Normal condition
No induced leak; otherwise comparable setup.
False alerts and whether normal remains normal.
A working bench prototype supports controlled chamber tests, recorded sensing data and a monitoring interface.

The acrylic chamber, replaceable test fixtures, pressure measurement and film sensors allow us to study different stages of an experiment.


Recorded one-second RMS features show different signal intensities during evacuation and air-release stages.

Video unavailable. Please check the media file.
Chamber and connected sensing hardware.
Video unavailable. Please check the media file.
Recorded signal and monitoring workflow.
Video unavailable. Please check the media file.
Prototype startup and indicator activity.
Building on recorded two-channel data, the next stage will compare sensor responses across known leak locations. The test matrix and number of repeats will be adapted to the experimental setup, with consistent conditions for each group of trials.
No induced leak; otherwise comparable setup.
False alerts and whether normal remains normal.
Known leak positions with repeated trials.
Whether the true region is ranked first or in the top two.
Vary opening size and introduce controlled disturbances.
Ranking stability and response time.
Record offline, missing or saturated channels.
Whether insufficient evidence is recognized.
Keep initial pressure, sensor mounting, recording window and acquisition settings consistent within each comparison.
Record the condition, known leak position, channel data and interruptions. Retain raw records separately from processed features.