Online Meter Calibration via Resistance Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting electrical non-conformities in low voltage networks, such as those described in US 2015/0241488, are limited in precision as they do not account for the calibration differences between meters and do not effectively detect load manipulations that alter the apparent connection resistance, leading to potential deception in consumption measurements.
Innovation Solution
A method for correcting consumption measurements by calculating voltage calibration coefficients using a cohort of samples from meters sharing a network, estimating network, connection, and total resistance values, and detecting anomalies in resistance values over time to refine calibration and identify non-conformities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If linear regression is used to link voltage variations to current variations at a single meter, then impedance correlation can be deduced, but the method mixes network resistance and meter connection resistance leading to reduced detection precision
Solution Approach 1:
The patent segments the total resistance measurement into two distinct components: network resistance (Zi) and connection resistance (ri). By analyzing voltage variations of a reference meter versus current variations of the target meter, the method separately determines network resistance from the reference meter's data, then subtracts it from the total resistance to isolate connection resistance. This segmentation resolves the contradiction by enabling precise anomaly detection through differentiated resistance analysis.
Solution Approach 2:
The patent introduces a reference meter as an intermediary element to measure network resistance independently. This reference meter serves as a mediator between the target meter and the network, allowing the system to distinguish between network-induced voltage variations and connection-induced variations. The reference meter's measurements act as a benchmark to correct the target meter's readings, thereby improving anomaly detection precision without sacrificing detection capability.
2Measurement precision
If voltage calibration coefficients are calculated using all available samples, then calibration accuracy improves, but load manipulations can deceive the measurement by lowering apparent connection resistance
Solution Approach 1:
The patent applies local quality by selecting specific cohorts of samples based on load conditions rather than using all available data uniformly. The method identifies and analyzes samples within specific load ranges where the relationship between voltage and current variations is most reliable for detecting connection issues. This selective sampling approach ensures that calibration is performed on data that reflects actual connection resistance rather than being skewed by load manipulation effects.
Solution Approach 2:
The patent performs preliminary analysis to identify and exclude samples that may be affected by load manipulation before calculating calibration coefficients. By pre-screening the data cohort to ensure it represents normal operating conditions, the method prevents deceptive load manipulations from corrupting the calibration process. This preliminary action protects the reliability of the calibration while maintaining its precision.
3Measurement precision
If differential voltage comparison between meters is used to detect anomalies, then smaller magnitude anomalies can be revealed, but calibration discrepancies between meters must first be corrected
Solution Approach 1:
The patent changes the parameter being measured from absolute voltage to corrected voltage that accounts for calibration discrepancies. By introducing calibration coefficients derived from the relationship between voltage variations and current variations, the method transforms raw voltage measurements into corrected values that eliminate systematic errors. This parameter transformation enables sensitive differential comparison between meters while the calibration correction complexity is managed through automated calculation based on observed electrical behavior.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
The disclosed method corrects consumption measurements supplied by meters presumed to be connected on the same network. The measurements taken by the meters at time intervals are collected in the form of samples used to determine the quantities of electricity in relation to the currents and voltages relating to the meters. A selection is performed of a cohort of samples selected from the samples considered to be valid, which correspond to a range of load or a range of load variation ratio transmitted by the network based on the electrical variables. Correction functions of the consumption measurements are defined on the basis of the cohort of samples selected, and the consumption measurements are adapted according to the correction functions. An anomaly corresponding to an electrical non-conformity of a meter can be detected by means of the method.