How an Energy Management System (EMS) Makes Decisions for an Energy Storage System
by Lindsey Paulk on Jun 21, 2023
energy management system
energy storage
energy storage management
ESS software
An energy management system (EMS) plays a crucial role in optimizing the performance and utilization of an energy storage system (ESS) and determining the most effective dispatch strategy for the system. Essentially, it makes the decisions for the system. A storage system controlled by a full-featured EMS is functionally synonymous with a self-driving car. High-level EMS software that works hand-in-hand with the battery management system (BMS) is essential to value creation in ESS applications.
Energy Toolbase’sAcumen EMS™ controls software, for example, uses artificial intelligence (AI) to predict and precisely discharge energy storage systems operating in the field. Acumen utilizes field operational and perfect foresight algorithms to constantly make swift decisions – a requirement when dispatching an ESS to extract the total economic value. These algorithms are what set up a system for success.
Field operational Algorithms
An EMS uses AI for event-based dispatch. Operational algorithms make decisions in real-time according to time-series forecasts that predict the value of site power demand and on-site renewable energy generation. The net power demand is calculated by determining the difference between these values. Acumen EMS forecasts site demand and renewable energy generation every 15 minutes for the upcoming 24 hours. Generating these new forecasts consistently enables the control system to enhance the battery asset’s charge and discharge schedule.
Additionally, control software leverages machine learning models to pinpoint time and weather-based patterns. Time-related factors include time of the day, day of the week, day of the year, and holidays. Weather-related factors, like ambient temperature, also play a role. The control software should also consider the battery’s state of charge (SOC) and other technical aspects.
Perfect Foresight Algorithms
Perfect foresight algorithms mathematically determine the best discharge profile using comprehensive information about the future, unlike operational algorithms that dispatch an EMS without any knowledge of the future. Energy Toolbase has created and refined cutting-edge perfect foresight algorithms on theETB Developer sales and modeling platform, enabling users to simulate practical ESS savings estimates by considering battery efficiency degradation.
“The intention with perfect foresight algorithms is to determine the maximum potential savings,” explains Prudhvi Tella, Energy Toolbase’s Engineering Manager in charge of Acumen EMS control strategies. “Predictions based on a perfect forecast essentially provide the maximum upper bound for the theoretical savings. Without the insights offered by perfect foresight algorithms, we cannot improve the accuracy of time-series forecasts over time or produce realistic simulations for sales proposals.”
Perfect foresight algorithms offer an ideal performance reference point for continued process development, as well as verification. To illustrate, Energy Toolbase conducted 3-, 6-, and 12-month performance reviews on recently deployed field assets based on perfect forecasts to confirm and substantiate customer savings. Conducting these reviews gives us the opportunity to learn from previous performance and strengthen performance moving forward.
You can view the performance reviews of Acumen EMS and review case studies in our full Monetize your Energy Storage Asset: Demystifying How Acumen EMS™ Reliably Dispatches to Achieve Optimal Financial ReturnsWhitepaper.
Customizing and Training the Model
AI and machine learning models predict future events by analyzing history. In the same manner, Acumen EMS produces forecasts according to the historical data of each site in order to properly train the model to operate in the field. Customer-provided historical usage data and location-based third-party weather information are key parameters used when beginning the training process.
“Our automated model training pipeline uses these raw input streams to encode features that Acumen EMS uses to create a forecast based on historical patterns,” explains Tella. “Ideally, we work with a year of historical data. However, Acumen’s models performs quite well with as little as 8 days of site data.”
There are specific scenarios where our team utilizes techniques to attend to the ‘cold start’ predicament, where historical data is not provided for new buildings, or when interval data is unavailable. In these situations, we employ our vast database of building load profiles to capitalize on forecast performance and accuracy whenever data is incomplete or unavailable. Energy Toolbase has spent years compiling these load profiles to fit almost any type of scenario.
Anyone developer working on a project with no historical data canschedule a call with our team to talk more. You can also set up an energy storage consultationhere.
