Artificial Intelligence and the Power Sector
- Jan 9, 2023 11:43 am GMT
Everyone has been talking of Sustainability in any business with due respect for environment. It is envisaged that the world has great potential to get transformed into sustainable and inclusive by 2050 i.e another almost 3 decades from now. In order to achieve this, the vital ingredient is solid economic growth capable of huge investments in sustainability and inclusion. Unlike the previous decades, growth alone cannot achieve this as innovative solutions will also be critical. Considering the importance of businesses in world economy which drive more than 70% of global GDP are championed to lead the innovation.
Current focus being net-zero emissions of GHG, natural capital and biodiversity, Global future hinges on three vital factors – sustainability, inclusion and growth. Inclusion is defined as economic empowerment, opportunity and progress to be shared by everyone. Economic activity would expand through greater productivity and innovation with financial resources and new forms of growth. This is not as easy as has been spelt out – any vision for future would bring with it tensions and trade-offs. However, this could help explore practical questions and urgent public debates that presumably are under way.
If one considers child mortality as an example for inclusion, we could say that the progress is indeed positive so far – the rate has been reduced almost by 60% since 1990 while the average number of years that is spent in school has increased by 50%. The share of world living in extreme poverty has also decreased by 70%. These are further supplemented by lower ozone depletion and in a few countries through energy efficiency and renewable energy sources.
There is however stark differences in growth, sustainability and inclusion among low income, middle income and high income countries which are quite obvious. The goal that people spend 12 years in school by 2030 in US falls short as 6.5 billion people less wealthy live in countries. Similarly, aspiration of digital inclusion (people using internet) equal to 100% worldwide which is smaller in many countries leaves it at three billion people with internet access. Aspects like life expectancy, child mortality, gender parity in labor participation and financial inclusion are others that go unfulfilled.
APPROACH ON SUSTAINABILITY:
Sustainability goals are no better – the goal towards net-zero GHG emissions by 2050 limiting temperature increase to 1.5 degree C, over pre-industrial levels is far from its reality. The demand that each year emission be no more than 3.0 metric tons of CO2 per capita by 2030 exceed in many countries leading to worldwide emissions to rise.
Artificial Intelligence (AI) and Machine Learning (ML) seem to be transforming hard-coded algorithms characterized by computers and processing a given input into an expected output.
If one wants to process historical data and predict manually or through traditional tools, it would perhaps, be tedious. Machine Learning does it faster and can even factor large number of variables as well. A code written by the manufacturer provides an alert for preventive maintenance at specified intervals but, fails to analyze whether the car or any of its components actually require maintenance. With AI, you can predict a specific component showing signs of malfunction as ML detects anomalous behavior based on historical data. Such advance warnings to help unexpected breakdown and such predictions are indeed applicable to different parts of manufacturing plants.
It is estimated that almost 75% of enterprises will utilize AI by the end of 2024 with good reasons:
Chatbots is one such application that is aimed at customer relationship for business users to facilitate an insight into production improvements. E-commerce is another extension where display of products suiting buyer’s preferences based on profile and purchase behavior. Computer vision using ML analyses uploaded images to organize products under different brands, categories and sizes. Retailers can benefit in not only forecasting demand but also pricing as well towards discounts without impacting overall profitability.
Similar applications are also possible in enhancing workplace productivity, better human resource management. Many are indeed conversant with the Netflix / Amazon shopping.
As can be seen, everyone – consumers / businesses seem to be enjoying the advantages of AI in combination with ML. In fact, social distancing did provide a great opportunity to this transformation. Simple statistics of AI/ML market rise from 3.2 billion in 2016 to 89.8 billion dollars by 2025 should provide an idea of its progress over the years. Their spread across consumer market would directly impact web development as well. The AI enabled chatbots are the next-generation digital assistants of the corporate world making customer – business relationships much stronger. It is also extended now to Natural Language Processing (NKP), identifying patterns of human conversations and converts them into actionable insights.
UMANG (Unified Mobile Application for New-Age Governance) launched in 2017 by Ministry of Electronics and Information Technology has been serving Indians with a single point of access to over 1200 major government services – education, COVID-19 vaccinations, Public transport and so many.
AI IN POWER SECTOR:
The advent of steam power propelled the world into the industrial Age. The use of power enabled mass production in the Second Industrial Revolution leading to advancements in transport, telecommunications and manufacturing. Digital age of the Third Industrial Revolution marked the proliferation of internet and internet of things facilitating widespread use of information technology to automate production. World Economic Forum says that we are into the fourth revolution which is distinct in velocity, scope and system impact with scientific and technological breakthrough. Geneva based public-private international cooperation which serves as adaptability platform drives the Fourth Industrial Revolution. ‘Software Engineer’ is artificial intelligence concept which will have widespread potential across industries. This has proved the potential to rethink integration of information, data analysis and insights to improve decision making.
Power industry is no exception as AI enabled digital solutions has been playing a major role. This is believed to be especially useful dealing with current issues like, DE- carbonization and decentralization. One thing is certain that AI application indeed is new and constantly evolving. McCarthy, the founding father of AI talked about ‘Science and engineering make intelligent machines’ in a conference of 1956. However, AI today has extensions in mathematics, statistics, cognitive science, philosophy and even linguistics.
Being human like (not human) it enables a vast set of capabilities as a tool. Artificial Intelligence and Technology Office (AITO) established by US Department of Energy (DOE) serves as a hub to co-ordinate AI development, delivery and adoption and even capability to solve problems. Machines rapidly learn from large data sets, solve problems and adapt to new data without human intervention which, is the beauty of AI. Computing power, access to data and improved algorithms help AI’s evolvement with growing opportunities.
Machine Learning used interchangeably with AI is data-driven computer algorithms that improve through experience. Currently, it encapsulates natural language processing (NLP), deep learning and neural networks. An Artificial Neural Network with multiple hidden layers is what ‘deep neural network’ is. Performance of one asset to performance of the fleet and eventually to the performance of entire electricity network (generation to consumption by consumer) is what is aimed at in the future.
It is believed that AI could take simulations to next level – helping utilities locate unstable areas of the grid and enhance field worker safety. There is no doubt that the power industry offers tremendous opportunity in daily operations, grid safety, reliability and many more. It would be interesting to see how many current challenges like leveraging low-carbon resources, forecast improvement and boosting autonomous operation are likely to be addressed. It is also hoped that AI has potential to innovation as well.
However, experts believe that AI in power sector is still at the infant stage and a long journey before one can reap potential widespread benefit. Not only the operational challenges but even standards for safety, quality and testing could prove challenging for integration. Technically sound evaluation of AI solutions need be seriously looked at prior to chancing AI in the power sector.
Among the various countries – UK, European Union, France, Germany and China who have already given considerable importance to AI in power sector, US seems to be heading the list with many properly oriented goals : National Artificial Intelligence Research & Development Strategic Plan; Sixteen separate agencies govern sectors of economy related to AI technologies; Major research hubs (San Francisco Bay area, New York-Boston); Largest market share for existing AI companies and large pool of qualified talents. Everyone believes that modernization of energy sector is indeed critical to economy of every nation. It is strongly felt, “Whoever control the strongest artificial intelligences controls the world”.
The following are a few areas of Energy Sector where, AI has potential:
Challenges of AI would be when fossil fuel generation and renewable share equal share of power in any country. While it can provide information related to meteorological data for a judicious use of renewable, it can also support smart grids help balance the power flow.
Challenges of AI in the energy sector:
Theoretical background: Being professional many don’t have sufficient technical background of how AI can help the industry as they rely more on time proven methods rather than risking. Looking at the contributions that AI is making in other fields, interest has been created of late, even in the power sector.
Practical expertise: Despite the in-depth theoretical knowledge, lack of AI-powered software is something that is not clearly missing. The sector presents many unknown domains in data loss, poor customization, system failure and unauthorized access. The cost or error deters the professionals taking risk in this new approach.
Infrastructure: The current status of outdated infrastructure despite voluminous data (disorganized, scattered across different formats and stored locally) and coping with it is another stumbling block. Add to this the vulnerability of outdated systems.
Finance: Though it may become economical over a period of time, the innovative smart technology is certainly not cheap currently. It does consume considerable time for experienced software service provider to develop and customize, adjust, manage and monitor. It is therefore, necessary for the energy sector to plan systematically starting with a good understanding on machine learning, their strategies and allocate impressive budget with a certain amount of risk factor before championing the technology.
AI seems to have attracted the attention as one of the most exciting technologies while acknowledging that the adoption varies considerably between not only countries but even industries. The governance of AI or creating policies addressing negative implications would perhaps continue until capabilities and impact of it on human lives are clearly defined.
With greater challenges that the sector is facing on ‘DE carbonization’ to contain global temperature, it is imperative that it has to start somewhere before turning into a boon to the sector. Hope someone would take the lead or software engineers get into the intricacies of the sector in developing appropriate software and experiment it on a trial basis.
Careful design along with the emerging new technologies would indeed be useful in automation of routine demand patterns that would lead to grapple with generation, transmission, distribution and consumption challenges in power sector of tomorrow.
Even if one considers the age groups around 2000 and forecast for 2050, there seems not much of difference in their percentages except marginally – 65+ goes up by almost 9% to the 2000 level of 7% but, age group of 20-64 shows just a marginal difference of just 3% compared to 2000 figure of 54%. But, 2050 would be the time that AI and ML would have invaded every sphere of human life and I wonder what would happen to 57% age group who may perhaps have to be idle.
Get Published - Build a Following
The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.
If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.