Parallel Sensor Calibration for Seamless Measurement Replacement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The process of replacing a sensor in a measurement system requires time-consuming calibration and may result in output value discrepancies due to instrumental differences, necessitating model modifications and potentially suspending the system for maintenance.
Innovation Solution
A measurement system utilizing two sensors, an existing and a new sensor, allows for simultaneous measurement and conversion of new sensor data to match the existing sensor's output, enabling calibration data generation while maintaining system operation, thus facilitating seamless sensor replacement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of repair
If a new sensor is installed to replace an old sensor, then maintenance is performed, but calibration time and effort increase
Solution Approach 1:
The system performs preliminary actions by having the new sensor operate in parallel with the old sensor before replacement. Calibration data is collected during this parallel operation period, so when the old sensor is finally replaced, the calibration is already complete or near-complete, eliminating the need for time-consuming post-replacement calibration
Solution Approach 2:
The old sensor acts as an intermediary during the transition period. It continues to provide reference measurements that serve as a mediator for calibrating the new sensor, allowing the new sensor to be calibrated against the old sensor's known characteristics without requiring external calibration equipment or procedures
2Measurement precision
If sensor characteristics differ between old and new sensors, then measurement accuracy may be maintained, but machine learning model modification is required
Solution Approach 1:
The system implements feedback by continuously comparing measurements from the old and new sensors during the parallel operation period. This feedback loop allows the system to automatically detect discrepancies between sensors and trigger recalibration or model adjustment procedures, maintaining measurement accuracy without manual intervention
Solution Approach 2:
The system automatically adjusts calibration parameters when sensor characteristics differ. By detecting changes in sensor output patterns and automatically modifying calibration parameters or scaling factors, the system maintains measurement accuracy without requiring complex machine learning model modifications
3Measurement precision
If the measurement system is suspended to modify the machine learning model, then model accuracy is improved, but system downtime increases
Solution Approach 1:
The system performs model adjustments and calibrations in advance during parallel sensor operation. By completing all necessary model modifications before the old sensor is actually replaced, the system ensures model accuracy is optimized while minimizing any suspension time to just the brief moment of physical sensor swap
Solution Approach 2:
The measurement system maintains continuous operation throughout the sensor replacement process. The old sensor continues measuring while the new sensor is being calibrated and integrated, ensuring uninterrupted measurement capability and eliminating system downtime associated with model modification
Data Source
AI summary
An easy-to-maintain measurement system including a first sensor that measures a target and outputs a predetermined physical quantity, a second sensor that measures the target and outputs the same type as the predetermined physical quantity, and a processor configured to (1) acquire a first value from the first sensor and a second value from the second sensor, (2) convert the first value to the value of the physical quantity by a first calibration data, (3) calculate second calibration data for the second sensor from the second value and the predetermined physical quantity obtained in step (2), (4) convert the second value to the physical quantity by the second calibration data, and (5) detect the state of the measurement target by an estimation model and the physical quantity obtained in step (4).