LC Stream Pressure Classification for Analyzer State Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated analyzers face errors and malfunctions due to increased complexity, leading to decreased productivity and unreliable measurement results, often requiring external service personnel for extended periods, which can be inefficient and costly.
Innovation Solution
A method for monitoring the state of a liquid chromatography stream by automatically classifying system pressure time series during sample injection, using machine learning algorithms to distinguish between different states and trigger appropriate responses, reducing the need for external expertise and minimizing downtime.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated analyzers are made more complex to increase throughput and functionality, then productivity is improved, but errors and malfunctions increase leading to decreased reliability
Solution Approach 1:
The system continuously monitors system pressure during injection processes and feeds this information back to the control unit, which automatically responds by pausing the workflow or alerting operators when anomalies are detected, thereby maintaining reliability despite increased system complexity
Solution Approach 2:
The automated analyzer performs self-diagnosis through automatic monitoring and classification of injection states, enabling the system to detect and respond to its own malfunctions without external intervention, thus maintaining reliability while supporting high throughput
2Reliability
If external service personnel are called to fix errors, then reliability is improved, but downtime increases leading to decreased productivity
Solution Approach 1:
The system automatically monitors injection processes, classifies states using machine learning models, and triggers appropriate responses including pausing workflows or alerting operators, enabling self-diagnosis and reducing the need for external service personnel
Solution Approach 2:
The system performs preliminary classification of injection states and identifies potential issues before they lead to complete system failures, allowing proactive responses that prevent extended downtime and reduce the need for external service intervention
3Reliability
If monitoring and classification systems are added to detect different injection states, then reliability is improved, but device complexity increases
Solution Approach 1:
The system replaces complex mechanical monitoring equipment with electronic pressure sensors and software-based machine learning classification algorithms, achieving reliable injection state detection through information processing rather than additional mechanical components
Solution Approach 2:
The monitoring system uses existing pressure measurement capabilities of the automated analyzer for their original purpose while simultaneously utilizing these same measurements for injection state classification and anomaly detection, making the existing system perform multiple functions without adding dedicated hardware
4Productivity
If automated monitoring and classification is implemented, then productivity is improved by reducing downtime, but use of energy increases due to continuous monitoring
Solution Approach 1:
The system continuously monitors pressure data during injection processes using existing operational measurements, rather than implementing separate continuous monitoring cycles, thereby maintaining productivity benefits while minimizing additional energy consumption by reusing already-acquired data
Data Source
Figure 1
Figure 2a~2c
Figure 3
AI summary
In one general aspect, the present invention relates to a method of monitoring a state of a liquid chromatography (LC) stream of an automated analyzer. The method includes automatically monitoring a system pressure of an injection assembly of the liquid chromatography stream to generate a time series of system pressures, classifying the time series in one of two or more predetermined classes indicating different states of the LC stream and triggering a response based on the classification result.