Integrating AI and machine learning with energy will help speed up the adoption of renewable sources
Covid-19 may have overwhelmed us, bringing lives and livelihoods to a standstill but it is not the biggest problem confronting the world.
A far greater crisis stares us in the eye threatening the very existence of humankind: Climate change. To speed up the energy transition process now it is necessary to integrate Artificial Intelligence (AI) and Machine Learning (ML) with energy. AI is not just about energy management; it can be an effective tool to counter climate change in consonance with our Sustainable Development Goals.
The energy sector often requires a huge infrastructural presence to operate. It also produces massive amounts of data. Artificial Intelligence can turn this data into insights, bring in efficiency and cut down costs. Major energy players right from the field of oil and gas to renewable energy are turning to AI to streamline operations. The US and Germany have already deployed such AI systems to bring in efficiency.
For instance, General Electric uses the AI analytics platform to monitor the performance of wind turbines. DeepMind, a subsidiary of Google, has been applying machine learning algorithms to 700 MW of wind power capacities in the US.
India has accelerated the pace of renewable energy installations to enable last mile connect for electrification. Increasing the share of renewable energy in the energy mix brings in a new set of problems such as grid stability. The country also has challenges in terms of managing the energy demand, which AI applications can minimise.
Applications of AI
Grid infrastructure and stability: The growing utilisation of renewable energy sources (RES) and their development in recent years poses key challenges for power system operators. The dependence on sun and wind, for instance, makes grids unstable. It may so happen that a cloudy day will not generate enough power to satisfy the energy demand or that on a sunny day power generation will exceed the demand.
With integration of AI, this pattern can be predicted early and thus grid adjustments can be made by automating operations accordingly. Automation of grids with real time control and advanced load control systems will bring in agility in operations. The smart grids and smart meters are the predominant features of the AI systems.
The current energy mix for power generation is quite diversified. In the past it was dominated by coal. In recent years, several other sources are being added. Solar and wind energy have increased their share significantly. Renewable energy sources now contribute a quarter of the share.
This also creates the potential to install hybrid energy systems. Especially for the construction of microgrids and minigrids which can operate in isolation. Hybrid energy systems are integration of various renewable energy source generators and battery storage system. This integration can be seamlessly achieved with AI systems.
Energy Storage: This is an integral part of renewable energy, especially when we speak of grid independent energy and uninterrupted power supply. Whether it’s solar or wind, the two major sources contributing predominantly into the energy mix, both have limitations as they operate according to weather conditions.
Artificial Intelligence finds many applications in energy storage systems. Remotely monitoring and maintaining the batteries is one of them. The smarter the energy storage, the more efficient would be the renewable energy system.
Again, with collection of data, predictive analytics can help better understand the performance and predict possible failures. Bringing AI into energy storage would increase the battery uptime, subsequently increasing ROI. Battery diagnostics and battery management are the major areas where AI can make a huge difference in terms of battery operations.
Transmission and Distribution
As energy demand increases in India, the distribution companies are supposed to bring in the quick response model, where the surplus energy being generated at one point is successfully diverted to a point where there is shortage of energy. With predictive analytics such calculations can be done beforehand. Integration of AI will reduce errors, improve predictability, and bring in balance to the automation of such processes.
Currently, power supply is largely dependent on central grids. As decentralisation of the grid happens, many small power plants will form the decentralised grid infrastructure and collectively respond to the energy demands.
AI would play a crucial role in the management of such systems. This would also reduce the losses during Transmission and Distribution (T&D) of energy. Deploying AI systems can also be crucial to monitor electricity theft, which is a major cause of T&D losses in India.
Energy infrastructure with integrated AI will be a complex system. It also generates apprehension, as AI would have the power to take decisions. The decisions an AI system will take need to be based on reasoning.
As machine learning is a complex process, it is very difficult to understand how and why certain decisions are being made (the concept is called explainable Artificial Intelligence – XAI). This is a new area of research and creates scope for researchers and AI enthusiastic entrepreneurs to explore more. AI decisions not being explainable is also a big impediment in autonomous vehicles. Once there is a breakthrough, power grids have potential to become completely autonomous with zero human intervention.
Smart Homes and Smart Buildings
Currently smart solutions or smart systems are retrofitted in buildings and homes in Indian cities. The smart building concept has a huge potential wherein the solutions could be installed in the construction phase itself, with reduced cost. Heating ventilation and air cooling (HVAC) is one area where AI plays a role for energy efficiency and smart operations. It can significantly reduce power bills for the end consumer.
Newer technologies are at the forefront of precise applications of energy management. As end consumers become aware of their consumption patterns, smart systems will enable conscious power consumption. This opens up the possibilities for AI to serve power distributors and end consumers and create an interactive model for energy management benefitting both.
The writer is Portfolio Manager at Social Alpha