisa Energética. Brasília: MME; EPE; 2007
6.Wanasinghe TR et al. DT for the oil and gas industry: Overview, research trends, opportunities, and challenges. IEEE Access. 2020;8:104175-104197. DOI: 10.1109/ACCESS.2020.2998723
7.International Electrotechnical Commission. Available from: https://iec.ch/homepage [Accessed: 27 April 2023]
8.ONS. Sistema Interligado Nacional – Rede de Operação (Portuguese). Available from: http://sindat.ons.org.br/SINDAT/Home/ControleSistema [Accessed: 02 June 2023]
9.Inductive Automation: What is SCADA? Available from: https://inductiveautomation.com/resources/article/what-is-scada [Accessed: 02 June 2023]
10.José Da Silva M, Melo De Souza S, Cavalcante De Lucena I, Da H, Santiago C, Galindo ES, et al. Data mining applied to abnormality prediction in electrical substation transformers. 2021. DOI: 10.37118/ijdr.23097.10.2021
11.Pinto SCD, Villeneuve E, Masson D, Boy G, Baron T, Urfels L. DT design requirements in downgraded situations management. IFAC-PapersOnLine. 2021;54(1):869-873. DOI: 10.1016/j.ifacol.2021.08.102
12.Fernandes SV, João DV, Cardoso BB, Martins MAI, Carvalho EG. DT concept developing on an electrical distribution system—An application case. Energies. 2022;15:2836. DOI: 10.3390/en15082836
13.Hayashi VT, Arakaki R, Fujii TY, Khalil KA, Hayashi FH. B2B B2C architecture for smart meters using IoT and machine learning: A Brazilian case study. In: IEEE - 2020 International Conference on Smart Grids and Energy Systems (SGES), Perth, Australia. 2020. pp. 826-831. DOI: 10.1109/SGES51519.2020.00152
14.Tjønn F. Are “digital twin through the life of a field.” In: Paper Presented at the Abu Dhabi International Petroleum Exhibition & Conference, Abu Dhabi, UAE. November 2018. DOI: 10.2118/193203-MS
15.Dattilo A, Vanderburg A, Shallue CJ, Mayo AW, Berlind P, Bieryla A, et al. Identifying exoplanets with deep learning. II. Two new super-earths uncovered by a neural network in K2 data. The Astronomical Journal. 2019;157(5):169. DOI: 10.3847/1538-3881/ab0e12
16.Morais D, Waldie M. How to implement tech in shipbuilding: Charting the course to success. In: SNAME Maritime Convention. Providence, RI, USA: The Society of Naval Architects and Marine Engineers; 2018
17.Parrott A, Warshaw L. Manufacturing meets its match. In: Deloitte Development – Industry 4.0 and the Digital Twin, New York, NY, USA, Tech. Rep. 2017. Available from: https://www2.deloitte.com/us/en/insights/focus/industry-4-0/digital-twin-technology-smart-factory.html
18.Parrott A, Warshaw L. Industry 4.0 and the DT. New York, NY, USA: Deloitte Development LLC; 2017
19.Zornio P. The control room is anywhere and everywhere: Putting the industrial internet of things to work offshore and beyond. In: Paper Presented at the Offshore Technology Conference, Houston, Texas, USA. April 2018. DOI: 10.4043/28943-MS
20.Sylthe O, Thornton B. The impact of digitalization on offshore operations. In: Paper Presented at the Offshore Technology Conference, Houston, Texas, USA. April 2018. DOI: 10.4043/28689-MS
21.Das R, Morris T. Modeling a midstream oil terminal for cyber security risk evaluation. IFIP Advances in Information and Communication Technology. 2018;542:149-175. DOI: 10.1007/978-3-030-04537-1 9
22.Mittal A, Slaughter A, Zonneveld P. Protecting the connected barrels—Cybersecurity for upstream oil and gas. New York, NY, USA: Deloitte Development LLC; 2017. Tech. Rep. Available from: https://www2.deloitte.com/tr/en/pages/energy-and-resources/articles/oil-and-gas-cybersecurity.html
23.Fujii TY, Hayashi VT, Arakaki R, Ruggiero WV, Bulla R, Hayashi FH, et al. A DT architecture model Applied with MLOps techniques to improve short-term energy consumption prediction. Machines. 2022;10:23. DOI: 10.3390/machines10010023
24.Sharma P, Hamedifar H, Brown A, and Green R. The dawn of the new age of the industrial Internet and how it can radically transform the offshore oil and gas industry. In: Proc. Offshore Technol Conferences. 2017. pp. 1-7
25.Daily J, Peterson J. Predictive maintenance: How big data analysis can improve maintenance. In: Supply Chain Integration Challenges in Commercial Aerospace. Cham, Switzerland: Springer; 2017. pp. 267-278
26.Ottonelli J, Cruz U, Rosa A, Andrade JC. Oportunidades e desafios do setor de energia solar fotovoltaica no Brasil. Revista Econômica Do Nordeste. version. 2021;52:8-26
27.Ottonelli J, Cruz U, Rosa A, Andrade J. Opportunities and challenges of the photovoltaic solar energy in Brazil. Review Economic NE, Fortaleza. 2021;52(4):8-26
Open access peer-reviewed chapter
