Automatic machine learning based prediction of baseline energy consumption
Inventors
Chandra Mohan, Ananda Vel Murugan • Balakrishna, Banuprakash • Chawla, Chhavi • Sandeep, Deepika • Pal, Rohil • Ghosh, Surjayan • Konala, Swaroop Reddy
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12755064-B2
Publication Date
2026-10-06
Expiration Date
Abstract
The present solution, approach or method, including an end-to-end automated data pipeline for data ingestion, storage, analysis, deployment, and a machine learning model maintenance. The present solution, approach or method, which computes a baseline using machine learning methods, may help in the following ways. Accurate real time estimation may help evaluate the deviation in the actual energy consumption, effectively identifying underlying root causes for an increase in actual consumption, as compared to the estimated energy. Triangulating the time of day and place of high energy consumption results in quicker resolution. Accurately quantifying energy savings may be helpful. Forecasting energy consumption in the future, may enable planning for future energy needs. Energy saving calculations may be done by comparing actual consumption versus baseline predicted consumption based for a specific baseline period. This solution may offer a configurable machine learning model, which takes on energy consumption patterns.
Core Innovation
The claimed method evaluates energy consumption by generating a building energy consumption baseline using a machine learning model, measuring building energy consumption, and comparing the measurement with the baseline. When the measurement exceeds the baseline, an alert is issued, and the model is retrained with updates, triggered events, and building changes. Predicted consumption across forecast horizons from one day to sixty days aids planning for energy demand.
The system utilizes meter, weather, occupancy, temporal, and building-ontology data as inputs for energy-consumption prediction, including for sites lacking historical meter data. Data extraction, validation, preparation, training, evaluation, model registration, and real-time prediction support the machine-learning evaluation alongside graph and delta table representations. Comparing measured consumption with the baseline generates alerts and actionable equipment insights.
Claims Coverage
The relevant claims comprise independent claim 1 and dependent claims 2–4, presenting six inventive features.
Generating an energy consumption baseline
Claim 1 generates an energy consumption baseline of a building using a machine learning model, with estimated predicted consumption in multiple forecast horizons varying from one day to sixty days.
Measuring and comparing energy consumption
Claim 1 measures energy consumption of the building and compares a measurement of the energy consumption with the energy consumption baseline.
Issuing an alert when consumption exceeds the baseline
Claim 1 provides that an alert is issued if the measurement of energy consumption exceeds the energy consumption baseline, and claim 2 specifies that the alert is issued automatically.
Retraining the machine learning model
Claim 1 retrains the machine learning model with updates, triggered events, and building changes to the building.
Highlighting actionable insights related to faulty equipment
Claim 3 describes highlighting actionable insights on an indicator to help the building management team address faulty equipment causing consumption to exceed the baseline.
Using occupancy information as a model factor
Claim 4 uses real-time or live occupancy information as an external model factor and correlates occupancy with energy consumption.
The claims cover generating a machine learning baseline across multiple forecast horizons, measuring and comparing energy consumption, conditional alerting, and retraining the model. The dependent claims add automatic alerting, indicator-based actionable insights concerning faulty equipment, and using occupancy information as a model factor.
Stated Advantages
Predicted consumption across forecast horizons from one day to sixty days aids in planning for energy demand in the building.
Indicator-based actionable insights help the building management team address faulty equipment causing consumption to exceed the baseline.
Documented Applications
Evaluating energy consumption by a building by comparing measured consumption with a machine-learning energy consumption baseline.
Planning for energy demand in a building using predicted consumption across multiple forecast horizons.
Helping a building management team address faulty equipment when energy consumption exceeds the baseline.
Interested in licensing this patent?