In Part 1 of this series, we discussed how the complexities of today’s energy grid are creating unprecedented challenges for transmission systems operators (TSOs), distribution system operators (DSOs), equipment providers, energy aggregators, and the communities they serve.
These challenges manifest in different ways, such as price volatility and grid unreliability. But at its core, the underlying problem stems from the difficulties utilities face when incorporating distributed, intermittent green sources in an aging grid that was originally designed for centralized power generation. Against this backdrop, it is becoming increasingly difficult to accurately predict supply, demand, and rates to make better-informed decisions about power distribution and storage. As a result of this guesswork, utility operators often find themselves generating either:
- Too much power supply – leading to network congestion and forfeit profits
- Not enough electricity – resulting in frequent brownouts and rolling blackouts
However, it is now possible to remove humans from the energy loop, allowing for greater accuracy and reliability:
- Veritone’s energy forecasting software (Forecaster) works in concert with utility management systems to correctly sense and predict future supply and demand – even as both continue to evolve in real-time.
- Veritone’s energy optimization software (Optimizer) combines these forecasts with device parameter data to output a highly-optimized mix across all distributed energy resources (DERs).
How Our Energy Forecasting and Optimization Software Works
Built on Veritone’s artificial intelligence (AI) platform (aiWARE), Forecaster and Optimizer leverage predictive models, rules, and adaptive machine learning to determine the ideal pricing and energy supply mix to satisfy grid demand in real-time. More specifically, these 2 powerful tools actively monitor current local weather conditions, utility rates, production levels, and device-specific parameters to accurately forecast where demand and prices are heading. This allows you to predict and optimize your facility’s own energy mix for cost-effective and reliable power delivery. And the forecasting software does this with real-time prediction models – both at the ecosystem-level all the way down to the device-level – resulting in extremely localized and accurate energy optimizations.
This is in sharp contrast to most predictive energy solutions that either use:
- Historical data to predict future energy demand
- Centralized, neural network-based algorithms that are CPU-intensive, expensive to deploy, and nearly impossible to scale
Veritone’s energy demand forecasting software “predicts the future using the present” by combining current data from multiple sources, including weather sensors, distributed grid devices, and load demand systems. This real-time information allows utilities to accurately and dynamically meet grid demand – seconds, days, and weeks into the future.
Our energy forecasting and optimization software does this with patented, Hamiltonian-based models that can be deployed either centrally or decentralized across a network of DERs. This approach enables distributed, autonomous edge device decisions using inexpensive processors that can be cost-effectively deployed and scaled on-premise, via Internet of Things (IoT) devices, or in the cloud.
Equally important, Forecaster and Optimizer both benefit from adaptive learning, which helps improve accuracy as incoming weather, load demand, and pricing data evolve over time. In fact, our patented Cooperative Distributed Inference (CDI) technology analyzes millions of independent variables in real-time as it learns to automatically adjust course based on whatever user-defined rules you’ve set for your facility.
Using Energy Forecasting and Optimization to Eliminate Uncertainty
Managing the energy mix has always been a complex task – prone to error, redundancy, and waste. But with growing public pressure to green the grid, there are now far more moving parts to manage. More problematic still is the speed with which renewables like solar and wind are coming online. While beneficial for the planet as a whole, managing distributed energy sources in a centralized grid is now beyond the scope of human capacity. And this gap will only widen as climate change, price volatility, and decarbonization continue to introduce ever-greater complexity and uncertainty.
But with AI powering your decisions, you can generate reliable and cost-effective electricity as you build accurate energy forecasts and deploy real-time optimizations – in the short-, medium-, and long-term. However, simply having this information is only so useful. You still need a way to quickly implement this optimized energy mix at the device level to ensure grid reliability and mitigate unnecessary costs.
Fortunately, Veritone’s AI energy solutions also come with additional in-built tools that provide real-time device control and synchronization – both of which we’ll discuss in the next installment.
In the meantime, contact us to learn how Veritone’s energy forecasting and optimization software can help you cut costs and improve grid reliability.