Chromatograph Virtual Curve Fitting With Segmented Plot Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing chromatography technologies face challenges in efficiently reducing arithmetic processing load and noise influence during virtual curve calculation, particularly when dealing with a small number of plots or a large number of noisy data points.
Innovation Solution
A data processing device that generates a smaller number of virtual curve calculation data points through bunching processing, followed by virtual curve calculation with reduced noise influence, using methods like Savitzky-Golay smoothing and non-linear least-squares regression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If curve fitting is executed using a large number of plot data points, then the fitting accuracy is improved, but the arithmetic processing load increases
Solution Approach 1:
The plot data is divided into multiple groups, with each group containing a predetermined number of data points. Virtual curve calculation is performed separately for each group, and the results are integrated to obtain the final virtual curve. This segmentation allows accurate curve fitting to be achieved through multiple small-scale calculations rather than one large-scale calculation, thereby reducing the arithmetic processing load while maintaining fitting accuracy.
2Productivity
If curve fitting is executed using a small number of plot data points, then the arithmetic processing load is reduced, but the curve fitting accuracy deteriorates
Solution Approach 1:
Instead of performing curve fitting on all plot data at once or on a single small subset, the data is segmented into multiple groups. Each group undergoes virtual curve calculation independently, and the resulting curves are integrated. This approach ensures that each individual calculation remains computationally lightweight while the collective result achieves high accuracy through the integration of multiple fitted curves.
3Measurement precision
If virtual curve calculation is performed on original plot data, then the curve fitting accuracy is improved, but the noise influence increases
Solution Approach 1:
The plot data is divided into multiple groups for separate virtual curve calculation. By processing data in smaller segments rather than as a single large dataset, the influence of noise on any individual curve fitting operation is reduced. The subsequent integration of multiple fitted curves from different groups further mitigates noise effects, as random noise tends to average out across multiple independent calculations.
Data Source
AI summary
To reduce an arithmetic processing load or an influence of noise at the time of virtual curve calculation processing, provided is a data processing device for a chromatograph, which is configured to execute data processing based on plot data measured by using a chromatograph, the data processing device including a virtual curve calculation data generator configured to obtain a smaller number of pieces of virtual curve calculation data than a number of pieces of the measured plot data; and an arithmetic processor (163) functioning as a virtual curve calculator configured to obtain a virtual curve based on the virtual curve calculation data.