Bayesian Network for Disaggregated Energy Consumption from Billing Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional energy management tools require frequent and high-frequency data collection, necessitating large amounts of data and the installation of special sensors, making them inconvenient, impractical, and inefficient for determining disaggregated energy consumption.
Innovation Solution
A Bayesian network model is trained using aggregated energy consumption data at low frequency intervals, allowing for the inference of energy consumption values for various sources without additional equipment, by incorporating user inputs and external data, and applying maximum a posteriori estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional energy management tools use high-frequency data collection, then measurement precision of energy consumption is improved, but device complexity and ease of operation deteriorate due to requiring special sensors and large data gathering
Solution Approach 1:
The patent replaces physical sensors and hardware-based measurement systems with a software-based Bayesian network model that processes billing data. This substitution eliminates the need for complex sensor installations while maintaining measurement capability through probabilistic inference from aggregate data patterns.
Solution Approach 2:
The patent introduces a Bayesian network model as an intermediary between aggregate billing data and disaggregated energy consumption estimates. This intermediary processes low-frequency billing data and user inputs to infer consumption patterns, avoiding the need for direct high-frequency measurements while achieving similar analytical goals.
2Measurement precision
If conventional energy management tools install special sensors, then measurement precision is improved, but ease of operation and ease of manufacture worsen due to installation requirements
Solution Approach 1:
The patent replaces physical sensor installations with a software-based Bayesian network model that processes billing data. This substitution eliminates the need for complex sensor installations while maintaining measurement capability through probabilistic inference from aggregate data patterns.
Solution Approach 2:
The patent uses billing data as a copy or proxy for direct measurements. Instead of installing sensors to capture raw consumption data, the system processes copies of billing information that already exist, inferring detailed consumption patterns from these aggregate representations.
3Measurement precision
If conventional energy management tools gather large amounts of data, then measurement precision is improved, but loss of time and productivity worsen due to data gathering requirements
Solution Approach 1:
The patent leverages billing data that has already been collected and processed by utility companies before it reaches the energy management system. This preliminary data collection and aggregation is performed in advance, eliminating the need for the end-user system to gather raw data and reducing processing time significantly.
Solution Approach 2:
The patent extracts useful information from pre-existing billing data that has already been aggregated and processed. By taking out and analyzing specific patterns from this pre-processed data using Bayesian inference, the system achieves energy consumption insights without investing time in raw data collection and preprocessing.
Data Source
AI summary
Various embodiments of the present disclosure can include systems, methods, and non-transitory computer readable media configured to train a Bayesian network model based on a given set of data. Information associated with a user can be received. The information can include aggregated energy consumption data at one or more low frequency time intervals. At least a portion of the information can be inputted into the Bayesian network model. A plurality of energy consumption values for a plurality of energy consumption sources associated with the user can be inferred based on inputting the at least the portion of the information into the Bayesian network model.