PIPELINE LEAK DETECTION USING SYSTEM IDENTIFICATION

摘要

Of all the methods available for systems dynamics identification, only one is capable of being used during the normal operation of the system. This method uses an artificially created signal which consists of a series of pulses which have statistical properties identical to that of white noise, but are periodic in nature. This is termed Pseudo Random Binary Sequence (PRBS), and because of its periodicy, requires much shorter sampling times than true white noise, to complete signal processing. This technique has been developed to detect leaks in hydraulic pipe lines, using a GENRAD Micro Modal analyser for signal processing. This paper describes the theoretical basis of the method which makes use of the imaginary part of the pressure frequency response function. Some experimental results are also presented, taken from an oil hydraulic pipe line, with a simulated leak. The test results show that this method is reasonably accurate, and has a significant potential as a general leak detection system for pipe lines.