Battery energy storage systems (BESS) have seen significant advancements over the past few years and with it, we have seen greater investor knowledge and confidence in the technology. It is widely understood that BESS is vital for enhancing grid resilience, levelling fluctuating power outputs, and addressing peak energy requirements and software development is crucial for maximising profitability.
Traditionally batteries have performed frequency response, providing grid stability with a millisecond response time. Here batteries use a classic control algorithm to adjust their output depending on the difference between the actual and desired grid frequency and provide ancillary services to the grid. Yet with continued technological improvements and cost reductions across the value chain, a more enhanced business model for BESS is Energy Arbitrage (Time Shifting). Energy Arbitrage allows for the purchasing of electricity during off-peak periods when it is cheaper and selling it back to the grid during peak times when it is more expensive, but there are several commercial challenges to overcome. Namely, faster asset degradation and higher capex volatility.
Despite the challenges, the prospects for BESS are bright. However, there is one area of BESS that is often overlooked: software and algorithms. Putting aside the challenges related to out-of-date grid infrastructure, as well as geopolitical and environmental questions surrounding supply chains, for BESS to have a future-proof profitable business model, tailored to individual markets, there needs to be significant focus and development concerning software.
David J.A Post and I went “inside the container” and discussed the challenges facing battery storage software. David J. Post is the President of the European Association for Storage of Energy and the Global Head of B2B Marketing and Sales at Enel X.
It is important to understand the difference between operations and maintenance software and energy bidding optimization software. Battery storage integrators are developing Predictive Maintenance software to help with charging-discharging to recommend better cycling patterns, minimise wear and tear, and reduce the degradation curve. In contrast, optimised bidding software relates to energy trading.
There is a natural challenge in assessing where software is at any point, given there are continual improvements. Generally speaking, software tools currently rely heavily on predictive models, meteorological predictions, and past data to forecast demand trends so that they can strategize the charge/release cycles. For markets with significant amounts of installed capacity, such as Australia, the UK, and parts of North America (e.g., California, and Ontario) there is already basic software that can optimise energy streams. Yet often players are still relying on an “Excel-type bidding approach.” Algorithms have the capacity for next-day pricing predictions but there is still a hesitancy to “let the machine take control” and permit automated energy price bidding, with limited manual oversight. Despite trust in automation being far higher than a few years ago and peak prediction algorithms becoming increasingly sophisticated, manual oversight, particularly on weather patterns, allows for lessons learnt to be implemented. Over time however we will see more and more automation in established markets.
As for the software readiness in other markets, such as Germany, Italy, and Spain, question marks remain.
There are several players in the BESS optimisation and trading space. For example, Habitat Energy was acquired by Quinbrook Infrastructure Partners and is expanding into the United States and Australia. Habitat focuses on optimising flexible generation and storage. Other players include Ascend Analytics, Enspired and DER.OS, which is an Enel X software platform. A trend among larger players is to acquire smaller software companies to then build out their in-house capabilities, as seen with Fluence.
Software will continually be developed to be more efficient and advanced, especially as improvements come to advanced machine learning. The improvement that needs to come is in the market itself and may require government intervention. New markets require new software, and despite leveraging off existing algorithms, it wouldn’t be fruitful to compare the Australian and Spanish markets for instance. Software development routinely runs over budget and having a few developers making their own (often in-house) software to have their own Intellectual Property (IP) is inefficient and expensive and relies on a significant (megawatt) critical mass to justify such investment. To support a faster scaling of BESS, more and more players are looking at various forms of collaboration to enter new markets to optimize future algorithm developments. One way new markets can be made more attractive and collaboration encouraged is through government incentives, be they economical or regulatory. For example, the decision to allow independent battery storage projects to be eligible for Inflation Reduction Act (IRA) funding enhanced the market conditions for storage in the United States.