CN111323481B - Large-scale structure activity redundancy detection method based on sound signals - Google Patents

Abstract

A method for detecting the excessive movable objects of a large-scale structure based on sound signals comprises the steps of placing a microphone array and a collection board card thereof on the large-scale structure, connecting the microphone array and the collection board card thereof with a pc by using a data transmission line and a usb port line to form a sound signal collection system, collecting the sound signals generated when the large-scale structure rolls, carrying out spatial filtering on the sound signals, and improving the signal-to-noise ratio preprocessing of the collected sound signals; calculating a time-frequency diagram of the sound signal received by the pc, comparing the time-frequency diagram with a time-frequency diagram of a normal sound signal, and calculating the kurtosis Q and the spectrum skewness P of the signal to serve as supplementary criteria; the microphone array is applied to the field of detection of the large-scale structure active redundancy for the first time, the effective filtering of the environmental noise is realized through the beam forming technology, and meanwhile, the judgment of the existence of the redundancy is realized; the invention realizes real-time non-contact monitoring, reduces the monitoring difficulty and enlarges the application range.

Description

Large-scale structure activity redundancy detection method based on sound signals

Technical Field

The invention relates to the technical field of redundancy detection, in particular to a method for detecting active redundancy of a large-scale structure based on a sound signal.

Background

The surplus refers to all substances existing in the product, which are generated from the outside or from the inside and are not related to the specified state of the product. It can cause the product to lose its prescribed function, causing product failure, which is a difficult problem that plagues the reliability of aerospace devices. In 3 months 1993, when the columbia space shuttle was prepared before flying at the kennedy space center in the united states, the solid booster rocket fuel was found to leak, and the hair line mixed in an O-ring was found after the rocket assembly was disassembled. In 2000-2007, the annual average percentage of the number of the redundancy problems in the development process of a certain model of a certain general college in China is 9.2%, and the flight test progress of the model is influenced by the redundancy problems.

With the development of the aerospace field in China, the problem of redundancy is more and more concerned. For example, in chinese patent CN102435672A, "an inspection table, and an automatic device and method for inspecting extra materials of space electronic equipment using the inspection table", a crank rocker mechanism is used to drive the electronic equipment placed on a working table to swing, and the sound and vibration generated by the movement of extra materials during the swinging process are collected by a sensor and transmitted to an upper computer for analysis and judgment. Chinese patent CN103472196A "a device for detecting excess material" uses a detecting device to drive a product to rotate around two axes, and judges whether excess material is generated by listening to the product in a mute room to make a sound and observing whether excess material moves inside the product.

The said patent greatly raises the efficiency of detecting the redundant material and raises the reliability of the space product, but it can only solve the problem of detecting the redundant material of small space product and has no power for detecting the redundant material of large structure (such as storage tank and shell section) in the field of aviation and spaceflight. China has few detection researches on large structure redundancy. For example, according to standard QJ2850A-2011 "aerospace product excess prevention and control", residual excess is reduced mainly by standardizing the processes of design, processing and assembly stages, and the provided excess detection methods include visual inspection methods, auditory inspection methods and other methods, and the methods have low detection efficiency and poor accuracy. Chinese patent CN102955178A "a method and apparatus for detecting excess material in spacecraft cabin" uses a handle to swing the spacecraft cabin fixed on the cantilever structure, and uses an observation method and a listening method to judge and check whether there is excess material in the spacecraft cabin, which has the disadvantages of low reliability of the cantilever structure, large influence of human subjective factors on the observation method and the listening method, and low detection precision. Chinese patent CN201555700U, a device for detecting the movable excess of a closed cylindrical structure, realizes the detection of the movable excess of the closed cylindrical structure, but has the disadvantages that the flange guide rail connecting rod needs to be manually rotated, the detection efficiency is low if a large-scale structure is met, and the stability of a piezoelectric transducer is poor.

Disclosure of Invention

In view of the above problems, an object of the present invention is to provide a method for detecting active redundancy of a large structure based on an audio signal, which can accurately and efficiently determine whether the large structure has the active redundancy.

The technical scheme of the invention is realized as follows:

a method for detecting active redundancy of a large structure based on sound signals comprises the following steps:

the method comprises the following steps that firstly, a large-scale structure (7) is placed on a trolley and clamped by using a rolling ring (4), the rolling ring (4) is driven by a motor (8) to drive the large-scale structure (7) to rotate around the axis of the rolling ring (4), and the rotating speed of the motor (8) is controlled by a control cabinet (9);

secondly, placing the microphone array and the acquisition board card (1) thereof on a large structure (7), and connecting the microphone array and the acquisition board card (1) thereof with a pc (5) by using a data transmission line (2) and a usb serial port line (3) to form a sound signal acquisition system;

step three, opening the pc (5) to set acquisition parameters of the sound signal acquisition system; when a control cabinet (9) is opened, a motor (8) drives a rolling ring (4) to drive a large structure (7) to rotate around the rolling ring (4), and meanwhile, a microphone array and an acquisition board card (1) thereof transmit acquired sound signals to a pc (5); software installed on the pc (5) processes the collected sound signals, when no redundancy is generated, the system continuously collects the sound signals until the specified time is exceeded, the large-scale structure (7) is judged that no redundancy is generated, when the redundancy is generated, the system alarms and reports the type and the position of the redundancy, and the redundancy detection is continuously carried out after the picking-up is finished.

Software installed on the pc (5) processes collected sound signals, namely the microphone array and the collection board card (1) thereof are used for collecting sound signals generated when the large-scale structure rolls, and spatial filtering is carried out on the sound signals, so that signal-to-noise ratio preprocessing of the collected sound signals is improved; the time-frequency diagram of the sound signal received by the pc is obtained, the time-frequency diagram is compared with the time-frequency diagram of the normal sound signal, compared with the sound signal collected under the condition of no redundancy, the components of various redundancies above a specific frequency contain higher energy, the components are used as a basis for judging whether redundancy exists or not and the types of redundancy, and in addition, the kurtosis Q and the spectrum skewness P of the signal are calculated to be used as supplementary criteria.

The software installed on the pc (5) processes the collected sound signals, and specifically comprises the following steps:

firstly, the signal is framed by a window function g (tau) which is a finite duration function in the time domain, the window function g (tau-t) being shifted to be centered on t by a shift parameter t, x (tau) g (tau-t) reflecting, for a given signal x (tau), that the signal is centered on a reference time t, in an interval

A change in range;

filtering each section of signal by using a high-pass filter with the cut-off frequency h1 ═ 500hz, and taking the obtained signal as an original signal A (i) to be analyzed and processed; filtering with a high-pass filter with a higher cut-off frequency h2 ═ 3000hz, and obtaining a signal denoted as b (i), where i ═ 1,2,3,4, and represents the kind of redundancy, and respectively: 1-no excess; 2, a nut; 3, a bolt; 4-copper wire; therefore, the presence or absence of excess and the type of excess can be determined by using an index S (i) of the energy synthesis of B and a, where S (i) is E (B (i))/E (a (i)), E (x) is the energy value of the signal x, S (t, i) is obtained by taking the synthesis index of x (τ) g (τ -t) with t as a parameter, and the formula is as follows:

S(t,i)＝E(h₂x(i,τ)g^*(τ-t))/E(h₁x(i,τ)g^*(τ-t)) (3)

in the formula, S (t, i) represents the energy ratio of the ith type redundancy signal after windowing, x (i, tau) is the ith type redundancy signal, h₂And h₁Respectively, a high-pass filter;

and finally, judging the type of the redundancy according to S (t, i), the kurtosis Q and the spectrum skewness P:

when the remainder is a copper wire, the number of intervals of S (2)/S (1) >1 is 60% or more and the number of intervals of S (2)/S (1) >2 is between 14% and 15.5%, there being substantially no interval of S (2)/S (1) > 6.7; the interval Q >3.5 and P >100 are below 5%, similar to the case without excess;

when the remainder is a nut, the number of intervals of S (2)/S (1) >2 is between 27.5% and 29%, the number of intervals of S (2)/S (1) >6.7 is around 10%, and continuous intervals of S (2)/S (1) >6.7 exist; the interval of Q >3.5 is about 20%, the interval of P >100 is about 10%, and continuous intervals exist;

when the redundancy is a bolt, the number of intervals of S (2)/S (1) >6.7 is more than 50 percent, and continuous intervals of S (2)/S (1) >6.7 exist; k >3.5 is greater than 80%, Q >100 is 100% and there are continuous intervals.

Compared with the prior art, the invention can realize the rotation of a large-scale structure; secondly, sound signals can be collected in a non-contact mode, and the detection of movable redundant objects in the rotation process of the object to be detected is realized; meanwhile, the size and the type of the redundancy can be comprehensively, comprehensively and accurately detected by introducing algorithm analysis.

Drawings

Fig. 1 and 2 are schematic views of equipment installation used in an embodiment of the detection method of the present invention.

FIG. 3 is a flow chart of the detection method of the present invention.

Fig. 4 is a diagnostic flow of the algorithm installed on the PC.

The labels in the figure are: the microphone array and the acquisition board card are 1, the data transmission line is 2, the usb serial port line is 3, the rolling ring is 4, the PC is 5, the car is erected 6, the large-scale cylindrical structure is 7, the motor is 8, and the motor control box is 9.

Detailed Description

The present invention will be described in detail with reference to the accompanying drawings.

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

The first embodiment is as follows:

a method for detecting movable redundant substances of a large-scale cylindrical structure comprises the following steps:

step one, the large structure 7 is placed on a trolley and clamped by using the rolling ring 4, the rolling ring 4 is driven by the motor 8 to drive the large structure 7 to rotate around the axis of the rolling ring 4, and the rotating speed of the motor 8 is controlled by the motor control box 9.

And step two, placing the microphone array and the acquisition board card 1 thereof as shown in fig. 1, and connecting the microphone array and the acquisition board card 1 thereof with the pc 5 by using a data transmission line 2 and a usb serial port line 3 to form a sound signal acquisition system.

Step three, opening the pc to set acquisition parameters of the sound signal acquisition system; when the motor control box 9 is opened, the motor 8 is driven to drive the rolling ring 4 to drive the large-scale structure 7 to rotate around the rolling ring 4, and meanwhile, the microphone array and the acquisition board card 1 thereof transmit acquired sound signals to the pc 5; and software installed on the pc 5 processes the acquired sound signals.

The sound signal processing and analyzing flow is as shown in fig. 4, a time-frequency diagram of the sound signal received by the pc is obtained, the time-frequency diagram is compared with a time-frequency diagram of a normal sound signal, and compared with a sound signal collected under the condition of no redundancy, components of various redundancies above 3000hz contain higher energy, which can be used as a basis for judging whether redundancy exists and the types of the redundancy. And in addition, the kurtosis Q and the spectrum skewness P of the signal are calculated as supplementary criteria.

We need to pre-process the signal in advance. In order to perform real-time signal analysis and diagnosis, a window function is used for framing the signal, the window function g (tau) is a finite duration function in the time domain, the window function g (tau-t) is moved to the center of t by a translation parameter t, and for a given signal x (tau), x (tau) g (tau-t) reflects the signal to be centered at a reference time t and in an interval

A change in range.

Filtering each section of signal by using a high-pass filter with the cut-off frequency of 500Hz, and taking the obtained signal as an original signal A (i) to be analyzed and processed; filtering by using a high-pass filter with the cut-off frequency of 3000Hz, and recording the obtained signal as B (i), wherein i is 1,2,3 and 4, which represent the types of the surplus objects and are respectively 1-no surplus objects; 2, a nut; 3, a bolt; 4-copper wire. Therefore, the presence or absence of excess and the type of excess can be determined using an index s (i) obtained by synthesizing the energy of B and the energy of a, where s (i) is an energy value of the signal x, E (B (i))/E (a (i)), and E (x). Taking t as a parameter, and taking a synthetic index of x (tau) g (tau-t) to obtain S (t, i), wherein the formula is as follows:

S(t,i)＝E(h₃₀₀₀x(i,τ)g^*(τ-t))/E(h₅₀₀x(i,τ)g^*(τ-t))(3)

in the formula, S (t, i) represents the energy ratio of the ith type redundancy signal after windowing, x (i, tau) is the ith type redundancy signal, h₃₀₀₀And h₅₀₀High pass filters with cut-off frequencies of 3000hz and 500hz, respectively, are shown.

And finally, judging the type of the redundancy according to S (t, i), the kurtosis Q and the spectrum skewness P.

And when no surplus objects are generated, the system continuously collects the objects until the specified time is exceeded, the large-scale structure 7 is judged to have no surplus objects, when the surplus objects are generated, the system alarms and reports the types and positions of the surplus objects, and the surplus object detection is continuously carried out after the surplus objects are picked up.

Example two:

a method for detecting movable redundant substances of a large-scale cylindrical structure comprises the following steps:

step one, the large structure 7 is placed on a trolley and clamped by using the rolling ring 4, the rolling ring 4 is driven by the motor 8 to drive the large structure 7 to rotate around the axis of the rolling ring 4, and the rotating speed of the motor 8 is controlled by the motor control box 9.

And step two, placing the microphone array and the acquisition board card 1 thereof as shown in fig. 2, and connecting the microphone array and the acquisition board card 1 thereof with the pc 5 by using a data transmission line 2 and a usb serial port line 3 to form a sound signal acquisition system.

Turning on a pc to set a sound signal acquisition system and set acquisition parameters by referring to the step 3; when the motor control box 9 is opened, the motor 8 is driven to drive the rolling ring 4 to drive the large-scale structure 7 to rotate around the rolling ring 4, and meanwhile, the microphone array and the acquisition board card 1 thereof transmit acquired sound signals to the pc 5; and software installed on the pc 5 processes the acquired sound signals.

The sound signal processing and analyzing flow is as shown in fig. 4, a time-frequency diagram of the sound signal received by the pc is obtained, the time-frequency diagram is compared with a time-frequency diagram of a normal sound signal, and compared with a sound signal collected under the condition of no redundancy, components of various redundancies above 3000hz contain higher energy, which can be used as a basis for judging whether redundancy exists and the types of the redundancy. And in addition, the kurtosis Q and the spectrum skewness P of the signal are calculated as supplementary criteria.

We need to pre-process the signal in advance. In order to perform real-time signal analysis and diagnosis, a window function is used for framing the signal, the window function g (tau) is a finite duration function in the time domain, the window function g (tau-t) is moved to the center of t by a translation parameter t, and for a given signal x (tau), x (tau) g (tau-t) reflects the signal to be centered at a reference time t and in an interval

A change in range.

Filtering each section of signal by using a high-pass filter with the cut-off frequency of 500Hz, and taking the obtained signal as an original signal A (i) to be analyzed and processed; filtering by using a high-pass filter with the cut-off frequency of 3000Hz, and recording the obtained signal as B (i), wherein i is 1,2,3 and 4, which represent the types of the surplus objects and are respectively 1-no surplus objects; 2, a nut; 3, a bolt; 4-copper wire. Therefore, the presence or absence of excess and the type of excess can be determined using an index s (i) obtained by synthesizing the energy of B and the energy of a, where s (i) is an energy value of the signal x, E (B (i))/E (a (i)), and E (x). Taking t as a parameter, and taking a synthetic index of x (tau) g (tau-t) to obtain S (t, i), wherein the formula is as follows:

S(t,i)＝E(h₃₀₀₀x(i,τ)g^*(τ-t))/E(h₅₀₀x(i,τ)g^*(τ-t))(3)

in the formula, S (t, i) represents the energy ratio of the ith type redundancy signal after windowing, x (i, tau) is the ith type redundancy signal, h₃₀₀₀And h₅₀₀High pass filters with cut-off frequencies of 3000hz and 500hz, respectively, are shown.

And finally, judging the type of the redundancy according to S (t, i), the kurtosis Q and the spectrum skewness P.

And when no surplus objects are generated, the system continuously collects the objects until the specified time is exceeded, the large-scale structure 7 is judged to have no surplus objects, when the surplus objects are generated, the system alarms and reports the types and positions of the surplus objects, and the surplus object detection is continuously carried out after the surplus objects are picked up.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.

And (3) displaying the result:

filtering each section of signal by using a high-pass filter with the cut-off frequency of 500Hz, and taking the obtained signal as an original signal A (i) to be analyzed and processed; filtering with a high-pass filter with a cut-off frequency of 3000Hz, and recording the obtained signal as b (i), wherein i is 1,2,3,4, which represents the types of the redundancy, and respectively: 1-no excess; 2, a nut; 3, a bolt; 4-copper wire. Therefore, the presence or absence of excess and the type of excess can be determined using the index S (i) of the energy synthesis of B and A.

The signal is segmented according to 10000 points in length, the energy ratio S (i), the kurtosis Q and the frequency spectrum skewness P are respectively calculated for each segment of signal, and the obtained results are shown in table 1:

TABLE 1 results of identifying the type of redundancy

As can be seen from Table 1, when the excess is a copper wire, the number of intervals of S (2)/S (1) >1 is 60% or more and the number of intervals of S (2)/S (1) >2 is between 14% and 15.5%, there being substantially no interval of S (2)/S (1) > 6.7; the interval Q >3.5 and P >100 are below 5%, similar to the case without redundancy.

When the remainder is a nut, the number of intervals of S (2)/S (1) >2 is between 27.5% and 29%, the number of intervals of S (2)/S (1) >6.7 is around 10%, and continuous intervals of S (2)/S (1) >6.7 exist; the interval of Q >3.5 is about 20%, and the interval of P >100 is about 10% and continuous intervals exist.

When the redundancy is a bolt, the number of intervals of S (2)/S (1) >6.7 is more than 50 percent, and continuous intervals of S (2)/S (1) >6.7 exist; k >3.5 is greater than 80%, Q >100 is 100% and there are continuous intervals.

Claims (1)

1. A method for detecting the activity redundancy of a large structure based on a sound signal is characterized by comprising the following steps:

the method comprises the following steps that firstly, a large-scale structure (7) is placed on a trolley and clamped by using a rolling ring (4), the rolling ring (4) is driven by a motor (8) to drive the large-scale structure (7) to rotate around the axis of the rolling ring (4), and the rotating speed of the motor (8) is controlled by a control cabinet (9);

secondly, placing the microphone array and the acquisition board card (1) thereof on a large structure (7), and connecting the microphone array and the acquisition board card (1) thereof with a pc (5) by using a data transmission line (2) and a usb serial port line (3) to form a sound signal acquisition system;

step three, opening the pc (5) to set acquisition parameters of the sound signal acquisition system; when a motor (8) is turned on through a control cabinet (9) to drive a rolling ring (4) to drive a large-scale structure (7) to rotate around the rolling ring (4), a microphone array and an acquisition board card (1) thereof transmit acquired sound signals to a pc (5); software installed on the pc (5) processes the acquired sound signals, when no redundancy is generated, the system continuously acquires the sound signals until the specified time is exceeded, the large-scale structure (7) is judged that no redundancy is generated, when the redundancy is generated, the system alarms and reports the type and the position of the redundancy, and the redundancy detection is continuously performed after the picking-up is finished;

software installed on the pc (5) processes collected sound signals, namely the microphone array and the collection board card (1) thereof are used for collecting sound signals generated when the large-scale structure rolls, and spatial filtering is carried out on the sound signals, so that signal-to-noise ratio preprocessing of the collected sound signals is improved; calculating a time-frequency diagram of the sound signal received by the pc, comparing the time-frequency diagram with a time-frequency diagram of a normal sound signal, wherein compared with the sound signal collected under the condition of no redundancy, components of various redundancies above a specific frequency contain higher energy to be used as a basis for judging whether redundancy exists or not and the types of the redundancy, and in addition, the kurtosis Q and the spectrum skewness P of the signal are calculated to be used as supplementary criteria;

the software installed on the pc (5) processes the collected sound signals, and specifically comprises the following steps:

firstly, the signal is framed by a window function g (tau) which is a finite duration function in the time domain, the window function g (tau-t) being shifted to be centered on t by a shift parameter t, x (tau) g (tau-t) reflecting, for a given signal x (tau), that the signal is centered on a reference time t, in an interval

A change in range;

filtering each section of signal by using a high-pass filter with the cut-off frequency h1 ═ 500hz, and taking the obtained signal as an original signal A (i) to be analyzed and processed; filtering with a high-pass filter with a higher cut-off frequency h2 ═ 3000hz, and obtaining a signal denoted as b (i), where i ═ 1,2,3,4, and represents the kind of redundancy, and respectively: 1-no excess; 2, a nut; 3, a bolt; 4-copper wire; therefore, the existence of excess and the type of excess can be judged by using an index S (i) of the energy synthesis of B and A, wherein S (i) is E (B (i))/E (A (i)), E (x) is the energy value of a signal x, S (t, i) is obtained by taking the synthesis index of x (tau) g (tau-t) with t as a parameter, and the formula is as follows:

S(t，i)＝E(h₂x(i，τ)g^*(τ-t))/E(h₁x(i，τ)g^*(τ-t)) (3)

in the formula, S (t, i) represents the energy ratio of the ith type redundancy signal after windowing, x (i, tau) is the ith type redundancy signal, h₂And h₁Respectively, a high-pass filter;

finally, judging the type of the redundancy according to S (t, i), the kurtosis Q and the spectrum skewness P:

when the remainder is a copper wire, the number of intervals of S (2)/S (1) >1 is 60% or more and the number of intervals of S (2)/S (1) >2 is between 14% and 15.5%, there being substantially no interval of S (2)/S (1) > 6.7; the interval Q >3.5 and P >100 are below 5%, similar to the case without excess;

when the redundancy is a nut, the number of intervals of S (2)/S (1) >2 is between 27.5% and 29%, the number of intervals of S (2)/S (1) >6.7 is about 10%, and continuous intervals of S (2)/S (1) >6.7 exist; the interval of Q >3.5 is about 20%, the interval of P >100 is about 10%, and continuous intervals exist;

when the redundancy is a bolt, the number of intervals of S (2)/S (1) >6.7 is more than 50 percent, and continuous intervals of S (2)/S (1) >6.7 exist; k >3.5 is greater than 80%, Q >100 is 100% and there are continuous intervals.

Images