Research On Multi-channel Weak Signal Detection Of Loose Particle

Seen as the key infrastructure components of military, aerospace and defense equipment, whether sealed electronic components can be stable or not is directly related to the reliability of the whole system. As an impurity in the inner of electronic component, the loose particles will have certain influence on the devices and even endanger the reliability of device because of its randomness and uncontrollability. Existing PIND system and method has a accuracy rate of up to 90% for loose particles of 0.02 mg and larger ones, But can do nothing with the weak signal produced by loose particles smaller than 0.02 mg. Therefore, the study on weak signal detection of loose particles is crucial for the development of loose particles detection and the elimination of hidden dangers of system reliability.For the problem of traditional test equipment can not effectively detect the weak signal of loose particles, weak signal acquisition module is designed in this paper. Analyze the causes of weak signals, and redesign each link of signal acquisition channels. Selecting high sensitivity resonant acoustic sensor with three channels, designing three-channel conditioning circuit and four-channel acquisition circuitry to realize the purpose of signal acquisition. Suppressing noise from the perspective of power source, transmission lines, PCB anti- jamming, shielding and grounding to meet the low noise requirement, which lays the foundation for signal processing.For the problem of a variety of noise contained in the signal of PIND test will interfere the weak signal detection, the characteristics of acoustic emission signals and PIND test signals are analyzed secondly, and it shows that the background noise and environment noise are the two most important sources of test signal noise. For background noise, wavelet analysis and thresholding means with improved threshold function is applied to filter the noise. For environmental noise, the Kalman filtering method with strong filtering effect on Gauss white noise is applied and then the denoising result are analyzed.For the problem of excessively small amplitude of weak signal, exiguous signal captured by single sensor, even the signal’s error, loss, distortion and so on due to the differences of sensor, a three-channel sound sensor is employed to receive the PIND signal. By analyzing the object of fusing loose particles weak signal, required results and the characteristics of three levels of data amalgamation to s elect the data level fusion method. Considering the poor objectivity and universality exists in the traditional weighted fusion method, a new multi-channel data algorithm based on quantum genetic algorithm is adopted to fuse three-channel loose particles weak signal. At last, the simulation and experimental results verify the effectiveness of proposed algorithm.For the weak signal of loose particles that is too small to identify or improve signal-to-noise ratio by software and hardware methods, a stochastic resonance method is formulated to enhance and extract the target signal. Currently study of stochastic resonance are mostly concentrated in the extraction of the period signal, while the loose particles signal herein is a typical non-periodic signal. Therefore, aperiodic stochastic resonance and its response to the results, its characterization and mechanism is the first to be studied. Through establishing the loose particles signal model to study the effect of various parameters on bistable system and th en to determine the appropriate conditions and parameters of loose particles weak signal detection. Experimental results show that the raised method for detecting loose particles weak signal has a 66.7% accuracy rate.The research of weak signal detection method of loose particle in this paper provides basis and method for detection sensitivity of loose particle, which can also provide a reference for weak signal detection of other areas.