Detecting Pressure Leakages in Hyperloop Pods Using Acoustic Emission Signals

Leak detection in a hyperloop pod is essential to guarantee the safety of passengers; however, traditional detection methods are slow and unreliable. Pressure sensors can only detect a leak after a pressure drop has already occurred, which means the effects of the leak may already be significantly affecting the pod. Additionally, these sensors are insensitive to smaller changes in pressure and experience variations due to the environment (temperature, vibrations, and movement). The hyperpod will require proactive solutions to the depressurization problem, not a reactive one.

I found a potential solution by reading a paper titled “Pressure vessel leakage detection method based on online acoustic emission signals” by Liu et al. In their study, they proposed a leakage detection method using acoustic emission (AE) signals to detect sounds that characterize gases moving in and out of vessels. Using signal processing, a system for online, reliable, and real-time detection of pressure changes using AE signals can be made.

AE signals are generated when leaks generate high-frequency waves that reflect and propagate inside the vessel. These waves can be captured by piezoelectric acoustic emission detectors that detect mechanical stress due to the waves and convert it into electric signals. However, using AE signals by themselves is not sufficient; they have strong noise interference and are sensitive to a broad range of frequencies.

To combat these disadvantages, Liu et al. implemented a Singular Value Decomposition (SVD) and Adaptive Singular Value Decomposition (ASVD) algorithm to filter and extract meaningful features in the context of pressure changes from the AE sensor data. SVD is a mathematical technique that involves taking signal matrices (found in the data collection step) and decomposing them into several components to separate the important signal characteristics from background noise. ASVD builds on this technique by updating the filtering method to take into account changing parameters. For example, changes in air vibration, turbulence, and environmental noise will be taken into account in the decomposition process to separate possible leakage events from other signals accurately.

The SVD and ASVD filtering techniques will be instrumental in the application of the hyperloop. The magnetic system, high speed, and passenger noise create vibrations that need to be excluded in analysis. By installing AE sensors along the pod’s hull and using the robust signal processing algorithms provided by Liu et al. and similar research, real-time information about the integrity of the pod body and possible risks can be provided. The viability of these algorithms in the context of hyperloop technology is still not certain, but I believe it’s a solution worth pursuing.