Extraction And Procession Of The Signal Of Broken Wires Of Wire Ropes Based On Magnetoresistive Sensor

As a kind of flexibility component, the wire ropes are applied widely. However, after used for a period, it is inevitable that some wire ropes may break. If the broken wires are not found out timely, severe affairs maybe happen. The reason that the wire ropes break in using is often that few broken wires product a mass of concentrative broken wires. Thus, finding out the broken wires early is very important to wire ropes' inspection. However, the resulted magnetic flux leakage is so weak that highly sensitive and well performing sensor should be adopted, and the signals produced by the broken wires must be processed precisely. These steps are of great importance to the wire ropes'inspection. In this paper, these technical problems are discussed and researched.In this research, the characteristic of several magnetic field sensors is compared. In view of the fact that the signal of broken wires is very weak, the magnetoresistive sensor that has very fine performance synthetically is chose, at the same time a stimulating circuit is designed. Taking into consideration of the inspection range of magnetoresistive sensor, under the status of universality, the magnetic leakage flux of the broken ropes is computed with the electromagnetic finite element method, and the characteristic of the magnetic leakage flux is described. The result can be used as the theoretical basis of magnetic testing system.Because the intensity of magnetic leakage flux of broken ropes is very weak,the output voltage of magnetoresistive sensor is very low, thus making it very difficult to discover the broken wires. Consequently, before the broken wires are recognized the output signal should to be amplified. In this sense, the amplifying circuit for the signal of broken wires is designed. Under the condition of laboratory, the source of noises in the signal is researched. According to the frequency spectrum, a filter is designed. How to process the signal is discussed with software, and the Fourier Transform and Wavelet Transform are compared, then the signal is decomposed to the third layer with Wavelet Transform. According to the characteristics of noise, low frequency part of the third layer is kept as the reconstructed signal, and the perfect signal of broken ropes is obtained fully with the method of smoothness processing. At the last, the absolute peak value and peak-peak value of the processed signal are selected as the input of BP neural network, and are used to recognize quantitatively the broken wires of wire ropes.