KR100456841B1 - Method for processing detection signal in duration test of vehicle - Google Patents

Abstract

In the endurance test device of the vehicle, the signals obtained from the sensors attached to various parts of the vehicle are separated and extracted into sound emission signals and vibration signals so as to obtain accurate test results for each part of the vehicle.

The present invention is a process for real-time processing the signal obtained from the sensor installed in each part of the vehicle to separate the acoustic emission signal and the vibration signal, and the wavelet conversion of the separated acoustic emission signal to decompose the step by step scale signal, step by step Analyze the scaled signal to determine whether there is damage in any part, and if it is determined that there is damage in any part, calculate the difference of arrival time of each sensor and then analyze the calculated result through neural network Detecting a location of damage and a corresponding component and a magnitude of damage; extracting a natural frequency by analyzing the separated vibration signal; and comparing the extracted natural frequency with a natural frequency set for the corresponding part. Detecting whether there is a change in the natural frequency, and a change in the natural frequency is detected. Comprises the step of determining to be a great damage and breakage in the area occurred.

Description

Detection signal processing method in vehicle endurance test {METHOD FOR PROCESSING DETECTION SIGNAL IN DURATION TEST OF VEHICLE}

The present invention relates to an endurance testing apparatus for a vehicle, and more particularly, to extract and analyze a signal obtained from a sensor attached to various parts of a vehicle into an acoustic emission signal and a vibration signal, thereby obtaining accurate test results for each part of the vehicle. The present invention relates to a detection signal processing method in a vehicle endurance test that can be obtained.

In general, the basic method of detecting damage or damage such as cracks generated in the vehicle body and chassis during the actual vehicle durability test or the target durability test of the vehicle is that the endurance tester observes each part through the naked eye on the vehicle that has traveled over a certain distance. It is common to extract the results.

Therefore, in order to inspect an invisible part of the vehicle after driving a certain distance in the durability test using a real vehicle, the vehicle has to be disassembled and the vehicle has to be periodically inspected.

In order to overcome this disadvantage, non-destructive inspection method that detects damage and damage in real time is used in aerospace, but there is no suitable method to apply it to automobiles. In the case of a simple signal processing method for frequency analysis of a signal acquired from a sensor installed at a site, there is a disadvantage in that quantitative damage detection is difficult.

The present invention has been invented to solve the above problems, the object of which is to detect the damage during the endurance test of the vehicle to each part of the vehicle (X-MBR, S / ABS, SPRING, ARM, LINK, body, etc.) A suitable piezoelectric ceramic sensor is attached, and signals obtained from the attached piezoelectric ceramic sensor are separated into acoustic emission signals and vibration signals through a wavelet transform, and each component of the vehicle is analyzed by acoustic emission signals and vibration signals. It provides accurate information about when the damage occurred, the rate of failure development, the extent of failure, and the final failure mode, and provides accurate life prediction and lightweight design modifications for the components.

1 is a schematic configuration diagram of an endurance testing apparatus for a vehicle according to the present invention.

2 is a flow diagram of one embodiment for performing detection signal processing in a vehicle endurance test in accordance with the present invention.

Figure 3 is a waveform diagram of an embodiment showing a step scale results for the acoustic emission signal among the signals detected in the vehicle endurance test according to the present invention.

Figure 4 is a waveform diagram of an embodiment showing a natural frequency extraction result through the spectrum analysis of the signal detected in the vehicle endurance test according to the present invention. Figure 5 is for any signal S detected in the vehicle endurance test according to the present invention. State diagram showing extraction of acoustic emission signal and natural frequency through wavelet transform.

The present invention for realizing the above object comprises the steps of real-time processing the signal acquired from the sensor installed in each part of the vehicle and then separating the acoustic emission signal and the vibration signal by applying the wavelet transform; Decomposing the separated sound emission signal into a step scale signal and determining whether any part is damaged by frequency component analysis; If it is determined that there is damage in any part of the sensor, calculating a difference in arrival time of each sensor and analyzing the calculated result through neural networks to detect a location of damage and a corresponding part and magnitude of damage; Spectral analysis of the separated vibration signal to extract a natural frequency, and then comparing the extracted natural frequency with a set natural frequency of a corresponding region to detect whether there is a change in natural frequency; When a change in natural frequency is detected, a detection signal processing method is provided in a vehicle endurance test comprising determining that a large damage and breakage has occurred in a corresponding region.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

As can be seen in Figure 1 is installed in each part necessary for the durability test of the vehicle, that is, X-MBR, S / ABS, SPRING, ARM, LINK, the vehicle body, etc., the deformation or crack of the corresponding parts as the durability test proceeds The sensor 10 for detecting the degree of damage and outputting the damage information, the amplifying unit 20 for amplifying the analog signal applied from the sensor 10 to a set level, and the amplified parts. Data is acquired by sampling the endurance test information in a set frequency unit, and a programming language (VISUAL C ++, MATLAB, LABVIEW) is set for processing the acquisition of signals in real time, and wavelet transform ( Wavelet Transform) to separate the sound emission signal and vibration signal, and wavelet transform the separated sound emission signal to decompose the signal by the scale step. Detecting information of the production and size, it comprises a diagnostic unit 30 for detecting an impairment or damage to the components by spectral analysis of the isolated vibration signal.

In the present invention including the functions described above, a signal processing process acquired by the sensor installed at each site is as follows.

Each part of the vehicle for durability test, that is, X-MBR, S / ABS, SPRING, ARM, LINK, vehicle body, etc. is installed with a sensor 10, preferably a piezoelectric ceramic sensor, and then subjected to durability test by running the vehicle. When the sensor 10 installed in each part detects a change in a signal due to deformation, crack, or breakage of the corresponding part and applies it to the amplifier 20 side, the amplifier 20 sets a predetermined signal. Amplify to the level of and apply to the diagnostic device (30) side.

The data acquisition board (not shown) in the diagnostic apparatus 30 reads out the signal of the sensor 10 that is input and samples at the set sampling frequency, preferably 1 MHz, to acquire the endurance test signal for each part, VISUAL C ++, A real time signal processing program set in MATLAB, LABVIEW, etc. allows the acquisition of information in real time (S101) (S102) (S103) .FFT Fourier Transform is a method commonly used in analyzing signals. ) And the wavelet transform are applied. The FFT transform is a method of analyzing various frequency components by amplitude based on a sine wave, and the wavelet transform is a transformed function, not a harmonic function, for a transformed signal. Is a method of decomposing a signal by applying as a basic function.

Accordingly, the signals acquired in real time from the sensor 10 mounted on each part of the vehicle, preferably the piezoelectric ceramic sensor, are wavelet-converted and separated into acoustic emission signals and vibration signals (S104) (S105) (S106). From the separated signal as described above, the acoustic emission signal is decomposed into a scale signal (step S107), and then it is detected whether there is a signal generation in a specific step for abnormal detection (step S108). As described above, when the wavelet transform is applied to the signal S measured from the sensor 10 mounted at an arbitrary place, the vibration signal A1 and the acoustic emission signal D1 are decomposed. From the low detail signal (highest frequency of D1), the signal is decomposed into higher and higher levels of detail signal. For example, as shown in the drawing, the vibration signal A1 and the acoustic emission signal D are measured with respect to the measured signal S. 1) primary separation, vibration signal A1 to vibration signal A2 and acoustic emission signal D2, and vibration signal A2 to vibration signal A3 and acoustic emission signal D3 In this case, the measured signal is S = A1 + D1 = A2 + D1 + D2 = A3 + D1 + D2 + D3. When the signal generation at is detected, the difference in arrival time between each sensor of the acoustic emission signal is calculated (S109), and the result of the analysis is performed by analyzing the calculated result using a neural network to detect the location of the damage, the component in which the damage occurs, and the size of the damage ( S110) (S111).

As described above, the principle of detecting the location of the damage, the part where the damage occurred, and the magnitude of the damage by the step-by-step signal after wavelet conversion is as follows.

When an arbitrary structure material for analysis is determined, the acoustic emission signal according to the failure mode of the structure has a constant frequency component, so that the step-by-step scale after wavelet transformation is decomposed into at least three stages as shown in FIG. From this, it is possible to determine whether damage to the structure occurs.

In addition, by analyzing the components of the vibration signal separated by the wavelet transform and signal decomposition in S104 to detect the natural frequency (S113) (S114).

Thereafter, the detected natural frequency is compared with the natural frequency for the corresponding part to detect whether a change in the frequency is generated (S115).

If it is determined that a change in the natural frequency has occurred in the above, it is determined that a large damage and breakage has occurred for the corresponding site, and the information about it is instructed to the endurance test facilitator through the monitor (S116).

As can be seen in Figure 4, since the natural frequency of the structure has a very low frequency range compared to the acoustic emission signal, A3 is a low frequency signal containing the natural frequency is decomposed from the original signal (PZT) obtained from the sensor The signal is analyzed by power spectrum to extract the natural frequency.

Subsequently, by comparing the natural frequencies for each of the preset parts with the natural frequencies extracted by spectral analysis, if there is a reduction due to the damage of the frequency, it is determined that large damages and damages to the corresponding parts are generated.

As described above, the present invention separates a signal detected from a sensor into an acoustic emission signal and a vibration signal, and then accurately analyzes the occurrence time, the propagation speed, and the final damage of the vehicle component by analyzing the scale of the acoustic emission signal and the neural network analysis. Accurate information about the mode can be obtained to provide reliability for endurance testing.

In addition, based on this, it is possible to accurately predict the life of the parts and to modify the design to reduce the weight, and to predict the propagation speed of the damage occurrence for each part, thereby eliminating endurance tests that should be performed periodically.

Claims (2)

A process of real-time processing a signal acquired from a sensor installed in each part of the vehicle, and separating the acoustic emission signal and the vibration signal by applying wavelet transform; Decomposing the separated sound emission signal into a step scale signal and determining whether any part is damaged by frequency component analysis; If it is determined that there is damage in any part of the sensor, calculating a difference in arrival time of each sensor and analyzing the calculated result through neural networks to detect a location of damage and a corresponding part and magnitude of damage; Spectral analysis of the separated vibration signal to extract a natural frequency, and then comparing the extracted natural frequency with a set natural frequency of a corresponding region to detect whether there is a change in natural frequency; Detecting a change in the natural frequency is a detection signal processing method in a vehicle endurance test comprising the step of determining that a large damage and breakage occurred in the area. delete