Sensor Arrangement Fault Bridging With ML Replacement Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor arrangements face challenges in ensuring reliable operation due to sensor defects, leading to potential downtime and economic consequences, particularly in industrial processes.
Innovation Solution
A method using machine learning to generate replacement sensor data from functional sensors to temporarily replace defective sensors, ensuring continuous process operation by training a data model with historical sensor data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor data from all sensors are used for process control, then measurement precision is improved, but reliability deteriorates when sensor defects occur
Solution Approach 1:
The patent creates virtual copies of defective sensor data by training a machine learning model on historical sensor data. The trained model generates replacement sensor data that mimics the expected output of the defective sensor, allowing the system to continue operating with accurate process control information even when physical sensors fail.
Solution Approach 2:
The machine learning model acts as an intermediary between the defective sensor and the process control system. Instead of directly using defective sensor data or completely stopping the process, the model mediates by generating plausible replacement data that maintains process continuity while preserving measurement precision.
2Reliability
If redundant sensors are added to ensure continuous operation, then reliability is improved, but device complexity increases
Solution Approach 1:
The patent replaces the mechanical approach of adding redundant physical sensors with an information-processing approach using machine learning. Instead of installing additional hardware sensors to provide backup data, the system uses a trained model to computationally generate replacement sensor data, significantly reducing device complexity while maintaining reliability.
Solution Approach 2:
The patent changes the parameter space from physical sensor quantity to computational processing capability. By transforming the problem from a hardware redundancy solution to a software-based data generation approach, the system achieves the same reliability goal with fewer physical components and reduced system complexity.
3Object-affected harmful factors
If sensor defects are detected and process stopped for safety, then harmful factors are reduced, but productivity decreases
Solution Approach 1:
The patent performs preliminary action by training the machine learning model on historical sensor data before defects occur. This pre-trained model is ready to immediately generate replacement data when a sensor failure is detected, eliminating the need to stop the process for model training or data collection, thus maintaining productivity while ensuring safety.
Solution Approach 2:
The system performs self-service by automatically detecting sensor defects and generating replacement data without requiring external intervention or process shutdown. The machine learning model autonomously compensates for defective sensors, allowing the process to continue operating safely and maintain productivity simultaneously.
Data Source
AI summary
A method for operating a sensor arrangement and an apparatus for data processing, which is suitable for carrying out such a method are provided. In addition, the invention relates to a sensor arrangement which is suitable for being used in such a method and/or interacting with such an apparatus, as well as to a device comprising such a sensor arrangement and/or such an apparatus.