Predictive Model for Energy Consumption Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current domestic energy consumption metering provides only coarse information, insufficient for detailed analysis or identifying opportunities to reduce energy consumption, as it lacks the accuracy needed for appreciable reductions in energy usage.
Innovation Solution
A computer-implemented method generates predicted energy consumption data for a first energy consumer based on energy consumption data from a plurality of other consumers, using a predictive model that defines relationships between predictor attributes and energy consumption data, allowing for detailed analysis and identification of energy-saving opportunities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current domestic energy consumption metering is used, then billing purposes are fulfilled, but detailed analysis and identification of energy-saving opportunities are not enabled
Solution Approach 1:
The patent introduces a predictive model as an intermediary between coarse metering data and detailed energy analysis. The model uses training samples from multiple consumers to generate detailed predicted energy consumption data, enabling fine-grained analysis without requiring complex physical metering infrastructure at each consumer premises.
Solution Approach 2:
The patent creates a virtual copy of detailed energy consumption patterns by training a predictive model on data from multiple consumers. This model copy can then generate detailed predictions for individual consumers based on their specific attributes, providing detailed analysis capability without direct complex measurement at each location.
2Measurement precision
If predictive modeling with multiple consumer data is used, then accurate energy consumption predictions are achieved, but data processing complexity increases
Solution Approach 1:
The patent performs preliminary actions by collecting and processing data from multiple consumers to create a trained predictive model in advance. This pre-trained model can then efficiently generate predictions for individual consumers without requiring complex real-time processing, as the heavy computational work has already been done during the training phase.
Solution Approach 2:
The predictive model serves multiple functions: it can predict total energy consumption, break down consumption by appliance categories, and provide detailed temporal patterns. This multi-functionality is achieved through a single unified modeling framework that processes training data from multiple consumers to create a versatile prediction system.
Data Source
AI summary
A computer-implemented method of generating predicted energy consumption data for a first energy consumer based on energy consumption data of a plurality of further energy consumers is disclosed. Prediction is based on a set of training samples, each training sample comprising predictor attribute data and energy consumption data for a respective one of the further energy consumers, as well as one or more predictor attribute values for the first energy consumer. A predictive model is generated based on the training samples, the predictive model defining a relationship between values of predictor attributes and energy consumption data of the training samples. Predicted energy consumption data for the first energy consumer is then determined using the predictive model and the predictor attribute values. A method of clustering energy consumption profiles is also disclosed.