Technology
Our thesis and development direction for an intelligence layer for distributed energy.
Everything on this page describes what Elenotic is researching and building toward.
The intelligence loop
We think of the system as a continuous loop. The final step feeds back into the first.
Observe
Collect information from solar systems, batteries, the grid, appliances, EVs, meters, weather and tariffs.
Understand
Maintain a current representation of the energy system: generation, consumption, battery state, grid status, appliance state and constraints.
Predict
Use machine learning to estimate solar generation, household demand, future energy availability and battery behavior.
Optimize
Decide how available energy should be allocated, weighing cost, solar utilization, battery health, backup needs, appliance constraints, preferences and comfort.
Act
Eventually coordinate controllable assets according to those decisions.
Learn
Use real-world outcomes and feedback to improve predictions and decisions over time.
Learn returns to Observe, and the loop repeats.
Neural networks help us understand what is likely to happen
We intend to use machine learning mainly for prediction and pattern recognition. A separate layer then decides what to do.
Solar forecasting
Historical generation, weather, time and system information go in; expected future solar generation comes out.
Demand forecasting
Historical household consumption, time and behavior patterns go in; expected future demand comes out.
Battery modeling
Historical battery behavior is used to build a better understanding of battery state and behavior.
Prediction is not enough
If a system predicts low solar production tomorrow, it still has to decide what to do today. Optimization balances competing objectives:
Prediction asks
What is likely to happen?
Optimization asks
What should we do about it?
Reinforcement learning: a research direction
Reinforcement learning (RL) lets a system learn sequential decision-making by interacting with an environment. We are exploring it as part of our long-term intelligence research.
Objectives we are interested in include lower energy cost, better solar utilization, better backup reliability, less unnecessary battery cycling and maintained comfort.
Hardware-agnostic by design
Elenotic is not meant to be another single-vendor solar ecosystem. The long-term vision is one intelligence layer that works across heterogeneous hardware and energy assets.
This represents our generic model
How we think about differentiation
Companies already work on solar optimization, home energy management, batteries and distributed-energy management. We do not claim to have invented AI energy management. Our thesis rests on four ideas.
Hardware-agnostic intelligence
One layer that could coordinate assets from different vendors.
System-level optimization
Coordinated decisions across generation, storage, consumption, backup and flexible loads, beyond monitoring.
Learning systems
Forecasting, optimization, machine learning and reinforcement learning together, instead of static rules alone.
Distributed-energy focus
Built around distributed energy from the start, not added later.
Elenotic is an early-stage energy technology company exploring the intelligence layer for distributed energy. We are researching and developing this technology; we do not yet have a commercial product, customers or deployments.