Filtering Electrical Consumption Curves for Device Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for monitoring and allocating electrical energy consumption in residential environments are intrusive, costly, and inefficient, typically allowing only around 80% of peak consumption to be accounted for, with a high need for manual learning phases and additional equipment installation.
Innovation Solution
A method that filters electrical energy consumption curves to allocate consumption to device categories without intrusion, using a hierarchical classification of devices based on consumption cycles, automatic learning, and specific power and time criteria to reconstruct device operation models, allowing for over 85% accurate allocation without prior knowledge of devices or manual learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If each electrical appliance is equipped with a wattmeter to record electrical power consumed, then the electrical consumption can be determined by device or by group of devices, but the cost is high both because of the additional equipment used and because of the need to install and configure the devices
Solution Approach 1:
The patent segments the aggregate load curve into individual device consumption patterns by analyzing transient signals and steady-state characteristics. Instead of measuring each device separately with wattmeters, the system divides the total consumption signal into component signals representing different devices based on their unique electrical signatures
Solution Approach 2:
The patent extracts individual device information from the aggregate load curve by identifying and separating transient signals characteristic of specific devices. The system takes out device-level consumption data from the combined signal without requiring physical separation or additional measurement equipment at each device
2Measurement precision
If a manual learning phase is carried out to establish a register of devices on the site, then the recognition of devices on the load curves can be improved, but this causes discomfort for the user and does not allow a change in consumption habits
Solution Approach 1:
The patent implements self-service by enabling the system to automatically learn and register devices without user intervention. The device learning and identification process occurs autonomously through analysis of electrical signatures and transient signals, eliminating the need for users to manually create device registers or provide information about their appliances
Solution Approach 2:
The patent performs preliminary automatic device learning and registration before consumption monitoring begins. The system proactively identifies and stores device characteristics in advance, so that when devices are later recognized on load curves, the matching can occur immediately without requiring user-initiated learning phases
3Quantity of substance
If known methods are used to determine the consumption of devices, then some information about device consumption can be obtained, but at best cases only around 80% of peak electrical consumption can be accounted for, and often much less than 80% on average
Solution Approach 1:
The patent uses feedback by continuously monitoring the load curve and comparing actual consumption patterns against identified device signatures. The system refines its device recognition and consumption allocation based on ongoing analysis, improving accuracy over time and achieving better than 80% accounting of peak consumption
Solution Approach 2:
The patent applies dynamics by adapting the device identification and consumption allocation process to changing conditions. The system dynamically updates device registers, adjusts recognition thresholds, and modifies consumption attribution based on evolving patterns in the electrical signals, allowing it to maintain high accuracy despite variations in device operation and addition of new devices
Data Source
Figure 1~2
Figure 3A~3D
Figure 4
AI summary
The invention relates to a method for analysing the electrical consumption of a plurality of electrical appliances operating on a consumption site, by filtering a demand curve representing the electrical consumption of said appliances according to time. Said method is characterised in that it comprises the following steps: before the filtering per se of the demand curve, the demand curve is recorded and digitalised in such a way as to obtain a demand curve digitalised by periods of time; a set of categories of appliances is defined, each category being defined by similar cycles of power variation according to the time; an algorithm is defined for each category of appliances, for filtering the demand curve for said category of appliances, said algorithm being able to extract the power variation cycles from the digitalised demand curve and to allocate the electrical consumption to said category of appliances; then during the filtering per se of the digitalised demand curve, the filtering algorithms for each category of appliance are used successively to identify and regroup the power variation cycles consumed by said electrical appliances, from the digitalised demand curve.