12th November 2019
The Emerging Need For Artificial Intelligence in Energy Storage
A tectonic shift for the Energy Industry is on the horizon.
We’re amidst one of the most exciting times as the grid storage market is finally taking off, giving rise to radical new opportunities with artificial intelligence in the energy storage sector for utilities, customers, policymakers and the economy.
Moreover, the potential of energy storage in shaping and smoothing variable generation and supporting changing demand will enable catalytic advancement in the energy sector, especially with renewable sources.
The energy storage market is expected to grow by US$97.8 Billion worldwide. Wood Mackenzie and Energy Storage Association analysts have predicted that in the next five years the total MWh deployed will grow by approximately 14 times!
The driving forces behind this shift are:
- The need to curb the volatility of wind and solar energy production
- Energy storage is no longer restricted to large projects. There is an evident growth in small-scale, residential and commercial solar projects in the past six years
- The development of energy storage technology coupled with a remarkable reduction in the cost of lithium-ion batteries, that are viable for projects of every size
The Need for Artificial Intelligence in Energy Storage
The sector is clearly steering in the direction, however, the development of energy storage technology still lags far behind. Today’s energy storage systems lack the infrastructure to understand and utilize the energy at high efficiency. The industry needs innovation and breakthrough in capacity planning, long-lifespan (high battery uptime), better ROI, etc. for energy storage.
There couldn’t be a better time for companies to act fast, optimize and differentiate themselves in the space or lose favor. Technology is progressing at an exponential rate. There have been monumental breakthroughs in the field of machine learning and deep learning – from chatbots to generative modelling. These concepts have allowed machines to process and analyze tonnes of information in a very sophisticated manner.
Applying Artificial Intelligence (AI) and machine learning (ML) algorithms can immensely advance the energy storage sector. AI-enabled energy storage will help collect and analyze the data, and by using simulations it can provide insights to optimize power utilization and predict possible failures.
How to Make Standalone Systems Intelligent
Applying Artificial Intelligence to standalone systems can make them smarter and more accessible. Most importantly, AI can enable the harvesting of renewable energy sources by improving the efficiency of power distribution, which is dependant on the production-consumption cycle of the end-consumer.
Adding battery-intelligent storage to a renewable energy installation will almost always increase the economic value. AI opens a lot of possibilities that help correctly optimize the system and maximize returns from a renewable energy storage system for the customer. It can help perform predictive analytics, machine learning, big data and grid-edge computing which is required to achieve these returns.
The intelligent storage will collect data, which can constantly be logged and analyzed on loads, power generation, weather, nearby grid congestion, etc. AI-enabled storage can drive real-time adaptive storage supply that generates increasing value for the customer and the grid.
As energy storage systems become smarter, it gets easier to harness renewable energy. Due to the intermittent nature of the sources, AI helps solve the issue of harnessing the irregular alternation of production phases of renewable energy.
The complexity of the situation is due to the complex dynamic nature of the production-consumption system. This makes it difficult to monitor and analyze the data in real-time.
The key to unlocking the value is having proper control of energy storage through intelligent platforms. Pairing renewable energy with AI-driven storage could be the change in paradigm towards a highly sustainable future.
What Do Battery Intelligence Platforms Offer?
Imagine having to swap brand new batteries from energy storage that may have been used a couple of times due to the rare blackouts. A telecom company in India had to swap their batteries across the country because of an uneven decline in the battery capacity. Factors like load, blackout duration, and even weather seem to affect the degradation but it is generally unclear as to which of these contribute to the degradation and in what measure.
Battery intelligence tools help diagnose the problem with to-the-minute real-time information. With the help of real-time geo-tagging and data visualization tools, every single battery can be checked, remotely, for battery health and reasons for downtime, if any.
For example, the image above displays the State of Charge (SoC) of a grid. A battery intelligence platform will conduct “Standard Cycling tests” for BMS benchmarking, assess and implement innovative methods to improve the SoC accuracy of the grid under varying situations.
Maximizing Uptime & Zero Downtime
he platforms use advanced machine learning (ML) algorithms to learn the battery energy consumption patterns from the grid, use digital twins to simulate a model, and help improve uptime by providing over-the-air updates.
With the help of advanced diagnostics, these platforms improve the battery performance by predicting the state of health (SoH) of the battery to avoid any possible downtime.
Predictive Maintenance (using customizable alerts)
Once the batteries have populated enough data, these platforms help in predicting (way ahead of time) possible failures, or the need to change batteries.
Alerts can be set for each and every battery asset depending on the parameter thresholds set by the energy storage operator. Some of the more advanced platforms let you tag multiple assets and multiple tags for every battery asset. This enables the operator to view only those batteries that fall under a specific tag – geography, operating temperatures, battery types, etc.
Planned Replacements – Residual Life Estimation
The predictive algorithms don’t have just the aforementioned capabilities, once it learns how a specific battery operates, it can also accurately estimate the residual life, identify the decline in the health of the battery and help you plan the replacements, well before its time. For a large scale deployment, it means to select, plan, and procure the batteries for locations around the world, which can be problematic if the requirement is urgent.
Learning so much about the battery can’t just sit inside a memory chip, there needs to be a switch that can action those recommendations and have a fruitful outcome for those who have put in years of hard work behind Machine Learning (ML) & Artificial Intelligence (AI). Only a handful of battery intelligence softwares have the capability to alter thresholds and limit the use of the battery to increase the life of the battery by up to 40%.
In conclusion, the need for AI in Energy Storage is now, so much more than ever. Integrating it with renewable energy will influence the future of the planet. The confluence of AI and energy storage is an effective starting point. However, the technology as of today is nascent as compared to the potential impact it can cause. This synergy can truly change the world and open up countless opportunities while improving sustainability.
Edison Analytics: Battery Intelligence & Analytics Platform2nd December 2020/