CN102749495A - Quantitative assessment method for stability of double-wire arc welding electrical signal - Google Patents

Abstract

The invention discloses a quantitative assessment method for the stability of a double-wire arc welding electrical signal. The specific method comprises the following steps of: (1) reading a current signal; (2) preprocessing collected data; (3) calculating a sample entropy sequence value SampEn(i) of an actually-measured signal under corresponding given parameters; (4) calculating a sample entropy average value SaEn of the actually-measured signal; (5) calculating a sample entropy value RSaEn of the actually-measured signal under the corresponding given parameters; (6) calculating a one-way stability assessment index CSaEn; and (7) calculating a double-wire electrical signal sample entropy index TCSaEn. According to the method provided by the invention, the sample entropies are introduced to a welding process stability assessment system for the first time, the influence of various pulse parameters on the sample entropy values is taken into comprehensive consideration, sample entropy-based welding process stability assessment indexes are self-designed, and a dedicated double-wire welding process stability quantitative assessment index is designed for the double-wire welding process. By using the method provided by the invention, the double-wire arc welding process stability degree can be well quantized.

Description

A kind of mariages arc-welding electric signal stability method for quantitatively evaluating

Technical field

The invention belongs to high-speed double-wire welding technology field, particularly a kind of method for quantitatively evaluating of mariages arc welding process current signal stability.

Background technology

The mariages pulse MIC welding is one and becomes when non-linear, strong and the process of strong coupling that unstable mechanism is complicated, and influence factor is numerous, does not also have complete theoretical explanation so far, for welding effect or the quality also unified comprehensive evaluation criterion of neither one at present.Generally adopt, qualitative evaluation mode such as bad, stable, instability.The shortage of quantitative evaluation standard has been brought control method research and shop characteristic difficulty analytically.

The signal that is used to detect research at present has electric arc sound, ultrasound wave, purplish red external radiation, visible light, electric signal and machine vision signal etc., wherein utilizes electric signal to carry out that welding process stability analysis cost is lower, effect is better.Electric signal has comprised the information of a large amount of reflection welding technological properties and welding quality, and the stability of electric signal is the key factor of restriction shop characteristic.The electric signal analysis generally is divided into temporal pattern analysis and statistical study again.The temporal pattern analysis generally is the angle from time domain and amplitude domain, the stability of time-dependent current voltage when analyzing in the welding process.Statistical analysis technique is at the amplitude domain of signal and frequency domain signal to be analyzed; It can reflect the stability of signal macroscopic view; But lost the varying information of signal; Can not the time dependent characteristic of reflected signal, so the analysis of welding process electric signal can not rely on statistical study fully, must be used in combination with other signal processing method.

South China Science & Engineering University utilizes electric signal that welding process has been carried out more research.People such as Xue Jiaxiang through information analyses such as electric current and voltage probability density distributions, shorted period time CO ₂The stability of weldering; Utilize FFT that electric signal is carried out spectrum analysis, proposition can be from the thinking of frequency angle analysis welding stability; The electric signal Virtual Analysis system that utilized industrial computer, data collecting card and VC platform development realizes the transient analysis and the statistical study of current/voltage.People such as Cao Biao, Lv Xiaoqing is to CO ₂The approximate entropy that carries out that welds current signal under different wires feed rate, voltage, the gas flow condition calculates and analyzes; The size of finding approximate entropy is not only relevant with molten drop transition frequency; Also the stability of short circuit transition welding termination process has contact more closely; Think that the stable more approximate entropy of welding process is big more, approximate entropy can as the short circuiting transfer estimation of stability a kind of criterion.People such as Nie Jing have set up aluminum Alloys Pulsed MIG Welding Based arc voltage approximate entropy generalized regression nerve networks forecast model, in order to the stability of measurement welding process, and think that the big welding process of approximate entropy is unstable, and the approximate little welding process of entropy is stable.

From above-mentioned research and method for estimating stability, at present to the evaluation many places of electric signal in the qualitative analysis stage, mainly concentrate in the statistical study of electric signal, the instantaneous state analysis is because record data difficulty comparatively, so present research is few.Entropy theory is applied to the research of welding process is at present domestic has only a few team, these a spot of researchs also concentrate on the stability of utilizing early stage approximate entropy theory to represent welding process.Research is also incomplete and ripe, mainly is to be in the qualitative analysis stage.Because approximate entropy receives data length and embeds dimension and influence greatlyyer, consistance is relatively poor relatively, Richman in 2000 on the basis of approximate entropy, propose a kind of new time series complicacy estimate method-Sample Entropy (Sample Entropy, SaEn).Sample Entropy and approximate entropy are similar, all are a kind of non-linear dynamic mathematic(al) parameters of measuring sequence complexity and statistic, represent certain seasonal effect in time series complicacy with a nonnegative number.Difference is that Sample Entropy is disregarded self matching value, has reduced the error of approximate entropy, and consistance and precision are better, and computing velocity is also than comparatively fast.

With regard to the domestic document of publishing at present, utilize Sample Entropy that the research that mariages pulse MIC welding technology stability carries out quantitative evaluation is not appeared in the newspapers as yet.

Consider that mariages arc-welding current signal is not only the simple superposition of monofilament signal, also considered contact each other and influence each other, this stability analysis to mariages arc-welding current signal has brought difficulty, and its quantitative test is also more complicated.Therefore study mariages arc-welding current signal method for quantitatively evaluating, can effectively instruct choosing of welding condition, for the process intelligent optimization and the Online Monitoring Control of welding are prepared.

Summary of the invention

The objective of the invention is to estimate difficulty to the Double Wire Welding technological specification; Be difficult to realize problems such as quantitative evaluation; A kind of electric signal Sample Entropy evaluation method that is applicable to the mariages arc welding process is provided; Designed and be used for the index that mariages arc-welding stability Sample Entropy is estimated, the quantitative evaluation of realization different process performance.

The object of the invention is realized through following technical scheme:

A kind of mariages arc-welding electric signal stability method for quantitatively evaluating is meant the calculated signals sample entropy after the denoising; Calculate the Double Wire Welding process stability evaluation index based on Sample Entropy of design voluntarily; Realize the quantitative evaluation of welding process stability, specifically comprise the steps:

(1) utilize the Wavelet analyzer of developing voluntarily to gather the current signal data of one period long period;

(2) data of being gathered are carried out pre-service, mainly comprise two steps, at first data are carried out wavelet packet filtering to reduce noise on analysis result's influence,, carried out the normalization processing secondly in order to prevent of the influence of electric signal size to the result;

(3) calculate Sample Entropy sequential value SampEn (i) under the corresponding given parameter of this measured signal;

(4) calculate the Sample Entropy average index S aEn that surveys electric signal;

(5) calculate sample entropy RSaEn under the corresponding given parameter of this measured signal;

(6) calculate single channel estimation of stability index CSaEn;

(7) calculate mariages electric signal Sample Entropy index TCSaEn.

The Sample Entropy algorithm of SampEn (i) in the said step (3) comprises:

Raw data is the sequence u (1) of N point, u (2) ... U (N), m is for embedding dimension.When embedding dimension when being m and m+1, preceding N-m vector of data sequence, needs are satisfied when 1≤i≤N-m, m dimensional vector X _m(i) meaningful.

(3.1) by one group of m n dimensional vector n of sequence number reconstruct

X _m(i)＝[u(i)，u(i+1)，…u(i+m-1)]i＝1，2，…，N-m+1 (1)

(3.2) definition X _m(i) and X _m(j) apart from d [X _m(i), X _m(j)] be of maximum difference in both corresponding elements:

$d[X_m\left(i\right),X_m\left(j\right)]=\underset{0<k<m-1}{\mathrm{<figure-callout\; id="0"\; label="max"\; filenames="HSA00000743223300021.png,HSA00000743223300022.png"\; state="\{\{state\}\}">max</figure-callout>}}\left(\right|u(i+k)-u(j+k)\left|\right)j\&NotEqual;i---\left(2\right)$

(3.3) for each i (1≤i≤N-m), add up d [X _m(i), X _m(j)] less than the number N of given threshold value r ^m(i), ask its with the distance total N-m ratio.

$B_r^m\left(i\right)=N^m\left(i\right)/(N-m)---\left(3\right)$

(3.4) ask mean value to all i.

$B^m\left(r\right)=\frac{\underset{i=1}{\overset{N-m}{\mathrm{\&Sigma;}}}{B}_r^m\left(i\right)}{N-m+1}---\left(4\right)$

(3.5) dimension adds 1, obtains the m+1 n dimensional vector n:

X _m+1(i)＝[u(i)，u(i+1)，…u(i+m)]i＝1，2，…，N-m (5)

(3.6) repeating step (2), (3), (4), try to achieve for the m+1 n dimensional vector n:

$B_r^{m+1}\left(i\right)=N^m\left(i\right)/(N-m-1)---\left(6\right)$

$B^{m+1}\left(r\right)=\frac{\underset{i=1}{\overset{N-m}{\mathrm{\&Sigma;}}}{B}_r^{m+1}\left(i\right)}{N-m}---\left(7\right)$

(3.7) subsequence corresponding sample entropy does in theory

SampEn(m，r，N)＝lim{-In[B ^m+1/B ^m(r)]} (8)

If N can be expressed as during for finite value

SampEn(m，r，N)＝-In[B ^m+1/B ^m(r)] (9)

Can know to (3.7) that by step (3.1) Sample Entropy calculating needs to confirm to embed dimension m, similar tolerance limit r, three parameters of sample length N, different m is also different with r corresponding sample entropy.

The computing formula of Sample Entropy average SaEn in the said step (4) is following:

$\mathrm{SaEn}=\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\mathrm{SampEn}\left(i\right)-\max \left[\mathrm{SampEn}\right]-\min \left[\mathrm{SampEn}\right]}{(n-2)}---\left(10\right)$

SaEn is the Sample Entropy average in the formula, and SampEn (i) is the sample entropy, and max [SampEn] is a Sample Entropy maximal value in the whole sequence, and min [SampEn] is a Sample Entropy minimum value in the whole sequence.

RSaEn value in the said step (5), different pulse parameter only need be calculated one, and this is that its sample entropy is constant because for the full sized pules signal, and average is consistent with each sample entropy.

The computing formula of CSaEn in the said step (6) is following:

$\mathrm{CSaEn}=\frac{\mathrm{SaEn}}{\mathrm{RSaEn}}\&times;\mathrm{SD}\left(\mathrm{SaE}{n}_{n-2}\right)---\left(11\right)$

CSaEn is the Sample Entropy evaluation index result of single channel electric signal in the formula, and RSaEn is the reference sample entropy of given standard signal, and SD (SaEnn-2) is for removing the standard deviation of minimum and maximum value back Sample Entropy sequence, and computing formula is following:

$s=\sqrt{\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{(x_i-\stackrel{\&OverBar;}{x})}^2}{n-1}}---\left(12\right)$

The computing formula of TCSaEn in the said step (7) is following:

TCSaEn＝max(CSaEn1，CSaEn2) (13)

TCSaEn is the final Sample Entropy estimation of stability index of mariages electric signal in the formula, and the final quantized result of the whole mariages signal stabilization of conduct of CSaEn index maximum in the two paths of signals is chosen in expression.

The present invention compared with prior art has following advantage and beneficial effect:

(1) the present invention has introduced the welding process estimation of stability with Sample Entropy.Used welding process estimation of stability is based on statistical evaluation more at present, or subjective assessment.Shang Weiyou is used for the stable research report of estimating of welding process with Sample Entropy.

(2) the present invention has taken all factors into consideration the influence of various pulse parameters to the sample entropy.Traditional welding process estimation of stability is mostly under the situation of other parameter constants, carries out univariate research, thereby analyzes in the influence of this factor to stability.Welding is a complicated process; Various parameters interknit each other; Therefore monofactorial research can not be satisfied the demand of technical study, and method of the present invention has been introduced this notion of reference sample entropy, and the sample entropy and the reference sample entropy of sample compared; Eliminated the systematic effects of pulse parameter, the welding stability evaluation is more tallied with the actual situation the sample entropy.

(3) the present invention has designed comprehensive evaluation index.Welding process stability can be reacted through the size and the Changing Pattern of Sample Entropy.In order to characterize the two situation of change comprehensively; These two parameters of Sample Entropy average and Sample Entropy sequence standard deviation have been introduced; In conjunction with the reference sample entropy, designed the Sample Entropy evaluation index of one-channel signal, can reflect of the influence of welding process stability to Sample Entropy comprehensively.

(4) the present invention has the Double Wire Welding characteristic.Double Wire Welding is different from monofilament weldering, can not think the simple superposition of monofilament weldering, the electric signal stability evaluation method specialized designs of the present invention's design be applicable to the evaluation index of Double Wire Welding, can obtain the final quantitative evaluation result of Double Wire Welding stability automatically.

Description of drawings

Fig. 1 is the process flow diagram of a kind of mariages arc-welding of the present invention electric signal stability method for quantitatively evaluating.

Fig. 2 is the current waveform figure of stabilization signal in the embodiment of the invention.

Fig. 3 is than the current waveform figure of stabilization signal in the embodiment of the invention.

Fig. 4 is the current waveform figure of unstable signal in the embodiment of the invention.

Fig. 5 is the Sample Entropy distribution plan of three mariages current signals in the embodiment of the invention.

Embodiment

For making the object of the invention, technical scheme and advantage clearer, below in conjunction with accompanying drawing and embodiment, the present invention is done further detailed description, but implementation method of the present invention is not limited thereto with the scope that requires protection.

Fig. 1 specifically comprises the steps: for the process flow diagram of a kind of mariages arc-welding of the present invention electric signal stability method for quantitatively evaluating

(1) utilize the Wavelet analyzer of developing voluntarily to gather the current signal data of one period long period;

(2) data of being gathered are carried out pre-service, mainly comprise two steps, at first data are carried out wavelet packet filtering to reduce noise on analysis result's influence,, carried out the normalization processing secondly in order to prevent of the influence of electric signal size to the result;

(3) calculate Sample Entropy sequential value SampEn (i) under the corresponding given parameter of this measured signal;

(4) calculate the Sample Entropy average index S aEn that surveys electric signal;

(5) calculate sample entropy RSaEn under the corresponding given parameter of this measured signal;

(6) calculate single channel estimation of stability index CSaEn;

(7) calculate mariages electric signal Sample Entropy index TCSaEn.

Index S aEn, CSaEn, TCSaEn are and the present invention is directed to the mariages arc welding process and specialized designs.

It is example that the present invention analyzes with the current signal quantitative evaluation of mariages pulse MIC welding, and test platform is made up of equipment such as the integrated mariages Arc Welding Power of DSP, travel mechanism controller and soldering test platform, the dynamic Wavelet analyzer of welding arc, mariages pulse MIC welding soft-switching inversion power supply, wire-feed motor, double-wire welding gun, water tanks.In test, utilize the dynamic Wavelet analyzer of welding arc that waveform is gathered and analyzed, by the Control Software realization control waveform of the integrated mariages pulse MIC welding soft-switching inversion power supply of developing voluntarily.Used test specimen is the Q235 steel, thick 8.0mm, and welding wire adopts H08Mn2SiA, and diameter is Φ 1.0mm, and blanket gas is a straight argon.Gas flow 15L/min, dry extension of electrode 12.0mm, distance is 8.0mm between the two root bead silk ends, dull and stereotyped built-up welding mode.

Fig. 2 to Fig. 4 is the current waveform figure of three samples in the present embodiment, and these three samples are preserved by Wavelet analyzer by above-mentioned platform collection, imports matlab and carries out showing after the wavelet packet filtering.Corresponding major parameter is following:

The major parameter of Fig. 2 is: leading welding wire Lead1 peak point current is 300A, and background current is 50A; Following welding wire Trailing1 peak point current is 290A, and background current is 75A; The front and back peak value of pulse time is 6ms, and the base value time is 20ms.

The major parameter of Fig. 3 is: leading welding wire Lead2 peak point current is 270A, and background current is 90A; Following welding wire Trailing2 peak point current is 300A, and background current is 85A; The front and back peak value of pulse time is 5ms, and the base value time is 10ms.

The major parameter of Fig. 4 is: leading welding wire Lead3 peak point current is 330A, and background current is 90A; Following welding wire Trailing3 peak point current is 280A, and background current is 80A; The front and back peak value of pulse time is 4ms, and the base value time is 10ms.

Visible from scheming, the stability of its welding process is successively by good variation.Sample 3 wherein shown in Figure 4 to follow the welding wire electric current chaotic, can't accomplish normal pulse welding.

Get N=2000, m=2, r=0.07 carries out 21 set of calculated, remove minimum and maximum value after, the Sample Entropy curve that obtains is Fig. 5.ILead1 representes the guiding welding wire current signal of sample 1 among the figure, and what ITrailing1 represented sample 1 follows the welding wire current signal; ILead2 representes the guiding welding wire current signal of sample 2, and what ITrailing2 represented sample 2 follows the welding wire current signal; ILead3 representes the guiding welding wire current signal of sample 3, and what ITrailing3 represented sample 3 follows the welding wire current signal.Can find out the variation along with current stability, the height of Sample Entropy curve and amplitude of variation all demonstrate certain variation.The current signal of ITrailing3 is extremely unstable, and it is bigger in Fig. 5, also to show as the sample entropy, and variation range is wider.

According to Fig. 5, behind the removal maximin, find the solution the Sample Entropy average SaEn of actual measurement electric signal, the result is following:

ILead1 SaEn＝0.0151

ITrailing1 SaEn＝0.0178

ILead2 SaEn＝0.0394

ITrailing2 SaEn＝0.0449

ILead3 SaEn＝0.0304

ITrailing3 SaEn＝0.0565

For the different parameters that weakens to the influence of electric signal sample entropy, by given pulse parameter 6 reference pulses are set, and calculate the reference sample entropy RSaEn of corresponding ideal signal, the result is following:

ILead1 RSaEn＝0.0063

ITrailing1 RSaEn＝0.0063

ILead2 RSaEn＝0.0121

ITrailing2 RSaEn＝0.0121

ILead3 RSaEn＝0.0126

ITrailing3 RSaEn＝0.0126

Can find out that from The above results for the different electric signal of pulse parameter, its Sample Entropy has big gap, under the situation of desirable pulse parameter, the RSaEn value of sample 3 is twices of sample 1, and this has explained the importance of the reference sample entropy that the present invention set forth.

According to the CSaEn parameter, it is following to obtain monofilament stability quantitative evaluation result:

ILead1 CSaEn＝2.39E-03

ITrailing1 CSaEn＝4.52E-03

ILead2 CSaEn＝6.19E-03

ITrailing2 CSaEn＝6.68E-03

ILead3 CSaEn＝5.07E-03

ITrailing3 CSaEn＝3.41E-02

Can find out that from CSaEn index situation of change first sample is highly stable, the CSaEn index of two current signals is all very little; The two paths of signals stability of second sample is more or less the same, and first sample is relatively more unstable relatively; The leading welding wire current signal of the 3rd sample is more stable, and is more stable than second sample, but follows very instability of welding wire, and unstable degree is more than 10 times of sample 1 leading welding wire electric current.

Calculating mariages quantitative evaluation standard TCSaEn at last is that TCSaEn1 is 4.52E-03, and TCSaEn2 is 6.68E-03, and TCSaEn3 is 3.41E-02.Its magnitude relationship is: 4.52E-03 (sample 1)＜6.68E-03 (sample 2)＜3.41E-02 (sample 3).Can find out the variation along with current stability, TCSaEn presents corresponding variation, the reaction strictly according to the facts of TCSaEn size the welding electric signal stability.The TCSaEn index of sample 3 is about 7.5 times of TCSaEn index of stabilization signal sample 1, has distinguished degree of stability effectively.Simultaneously through the Sample Entropy quantitative evaluation find each signal to follow the leading relatively welding wire of welding wire electric current more unstable, corresponding CSaEn index is bigger, for the double wire welding technical study provides new clue.This difference is indiscoverable through naked eyes, the significance of the welding process stability indicator that this also further specifies the present invention and is designed.

Above-mentioned quantitative analysis conclusion is consistent with the current stability analysis conclusion of Fig. 2 to Fig. 4.And the TCSaEn index result of calculation of 3 sample current shows, method for quantitatively evaluating of the present invention is consistent with the waveform actual conditions, can correctly reflect the stability of welding process.

Therefore can estimate the stability of Double Wire Welding electric signal effectively through the Sample Entropy algorithm; The evaluation index TCSaEn of design has taken all factors into consideration the difference, electric signal different parameters of mariages pulse MIC welding and the monofilament pulse MIC welding influence to Sample Entropy voluntarily; Can estimate the welding process electric signal stability preferably, can be used as an index of weld procedure specification performance synthesis evaluation model.

This embodiment explains that a kind of mariages arc-welding of the present invention electric signal stability method for quantitatively evaluating tallies with the actual situation, and has actual application value preferably.

In this instructions, the present invention is described with reference to its certain embodiments.But, still can make various modifications and conversion obviously and not deviate from the spirit and scope of the present invention.Therefore, instructions and accompanying drawing are regarded in an illustrative, rather than a restrictive.

Claims (6)

1. mariages arc-welding electric signal stability method for quantitatively evaluating; It is characterized in that gathering current signal; Electric signal after the denoising is calculated its sample entropy; Calculate the Double Wire Welding Sample Entropy quantitative evaluation index of design voluntarily on this basis again, realize the quantitative evaluation of electric signal, specifically comprise the steps:

(1) utilize the Wavelet analyzer of developing voluntarily to gather the current signal data of one period long period;

(2) data of being gathered are carried out pre-service;

(3) calculate Sample Entropy sequential value SampEn (i) under the corresponding given parameter of this measured signal;

(4) calculate the Sample Entropy average index S aEn that surveys electric signal;

(5) calculate sample entropy RSaEn under the corresponding given parameter of this measured signal;

(6) calculate single channel estimation of stability index CSaEn;

(7) calculate mariages electric signal Sample Entropy index TCSaEn.

2. according to the said a kind of mariages arc-welding electric signal stability method for quantitatively evaluating of claim 1; It is characterized in that; Pre-service in the step (2) mainly comprises two steps; At first data are carried out wavelet packet filtering to reduce noise on analysis result's influence,, carried out the normalization processing secondly in order to prevent of the influence of electric signal size to the result.

3. according to the said a kind of mariages arc-welding electric signal stability method for quantitatively evaluating of claim 1, it is characterized in that the RSaEn in the step (4) is meant the sample entropy of the ideal pulse signal that calculates according to the set-point of surveying sample.

4. according to the said a kind of mariages arc-welding electric signal stability method for quantitatively evaluating of claim 1, it is characterized in that the SaEn in the step (5) calculates by following formula:

$\mathrm{SaEn}=\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\mathrm{SampEn}\left(i\right)-\max \left[\mathrm{SampEn}\right]-\min \left[\mathrm{SampEn}\right]}{(n-2)}$

5. according to the said a kind of mariages arc-welding electric signal stability method for quantitatively evaluating of claim 1, it is characterized in that the CSaEn in the step (6) calculates by following formula:

$\mathrm{CSaEn}=\frac{\mathrm{SaEn}}{\mathrm{RSaEn}}\&times;\mathrm{SD}\left(\mathrm{SaE}{n}_{n-2}\right)$

6. according to the said a kind of mariages arc-welding electric signal stability method for quantitatively evaluating of claim 1, it is characterized in that the TCSaEn in the step (7) calculates by following formula:

TCSaEn＝max(CSaEn1，CSaEn2)

Images