EV Charging Session Detection From Household Net Profiles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Utilities face challenges in identifying and estimating the impact of residential electric vehicle (EV) charging on distribution grids due to the lack of dedicated meters for residential chargers, relying instead on power net metering that measures net load/production.
Innovation Solution
A method and system that receive a net profile for a whole household, estimate the EV charger size, identify candidate windows for EV charging, and classify these windows using a machine learning model to detect EV charging sessions without requiring separate EV charger metering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If power net metering is used to measure net load/production, then installation complexity is reduced and cost is lowered, but measurement precision for EV charging detection deteriorates
Solution Approach 1:
The patent segments the aggregate net load profile into individual appliance load profiles using machine learning models. By dividing the overall power consumption into separable components (HVAC, water heater, EV charger, etc.), the system can detect EV charging sessions from net metering data without requiring dedicated EV meters, thus resolving the contradiction between simple installation and precise measurement.
Solution Approach 2:
The patent introduces machine learning models as an intermediary between the net metering data and EV charging detection. These models act as a mediator that processes the aggregate net load profile and extracts EV charging information, enabling precise EV detection while maintaining the simplicity of net metering infrastructure.
2Measurement precision
If dedicated meters for residential chargers are installed, then measurement precision for EV charging is improved, but device complexity and installation cost increase
Solution Approach 1:
The patent creates a virtual copy of the EV charging load profile by training machine learning models on net metering data. Instead of installing physical dedicated meters, the system generates a digital replica of EV charging patterns through algorithmic processing, achieving the measurement precision of dedicated meters without their associated complexity and cost.
Solution Approach 2:
The patent replaces the mechanical/physical metering system (dedicated EV chargers meters) with an information-processing system (machine learning models). This substitution eliminates the need for additional physical measurement devices while achieving equivalent or superior detection capabilities through computational analysis of net metering data.
3Measurement precision
If machine learning models are used to classify candidate windows, then EV charging detection accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent performs preliminary processing by generating candidate windows that are likely to contain EV charging sessions before applying the machine learning classification model. This pre-filtering step reduces the number of data points requiring complex computational analysis, thereby improving detection accuracy while managing computational complexity through staged processing.
Solution Approach 2:
The patent applies different processing strategies to different portions of the data: simple candidate window generation for initial filtering, and sophisticated machine learning classification only for promising candidates. This localized application of computational complexity ensures high detection accuracy where needed while minimizing overall processing burden.
Data Source
AI summary
For detecting electric vehicle charging sessions in a whole household net profile, a first step estimates an electric vehicle charger size from the net profile. A second step identifies candidate windows associated with the charging of an electric vehicle based on the estimated charger size. In a third step, a machine learning model classifies the candidate windows into charging windows and unrelated windows, wherein each charging window indicates an electric vehicle charging session. The method needs only net-metering as input. No separate EV charger metering is needed for the detection. In an embodiment of the method, the machine learning model is picked from a set of machine learning models, depending on the estimated charger size. Each model in the set of machine learning models has been trained for a different charger size. The machine learning model receives as input suitable features that have been computed for the candidate windows.