Fluid Disruption Detection Using Eigenvector Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for detecting disruptions in fluid connections between extracorporeal blood processing systems and the cardiovascular system face challenges in accurately distinguishing between pump pulses and heart or breathing pulses, leading to difficulties in reliably detecting venous needle dislodgement and other potential malfunctions, especially when pump pulses overpower physiological signals.
Innovation Solution
A monitoring device that utilizes a source separation algorithm to process pressure signals from pressure sensors, computing eigenvectors and eigenvalues to identify disruptions without the need for extensive signal filtering, allowing for robust and efficient detection of fluid connection disruptions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional filtering techniques are used to remove pump pulses from pressure signals, then the signal can be cleaned, but the filtering complexity increases and stability issues are introduced
Solution Approach 1:
The patent transforms the pressure signal from time domain to frequency domain using Fourier transform, then identifies and removes pump pulse frequencies by analyzing spectral peaks. This parameter transformation approach changes the signal representation to make pump pulse removal more reliable without introducing complex filtering algorithms, thereby resolving the contradiction between detection reliability and system complexity
Solution Approach 2:
The patent replaces conventional mechanical filtering approaches with a spectral analysis-based method. Instead of using complex filter hardware or algorithms to remove pump pulses, the system uses Fourier transform to identify pump pulse frequencies and subtracts them mathematically, achieving more stable and reliable signal cleaning with simpler implementation
2Measurement precision
If pump pulses are strongly present in the pressure signal, then the blood pump is functioning, but the heart and breathing pulses become difficult to detect
Solution Approach 1:
The patent performs preliminary spectral analysis of the pressure signal to identify pump pulse frequencies before attempting to detect heart and breathing pulses. By first characterizing the pump pulse spectrum and then removing or suppressing these frequencies, the method prepares the signal in advance to enable accurate detection of the weaker physiological pulses that would otherwise be obscured by pump interference
Solution Approach 2:
The patent converts the harmful effect of strong pump pulses into a useful feature by using the consistent spectral signature of pump pulses as a reference. The identified pump pulse frequencies are subtracted from the original signal, transforming the interfering pump pulses into a known quantity that can be removed, thereby revealing the hidden heart and breathing pulses for accurate VND detection
Data Source
Figure 1~2(b)
Figure 3(a)~5
Figure 6~7(d)
AI summary
A monitoring device detects a disruption of a fluid connection between first and second fluid containing systems using one or more pressure sensors arranged in the first fluid containing system to detect first pulses from the first fluid containing system and second pulses from the second fluid containing system. The monitoring device receives (501) pressure signal(s) from the pressure sensor(s), populates (504) signal vectors by signal segments in the pressure signal(s) and computes (505) one or more eigenvectors and/or one or more eigenvalues for the signal vectors by a source separation algorithm. The monitoring device detects (506) the disruption based on a monitoring parameter, which is computed as a function of the eigenvector(s) and/or eigenvalue(s) to be responsive to the second pulses in the pressure signal(s). The monitoring device may be associated with or included in an apparatus for extracorporeal blood processing, such as a dialysis machine.