一种液体物质成分检测方法、系统和存储介质
By converting liquid-solid triboelectric signals into a spectrum and utilizing a convolutional neural network, the problem of traditional methods being insensitive to subtle differences is solved, achieving high-precision detection of liquid substance composition.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING NORMAL UNIV AT ZHUHAI
- Filing Date
- 2026-06-17
- Publication Date
- 2026-07-17
AI Technical Summary
Existing methods for detecting liquid components are ineffective in distinguishing subtle differences in liquid-solid triboelectric signals, especially in terms of low accuracy and robustness in differentiating solution mechanical properties and inorganic solutes. Traditional time-domain analysis methods are also insensitive.
The original electrical signal is converted into a spectrum signal. The frequency domain energy distribution is extracted by discrete Fourier transform and overlapping framing strategy. Combined with convolutional neural network to autonomously learn the characteristics of liquid substance composition, a nonlinear mapping relationship from spectrum features to substance classification is constructed.
It significantly improves the accuracy and robustness of liquid substance component identification, and can efficiently distinguish between electrolytes and non-electrolytes, different metal ions and other similar substances, achieving efficient and low-cost end-to-end detection.
Smart Images
Figure CN122409484A_ABST