Pressure change leak detection using exponential analysis

Abstract

The pressure change leak test detects a leak from the pressure drop observed after the container under test is filled with compressed gas and then closed. This method has the serious disadvantage of being sensitive to the temperature variation in the target container. We propose a new method using exponential analysis to compensate for exponential temperature variation in pressure change leak detection. In the proposed method, two successive leak tests are performed at different initial compression pressure. Exponential parameters of the temperature variation in the gas within the container are determined from the pressure signal in the first test, and the pressure in the second test is predicted using these parameters. The leak in the container is estimated from the difference between the predicted pressure and the measured one. Experimental results using a prototype leak detector and model piping showed that leak can be successfully detected without being affected by the given exponential temperature variation. With the proposed method, it was able to shorten the total test time by as much as 65% in comparison with the case in which a conventional leak test is performed after waiting until the exponential temperature variation settles. An algorithm using Discrete Fourier Transform for estimating exponential parameters from finite length data of pressure signal is also presented.