Automatic Analyzer Dispensing Abnormality Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automatic analyzers face challenges in accurately detecting dispensing abnormalities due to biased reference database configurations and increased calculation complexity, particularly when dealing with small pressure variations from issues like bubble sucking, which can lead to erroneous determinations and increased computational demands.
Innovation Solution
An automatic analyzer with a control unit that calculates feature quantities from pressure waveforms using a linear combination formula with optimal coefficients to determine normal suction or ejection of samples, reducing reliance on database balance and minimizing calculation complexity by employing a display unit for abnormality indication and controlling analysis operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Mahalanobis distance is used for dispensing abnormality detection, then determination can be performed based on reference database, but determination performance depends on configuration balance of reference database data and erroneous determination rate increases when data is biased
Solution Approach 1:
The invention changes the detection parameter from Mahalanobis distance (which requires balanced reference database configuration) to a pressure differential parameter (difference between suction pressure and ejection pressure). This parameter change makes the detection method independent of reference database balance, thereby improving both detection accuracy and determination reliability without being affected by data configuration bias.
2Measurement precision
If Mahalanobis distance calculation is performed with increased feature quantities, then determination performance may improve, but calculation amount increases enormously
Solution Approach 1:
The invention extracts only the essential feature quantities needed for dispensing abnormality detection - specifically the suction pressure and ejection pressure values - and calculates the abnormality based on their differential. This extraction approach avoids the need for numerous feature quantities and complex Mahalanobis distance calculations, significantly reducing calculation complexity while maintaining detection precision.
3Object-affected harmful factors
If liquid level detection technique is used to minimize probe immersion depth, then contamination may be reduced, but bubbles or liquid film on liquid level are erroneously detected leading to dispensing abnormality
Solution Approach 1:
The invention replaces the mechanical/capacitive liquid level detection system (which is susceptible to false detection by bubbles or liquid film) with a pressure-based detection system. By measuring suction pressure and ejection pressure differentials, the system can reliably detect dispensing abnormalities without being misled by surface phenomena, thereby maintaining low probe immersion depth while improving detection reliability.
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
This approach enables high-accuracy dispensing abnormality detection without compromising determination performance or increasing calculation loads, providing stable and efficient analysis operations despite variations in reference database configurations.
Implementation Method 1
a pressure sensor that detects internal pressure of the sample dispensing nozzle
Data Source
AI summary
An automatic analyzer which is capable of detecting a dispensing abnormality with a high degree of accuracy without causing the decrease in the determination performance or the increase in the calculation amount caused by the configuration balance of the reference database is implemented. A dispensing nozzle of a sample dispensing mechanism 50 is immersed in a dispensing target contained in a specimen container 11 and sucks the dispensing target, and internal pressure of the dispensing nozzle of the sample dispensing mechanism 50 of ejecting the sucked dispensing target to a reaction container 41 is detected through a pressure sensor 54. A plurality of feature quantities are extracted from a waveform of the detected pressure, and a determination result is output through a linear combination formula using an optimal coefficient for a determination function that receives a plurality of feature quantities and outputs one value. The determination result indicates whether or not dispensing of the sample dispensing mechanism 50 is performed normally in accordance with the magnitude of the output result of the determination function.