Automatic selection of a suitable coherence frequency domain [articol]

Abstract

The improvement of techniques which allow pinpointing of leaks in pipe transportation networks is a priority for companies and authorities around the world. The flow of liquids (water) trough a pipe generates specific auditive (noise) signals. If the pipe has leakage points or other faults, then we face problems of liquid loss. The liquid which comes out of the pipe generates specific leak signals transmitted in the material of the pipe and in the liquid which is inside the pipe. These leak signals contain parasite information coming from other noise sources like pipe elbows or junctions. The signals must be filtered in order to determine the best frequency domain in which they should be analyzed and to remove any unwanted frequencies. The domain of interest must be the one where the signals are most coherent. At the same time, the limits of the filter must be determined in an automatic way in order to ease the work of the person who uses the application. After the domain of interest is found, the Cross Correlation Function (CCF) between the signals can be calculated in order to determine the delay between the two signals. This paper presents a way in which one can establish the frequency domain which is most suitable for analyzing leak signals. We are interested to show that the quality of the CCF grows after the automatic filtering process is used. The programs used in this paper were implemented with the help of Matlab 7.5 functions.