Research On Vibration Measurement Technology Of Tip-timing System Based On GRNN

The solution is desiderating for power industry, energy industry, aviation and shipping to real-timely measure the high-speed rotating blade vibration. The conventional contacting measurement technique is difficult to measure the vibration of all blades in one stage at the same time. So the scholars of domestic and abroad have being researching a new non-contact measurement technique using blade tip-timing method, which uses of the feature that the time of rotating blade arriving at sensors is different to count blade vibration amplitude and frequency.The blade tip-timing technique have become very mature in the hardware owing to the rapid development of laser technique and electron technique, but the problems of it's data processing and sensor installing method still haven't been completely solved, which seriously blocks the development of the blade tip-timing technique.This treatise analyses all the kinds of the data processing methods of blade tip-timing technique and indicates each strongpoint and shortcoming. Hereby, it is concluded that data processing methods that exists at present can not simultaneously process synchronous vibration data and asynchronous vibration data. The dissertation puts forward a novel data processing method for tip-timing technique based on artificial neuron net(ANN), which is combined with how to install sensors and accords as the feature that ANN can fit any function curves to process data, not only can it simultaneously process synchronous vibration data and asynchronous vibration data , but also overmatches other methods in the precision.In the meantime this treatise testifies it's feasibility by simulation and actual experiment. In terms of actual experiment, it attains to the experiment aim by comparing the measure frequency, by the method of using strain pieces, with that ,by tip-timing technique combined with the novel data processing method.A real-time measurement program is programmed after the novel dataprocessing method is testified, whose data capture, display and processing are achieved by use of LABVIEW, which has reduced using the experiment instrument and the experiment cost.