CN1267579A - Sound signal sensing method and system for the perforative state of plasma arc welding bath - Google Patents

Abstract

The invention belongs to the field of plasma arc welding quality control. In the present invention, by setting up a microphone near the plasma welding torch, the sound signal s(t) in the plasma arc welding process is collected and stored in the computer in real time by using a data collection system; Quantitative sensing signal A _s of whether the molten pool is perforated or not; Finally, compare A _s with a given threshold to judge the perforation state of the molten pool at that time. This method can effectively sense whether the molten pool is perforated, has the advantages of simple system structure and low cost, and is easy to be popularized in production practice.

Description

Acoustic signal sensing method and system for plasma arc welding molten pool perforation state

The invention belongs to the technical field of plasma arc welding quality control.

In the high-energy beam processing technology known as "the processing technology of the 21st century", the plasma arc has the characteristics of simple equipment and low cost. In industrial production, especially in aerospace manufacturing, it is widely used in welding and easy oxidation. Aluminum, copper and other non-ferrous metals and their alloys, stainless steel, high-temperature alloys, titanium and titanium alloys, and refractory active metals. Due to the high energy density and strong penetrating power of the plasma arc, the arc center diameter of the arc is thin, the arc center and the arc flame of the plasma arc heat the workpiece 10 at the same time during welding, and the arc can dig deep into the molten pool to excavate the liquid metal, thereby causing A penetrating small hole 11, the arc passes through the small hole to form a tail flame 12 on the back of the weld, and then as the welding torch 14 moves forward, the liquid metal in the molten pool relies on the support of surface tension, along the direction of the molten pool The solid wall flows towards the tail of the molten pool, cools and crystallizes to form a weld 13, as shown in FIG. 1 . It can be seen that the stability of the small hole state in the plasma arc welding process is the key to ensure the complete penetration and forming quality of the weld. However, the fluctuation of various specification conditions in actual production is inevitable, which will inevitably affect the stability of the "small hole effect". For example, the small hole is closed during the welding process, resulting in defects such as incomplete penetration and seriously affecting the quality of welded products. Therefore, the research on sensing technology for automatic and real-time detection of "small hole" state is proposed, so as to lay the foundation for the control of the penetration quality of plasma arc welding seam.

Judging from the current situation of plasma welding production and research at home and abroad, there are several methods for detecting the state of the small holes in the molten pool, as shown in Figure 2. The detection methods on the back of the workpiece include: the back light signal measurement method, the back arc tail flame conduction method and the back sound signal detection method. The backside optical signal detection method is shown in FIG. 2 a . This method uses various photoelectric devices 201 arranged on the backside of the workpiece 200 to detect the intensity of the radiated light on the backside of the weld, so as to know whether the molten pool is perforated. The back plasma flame conduction method is shown in Figure 2b. In this method, a copper plate 203 is placed on the back of the workpiece 204 and kept at a certain distance from the workpiece. When the molten pool is perforated, the plasma flame 205 passing through the small hole reaches the copper plate 203. On the other hand, due to the conductive characteristics of the plasma, the voltage can be obtained on the resistor R after the external power supply is applied, and it is rectified to output U _o , and the size of U _o has a certain relationship with the size of the small hole. The rear sound signal detection method, as shown in Figure 2c, is mainly used for welded pipes. The microphone 206 is placed on one side of the pipe to detect the sound signal inside the pipe, and judge whether there is a small hole according to the change of the sound signal. The common feature of the above various detection methods is that the sensor is placed on the back of the workpiece, so its application to some complex structure welding is limited. The methods for detecting from the front of the workpiece include spectral analysis and CCD imaging. The spectral analysis method is shown in FIG. 2d. In this method, the optical signal obtained by the area array CCD is transmitted to the spectrum analyzer through the optical fiber 209, and the spectral line intensity of hydrogen relative to argon in the arc signal 211 on the front of the workpiece 210 is detected by the spectrum analyzer. To judge whether there are small holes in the molten pool. The CCD imaging method is shown in Figure 2e. The image of the small hole is directly observed through the CCD. However, due to space constraints, the image of the small hole in the molten pool captured by the CCD on the front of the workpiece is distorted, which will bring trouble to the subsequent processing. These two methods are still in the research stage due to complex structures, expensive equipment or complicated processing methods, and there are certain difficulties in their application in production.

The present invention aims to overcome the deficiencies of the prior art. By using the sound signal in the welding process, it proposes a sound sensing method and device for the perforation state of the plasma arc welding molten pool, in order to realize the automatic penetration of the plasma arc welding of medium and thick plates. Controls provide the necessary feedback information. Because the sound signal is used as the sensing signal source, it has the advantages of simple system structure and low cost, and is easy to be popularized in production practice.

A sound signal detection method for the perforation state of the plasma arc welding pool proposed by the present invention is characterized in that it comprises the following steps:

1) Use the sound microphone to obtain the sound signal s(t) during the plasma arc welding process, and collect the sound signal s(t) in real time through the signal acquisition system and store it in the computer;

2) Process the collected sound signal sequence s(n) with the A algorithm to obtain a quantitative sensing signal A _s of whether the molten pool is perforated or not, which can be recognized by the machine;

3) Compare A _s with the threshold value of whether the molten pool is perforated or not, so as to judge the perforated state of the molten pool at that time.

The A algorithm for processing the sound signal includes the following steps:

1) Design and analyze the rectangular window function g(n-m), and perform discrete windowed Fourier transform (DWFT) on the sound signal sequence s(n) obtained by the computer.

(式中f₁，f₂——声音信号的频率，f₁＝0Hz，2) For the discrete windowed Fourier transform result DWFT _s (m, k), do the following:

(where f ₁ , f ₂ - the frequency of the sound signal, f ₁ = 0Hz,

f ₂ =100Hz), A(m) is obtained, and the sound sensing signal A _s is obtained after smoothing.

In the above-mentioned sound signal detection method, the criterion for judging whether the molten pool is perforated or not is as follows: setting the threshold of whether the molten pool is perforated or not and the critical value of the transition state of the molten pool, when the A _s signal amplitude is higher than the set When the critical value of the molten pool transition state is reached, the molten pool is considered to be in the transition stage from non-perforation to perforation or from perforation to non-perforation; when the amplitude of the A _s signal decreases and is lower than the threshold of whether the molten pool is perforated or not, it can be confirmed that the molten pool has been perforated. The normal perforation plasma arc welding is realized; when the A _s signal amplitude is lower than the critical value of the transition state of the molten pool and higher than the threshold of whether the molten pool is perforated or not, the molten pool is judged to be in the non-perforated state.

The principle of the present invention is as follows: sound wave is a kind of mechanical wave, which is generated by the vibration of solid or the pulsation of fluid. During the welding process, the welding power supplies arc energy, which acts on the molten pool through the arc, and part of the energy will be expressed in the form of molten pool vibration and plasma airflow pulsation. When the state of the molten pool changes during the welding process, the vibration of the molten pool Conditions change accordingly, resulting in changes in the vibration of the molten pool and the pulsation characteristics of the plasma gas flow. This change will also be reflected in the sound signal, so it is possible to find information reflecting the state of the molten pool through the analysis and processing of the sound signal Come.

The realization principle of the method of the present invention is shown in FIG. 3 . Firstly, the high-precision microphone is used to obtain the sound signal s(t) during the plasma arc welding process, and the sound signal s(t) is collected and stored in the computer in real time through the signal acquisition system. The sound signal itself cannot be used as a sensing signal to directly determine whether there is a small hole in the molten pool, but the signal carries characteristic information that can reflect the perforated and non-perforated state of the molten pool. The characteristic information is extracted through the developed sensing signal extraction algorithm (Algorithm A) processing, the quantitative molten pool perforation state sensing signal A _s that can be recognized by the machine is obtained. Further combined with the designed criterion for judging whether the molten pool is perforated or not, the A _s signal can be used to detect whether there are small holes in the molten pool. The sensing signal extraction algorithm and the threshold judgment method are realized by pre-programmed software stored in the computer.

The sound signal acquisition system in the plasma arc welding process of the present invention is made up of sound sensor, data acquisition card, computer and relevant interface, and is stored in the corresponding sound signal acquisition processing software in the computer, as shown in Figure 4, among the figure, sound The signal is input to the input terminal of the data acquisition card through the sound sensor, and the data acquisition card is directly inserted into the corresponding slot of the computer. Wherein, the sound signal is directly acquired by the microphone 43 placed near the welding torch, and the data acquisition card is combined with the data acquisition application program developed by the present invention to complete signal acquisition and analog-to-digital conversion, and utilize the interface between the acquisition card and the computer to store the signal in the computer .

The algorithm and corresponding signal extraction method for extracting molten pool perforation state sensing signal from the characteristic information carried by the sound signal according to the present invention are as follows:

1) Design a reasonable analytical rectangular window function g(nm): In the formula, i——analyze the moving step of the rectangular window function, which can be taken as an integer according to the actual situation

N - the number of Fourier transform points

n, m——Natural numbers, n=0, 1, 2,..., m=0, 1,...

2) Perform discrete windowed Fourier transform (DWFT) on the sound signal sequence s(n) obtained by the computer: $\mathrm{DWF}{T}_{\mathrm{the\; s}}(m,k)=\underset{\mathrm{no}=0}{\overset{\∞}{\mathrm{\Σ}}}\mathrm{the\; s}\left(\mathrm{no}\right)g(\mathrm{no}-m)e^{-j\frac{2\mathrm{\π}}{N}\mathrm{nk}}......\left(2\right)$

In the formula, k—k takes 0, 1, ..., N-1

3) For the discrete windowed Fourier transform result DWFT _s (m, k), do the following: $A\left(m\right)=\underset{k=f_1{T}_{\mathrm{the\; s}}N}{\overset{f_2{T}_{\mathrm{the\; s}}N}{\mathrm{\Σ}}}\left|\mathrm{DWF}{T}_{\mathrm{the\; s}}(m,k)\right|......\left(3\right)$

In the formula, f ₁ , f ₂ —— the frequency of the sound signal, f ₁ = 0Hz, f ₂ = 100Hz, get A(m), and smooth it

After processing, the sound sensing signal A _s is obtained.

The criterion for judging the perforation state threshold of the molten pool designed by the present invention is shown in Fig. 5. The abscissa in the figure is time, and the ordinate is the amplitude of the A _s signal. The variation of the A _s signal with time is shown in curve C. It can be seen from the figure that the A _s signal can reflect the state of the molten pool: when the molten pool is in the perforation stage, the signal amplitude is the lowest, as shown in the BD section on the time axis; when the molten pool is not perforated, the signal amplitude is higher, It is shown before point A or after point E on the time axis; and when the melt pool is in the transitional stage of perforation, the signal amplitude is the highest, as shown in section AB or section DE on the time axis. The process of judging whether the molten pool is perforated or not by the threshold method is as follows: when the A _s signal amplitude is higher than the set critical value _T1 of the molten pool transition state, the molten pool is considered to be in the transition stage from no perforation to perforation or from perforation to no perforation ; When the A _s signal amplitude is lower than the threshold _T2 of whether the molten pool is _perforated or not, it is confirmed that the molten pool has been perforated, and the normal perforation type plasma arc welding has been realized; When the critical value T ₁ of the molten pool is higher than the threshold T ₂ of whether the molten pool is perforated or not, it is determined that the molten pool is not perforated.

The detection method of the invention can effectively sense the state of whether the molten pool is perforated, and can provide necessary feedback information for realizing the automatic penetration control of the plasma arc welding of the medium-thick plate. Since the sound in the welding process is used as the sensing signal source, it has the advantages of simple system structure and low cost compared with other detection methods, and is easy to promote in production practice

Brief description of the drawings:

Figure 1 is a schematic diagram of the perforated plasma arc welding process.

Fig. 2 is a schematic diagram of several methods for detecting the perforation state of the molten pool in the prior art; wherein,

Figure 2a is the back light signal measurement method,

Figure 2b is the back plasma flame current conduction method,

Figure 2c is the detection method of the back sound signal,

Figure 2d is the spectral analysis method,

Figure 2e shows the CCD imaging method.

Fig. 3 is a functional block diagram of the method of the present invention.

Fig. 4 is a schematic diagram of the composition of the sound signal acquisition system of the present invention.

Fig. 5 is a schematic diagram of the criterion for judging whether the molten pool is perforated or not according to the present invention.

FIG. 6 is a software flow diagram for implementing the A algorithm in this embodiment.

FIG. 7 is a schematic diagram of the composition of the sound signal acquisition system of this embodiment.

Fig. 8a is a schematic diagram of the structure of the back grooved workpiece of this embodiment,

Figure 8b is a diagram showing the detection results of the perforation state of the molten pool in this embodiment

Fig. 8c is a curve diagram of the sensing signal A _s obtained in this embodiment,

A sound signal sensing method for the perforation state of the plasma arc welding pool designed by the present invention and its system embodiment are described in detail in conjunction with the accompanying drawings as follows:

Considering changes in heat dissipation conditions, workpiece thickness, and welding specifications in actual production, this embodiment adopts back-grooved workpieces, and the set welding specification parameters ensure that the welding arc does not perforate the molten pool at the workpiece thickness of 6 mm, and melts at the 4 mm. Pool piercings were welded on stainless steel workpieces using an inverter welder.

The sound signal acquisition system of this embodiment is composed of a sound sensor, a data acquisition card, a computer and related interfaces, as shown in FIG. 7 . The sound signal is acquired by the microphone 71 placed near the welding torch, and then input to the input terminal of the data acquisition card, and the data acquisition card is directly inserted into the corresponding slot of the computer. The microphone adopts high-precision sound pressure sensor JL-1, and the response frequency bandwidth is up to 20kHz. The voltage output by the microphone has a relatively strict linear relationship with the sound pressure of the sound signal, which ensures that the voltage sent to the data acquisition card can accurately reflect the welding sound pressure in the process. The data acquisition card model is "PCI-9118HG", the precision is 12 bits, the highest acquisition speed is 330kHz, and the continuous acquisition data capacity can reach 64M each time. This card combines the data acquisition application program developed by the present invention to complete signal acquisition, analog-to-digital conversion, and The signal is stored in the computer by using the interface between the acquisition card and the computer. The computer uses PII266, memory 128M, and is based on the Window98 operating platform.

The sound signal sensing method for the perforation state of the plasma arc welding molten pool in this embodiment is as follows: first, the above-mentioned signal acquisition system is used to collect the sound signal during the welding process in real time at a sampling frequency of 50 kHz, and input it into the computer. Subsequently, the sound signal A algorithm designed by the present invention is used to process the data stored in the computer, and Fig. 6 is a flow chart of implementing the algorithm with software.

；对s(n)作加窗傅里叶变换 ；对DWFT_s(m，k)作处理判断m≥M，若为假则m＝m+1，重复上述运算；若为真，则退出循环处理。对A(m)进行平滑处理，即得A_s(m)信号，如图8c所示，图8c中横坐标为时间，单位为s，纵坐标为A_s信号的幅值。The algorithm flow is as follows: read in the collected sound signal sequence s(n); take the analysis rectangular window function moving step i=1024, divide s(n) into segments M=s(n)/i-1; give m, N assignment (m=1, N=4096). Enter the loop to calculate the analysis window function

;do windowed Fourier transform on s(n) ;Process DWFT _s (m,k) It is judged that m≥M, if it is false, then m=m+1, repeat the above operation; if it is true, then exit the loop processing. A(m) is smoothed to obtain the A _s (m) signal, as shown in Figure 8c, where the abscissa in Figure 8c is time in s, and the ordinate is the amplitude of the A _s signal.

Table 1 lists the threshold value of the molten pool perforation and the critical value of the molten pool transition state of the A _s signal when welding stainless steel under different welding power conditions.

Table 1 Sensing signal Threshold T ₂ for whether the molten pool is perforated or not Critical value T ₁ of molten pool transition state Inverter welding machine Magnetic Amplifier Welding Machine Inverter welding machine Magnetic Amplifier Welding Machine _As 25～38 25～35 80 70

_After processing the sound signal with the A algorithm to obtain the sensing signal A _s , find out from Table 1 the threshold value T ₂ (25-38) and The critical value T ₁ of the molten pool transition state (80), according to the threshold judgment criterion of the molten pool perforation state, when the A _s signal amplitude is higher than the set critical value T ₁ of the molten pool transition state = 80, the molten pool is considered It is in the transition stage from non-perforation to perforation or from perforation to non-perforation; when the A _s signal amplitude is lower than the threshold T ₂ =25-38 for whether the molten pool is perforated or not, it is confirmed that the molten pool has been perforated and normal plasma has been realized arc welding; and when the amplitude of the A _s signal is lower than the critical value T ₁ =80 for the transition state of the molten pool and higher than the threshold T ₂ =25-38 for whether the molten pool is perforated or not, it is determined that the molten pool is in a non-perforated state. The results of using the A _s signal to detect the perforation state of the molten pool are marked in Fig. 8b, where the AB segment is the unperforated stage, the BC and DE segments are the transition stage, and the CD segment is the perforated stage. It can be seen that the detection results are consistent with the display of the structure of the back grooved workpiece on the perforation state of the molten pool. In addition, in order to provide a synchronous reference signal for the processing of the sound signal, a phototransistor 70 (model 3DU33) is mounted on the back of the weld seam as shown in FIG. 7 . During the welding process, the phototransistor and the welding torch move synchronously. The output voltage signal and the sound signal are sampled at a rate of 50kHz and sent to the computer. When the perforation state of the molten pool changes, the output signal of the phototransistor will have a step change. Comparing the processing result of A algorithm with the output voltage signal of the phototransistor, the coincidence effect is good. This shows that the detection technology can accurately sense the perforation and non-perforation states of the molten pool.

Table 1 shows that both the type of welding machine and the parameters of the welding specification will lead to differences in the threshold of the small hole sensing signal. Therefore, in actual use, it is necessary to select an appropriate sound signal judgment threshold according to the specific welding implementation conditions.

Claims (4)

1. A sound signal detection method of a plasma welding pool perforation state, characterized in that, comprising the following steps: 1) Use the sound microphone to obtain the sound signal s(t) during the plasma arc welding process, and collect the sound signal s(t) in real time through the signal acquisition system and store it in the computer; 2) Process the collected sound signal sequence s(n) with the A algorithm to obtain a quantitative sensing signal A _s of whether the molten pool is perforated or not, which can be recognized by the machine; 3) Compare A _s with the threshold value of whether the molten pool is perforated or not, so as to judge the perforated state of the molten pool at that time. 2. sound signal detection method as claimed in claim 1 is characterized in that, said A algorithm that processing sound signal is processed comprises the following steps: 1) Design and analyze the rectangular window function g(n-m), and perform discrete windowed Fourier transform (DWFT) on the sound signal sequence s(n) obtained by the computer; 2) For the discrete windowed Fourier transform result DWFT _s (m, k), do the following: In the formula, f ₁ , f ₂ —— frequency of sound signal, f ₁ = 0Hz, f ₂ =100Hz, A(m) is obtained, and the sound sensing signal A _s is obtained after smoothing. 3. The sound signal detection method as claimed in claim 1, characterized in that, the criterion for judging whether the molten pool is perforated or not is as follows: setting the threshold of whether the molten pool is perforated and the critical value of the molten pool transition state, when When the amplitude of the A _s signal is higher than the set critical value of the molten pool transition state, the molten pool is considered to be in the transition stage _from non-perforated to perforated or perforated to non-perforated; threshold, it can be confirmed that the molten pool has been perforated, that is, normal perforated plasma arc welding; and when the A _s signal amplitude is lower than the critical value of the molten pool transition state and higher than the threshold of whether the molten pool is perforated or not, it is determined that The molten pool is in an unperforated state. 4. a plasma arc welding sound signal acquisition system in the method as claimed in claim 1, is characterized in that, by sound sensor, data acquisition card, computer and relevant interface, and be stored in corresponding sound signal acquisition processing software in the computer Composition, wherein the sound signal is first introduced into the input end of the sound sensor, the output end of the sound sensor is connected to the input terminal of the data acquisition card, and the data acquisition card is directly inserted into the corresponding slot of the computer.

Images