A Novel Looseness Detection Method for Hydraulic Pipeline Clamp Based on Statistical Analysis

The pipeline is widely used in various kinds of mechanical equipment, usually fixed by the clamps. It is great important to detect the fixed clamps for the operation of equipment by signal processing. The time domain signal is quite popular in the field of signal processing. There are generally two major types of methods for analyzing time-domain signals. The one is that the time domain signal is converted into other domains such as frequency domain and time frequency domain; the other is using statistical analysis methods to analyze signals. This paper presents a new feature extraction method based on statistical analysis, which is called the number of conformation patterns (CPN). The phase space reconstruction of time series is carried out, and the numbers of conformation patterns with different embedding dimensions are counted to be features for the signal. Through the simulation analysis, the characteristic values are sensitive to the frequency of the signal. This method is used to analyze the experimental data of the looseness of the pipeline clamp, and the results show that CPN can effectively detect the different degrees of clamp looseness.