Radiation Detector State Inference Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional radiation monitoring devices face challenges in accurately determining soundness due to the complexity of determining a large number of detectors outputting complicated signals, leading to inefficient and time-consuming maintenance processes.
Innovation Solution
An inference device utilizing machine learning to infer the operation state of radiation detection devices through a trained model based on feature quantities of state signals, enabling accurate and simplified determination of device soundness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional spectrum measurement methods are used to confirm soundness of radiation monitoring devices, then measurement precision is maintained, but device complexity increases and determination time increases
Solution Approach 1:
The patent replaces conventional mechanical/signal-processing-based spectrum measurement methods with machine learning inference. The inference device uses a trained model that processes detector signals through neural networks to directly determine soundness, eliminating the need for complex manual spectrum analysis and reducing determination complexity while maintaining accuracy.
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between raw detector signals and soundness determination. The trained model acts as a mediator that processes complicated signals from multiple detectors and outputs soundness status, simplifying the determination process while maintaining measurement precision.
2Measurement precision
If conventional spectrum measurement methods are used to confirm soundness of radiation monitoring devices, then measurement precision is maintained, but determination time increases
Solution Approach 1:
The patent applies preliminary action by pre-training the machine learning model using historical data and conventional measurement results. The trained model captures patterns and relationships from past measurements, enabling rapid inference without requiring time-consuming real-time spectrum analysis during actual soundness determination.
Solution Approach 2:
The patent replaces time-consuming conventional spectrum measurement processes with fast machine learning inference. The neural network model processes signals much faster than traditional methods, significantly reducing determination time while maintaining the same measurement precision through learned patterns from training data.
3Reliability
If the number of detectors is increased to improve monitoring capability, then monitoring performance is improved, but determination complexity increases
Solution Approach 1:
The patent applies universality by designing a single machine learning model that can process signals from multiple types of detectors with different characteristics. The model is trained to handle varied detector outputs and provides unified soundness determination for the entire radiation monitoring device, eliminating the need for separate processing for each detector type.
Solution Approach 2:
The patent replaces complex manual signal processing for multiple detectors with machine learning inference. The trained neural network automatically processes and correlates signals from numerous detectors, reducing determination complexity while maintaining improved monitoring performance through comprehensive multi-detector analysis.
Data Source
AI summary
Provided is an inference device that can swiftly and accurately infer the operation state of a radiation detection device. An inference device includes: data acquisition circuitry which acquire, as training data, a state signal outputted from a radiation detection device and indicating an operation state of the radiation detection device; and an inference circuitry which outputs the operation state of the radiation detection device on the basis of the state signal acquired by the data acquisition circuitry, using a trained model for inferring the operation state of the radiation detection device, the trained model being constructed through machine learning based on a feature quantity of the state signal.