Multimodal Fluid Sensing With AI for Selective Composition Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fluid sensors, particularly optical and semiconductor sensors, face limitations in selectivity and stability, necessitating improved methods for accurately determining fluid composition.
Innovation Solution
A method involving multiple sensing modalities, including photoelectrochemical and optical sensors, is used to generate training data for an artificial intelligence engine, which is trained to infer fluid composition based on combined sensor outputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical sensors are used for fluid composition detection, then sensitivity and multi-constituent detection capability are improved, but selectivity is limited due to narrow wavelength sensitivity and environmental noise
Solution Approach 1:
The patent combines multiple sensing modalities (optical sensors with semiconductor sensors) into a hybrid sensing system. The optical sensors provide sensitive detection across multiple wavelengths, while the semiconductor sensors add selectivity through their material-specific responses to different fluid constituents. This merging of complementary sensing approaches resolves the contradiction between sensitivity and selectivity.
Solution Approach 2:
The sensing system is designed to perform multiple functions simultaneously: optical sensors detect a broad range of constituents across different wavelengths, while semiconductor sensors provide selective detection of specific compounds. The AI engine integrates these multiple functions to achieve both high sensitivity and high selectivity in fluid composition analysis.
2Measurement precision
If semiconductor sensors are used for fluid composition detection, then sensitivity is improved, but long-term stability deteriorates
Solution Approach 1:
The patent merges semiconductor sensors (high sensitivity, lower stability) with optical sensors (good stability, high sensitivity) into a hybrid system. The AI engine processes data from both sensor types, leveraging the strengths of each to achieve both high detection sensitivity and long-term operational stability that neither sensor type could achieve alone.
Solution Approach 2:
The AI engine continuously processes sensor data and adjusts its predictions based on patterns learned from training data. This feedback mechanism allows the system to compensate for drift or degradation in sensor performance over time, maintaining long-term stability while preserving the high sensitivity provided by the semiconductor sensors.
3Measurement precision
If multiple sensing modalities are combined with AI processing, then fluid composition analysis accuracy is improved, but system complexity increases
Solution Approach 1:
The AI engine serves as a universal processing platform that handles data from multiple sensor types (optical and semiconductor sensors) with different modalities. Rather than requiring separate processing systems for each sensor type, the single AI engine integrates and analyzes all sensor inputs, achieving high composition analysis accuracy while managing system complexity through centralized intelligent processing.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The approach enhances the accuracy and reliability of fluid composition analysis by leveraging synergies between different sensing modalities, improving selectivity and stability.
Implementation Method 1
Semiconductor materials are activated through exposure to photons with a minimum energy that is larger than the semiconductor band gap
Implementation Method 2
Optical sensors detect radiation (e.g., visible and/or UV light and/or infrared light) that has interacted with a fluid
Data Source
AI summary
Training data is generated over a plurality of trials. In each trial, training information about a training fluid from sensors having at least two different sensing modalities is obtained. After generating training data, an artificial intelligence engine is trained on the training data. After training the artificial intelligence engine, the artificial intelligence engine infers a composition of a fluid based at least in part on deployed sensors. The artificial intelligence engine can be further trained using legacy sensors at an industrial facility.