A flash butt welding quality evaluation method based on welding process characteristics

Abstract

The application discloses a flash butt welding quality evaluation method based on welding process characteristics, comprising the following steps: (1) collecting signals of the welding process through a current sensor, a displacement sensor and a top forging force sensor, determining a welding signal parameter sequence, and collecting a welding state sequence through a welding control system; (2) extracting features of the welding parameter sequence obtained in step (1) to obtain welding parameter feature values; (3) classifying the features collected; and (4) establishing a two-level index quality evaluation model to evaluate the welding quality. The application extracts typical features in different welding stages, evaluates the welding quality from multiple aspects, and guarantees the accuracy and scientificity of the evaluation. The two-level index evaluation model solves the problem that the flash butt welding process is affected by mutual coupling of multiple factors, the decisive influence of the primary index on the quality evaluation makes the evaluation process more simple, and the secondary index makes the evaluation result more accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of welding, in particular to a flash butt welding quality evaluation method based on welding process characteristics. BACKGROUND

[0002] As a kind of efficient welding means, flash butt welding is widely used in mechanical and electrical, building, transportation, oil drilling and metallurgical industry, etc., especially for large cross-section welded parts, such as pipe with diameter over 500mm, section over 20000mm 2 of profile can be welded. In the process of flash butt welding, the oxides and impurities on the end face of the welded workpiece are taken out by flash spark or squeezed out with liquid metal, so that high-strength high-quality butt joint can be obtained.

[0003] With the increasing application of flash butt welding technology, the requirement for welding quality is becoming more and more strict. Since flash butt welding is a process affected by multiple parameters coupling and discontinuous, it is difficult to establish a complete welding quality prediction model for the welding process. Chinese patent 2021115254636 discloses a real-time evaluation method and device for flash butt welding quality of automobile wheel rim. The invention collects information of the welding process through sensors, and transmits the obtained digital signals to an industrial computer after conversion. The industrial computer extracts the peak-to-valley value, average value in time domain, and main frequency and amplitude in frequency domain as characteristic values. Finally, by comparing the characteristic values with the allowable fluctuation range of the preset welding parameters, it is judged whether the welding quality is abnormal. The disadvantage of this method is that when analyzing the digital signal in time domain, since the welding machine is unknown at this time, it cannot be segmented according to the different stages of welding. Therefore, when tracing the cause of welding defects, it cannot be traced to the specific welding stage. Thus it cannot provide technical reference for later parameter adjustment.

[0004] Chinese patent 2016106623024 discloses a method and device for evaluating the quality of ultrasonic metal welding and an ultrasonic metal welding machine. The invention divides the welding process into several sub-time periods according to the preset time, and then obtains the average speed of the ultrasonic wave emission end depression in all sub-time periods. Finally, the average value of the average speed in all sub-time periods is calculated to obtain the average speed of the ultrasonic wave emission end depression in the whole welding process. The disadvantage of this method is that the characteristic value extracted can only reflect the overall characteristics of the welding process, but the local characteristics such as the minimum depression speed which has a decisive influence on the welding quality cannot be reflected. Therefore, when using this characteristic to evaluate the quality, the result cannot be used as a direct basis for judging the welding quality. SUMMARY

[0005] Invention purposes: The present application aims to provide a method for evaluating the flash butt welding quality based on the welding process characteristics by establishing a two-level index evaluation model.

[0006] Technical solutions: The flash butt welding quality evaluation method based on the welding process characteristics comprises the following steps:

[0007] (1) The welding process is signal collected by the current sensor, displacement sensor and upsetting force sensor to determine the welding signal parameter sequence, and the welding state sequence is collected by the welding control system;

[0008] (2) The welding parameter sequence obtained in step (1) is feature extracted to obtain the welding parameter characteristic value;

[0009] (3) The collected features are classified;

[0010] (4) A two-level index quality evaluation model is established to evaluate the welding quality.

[0011] Further, step (1) comprises the following steps:

[0012] (1.1) The welding parameter sequence is obtained by signal collection of the welding process by the sensor, and the welding signal parameter sequence comprises the welding time sequence T 1×n =[t0... t i t i+1 ... t n ], the welding current sequence C 1×n =[c0... c i c i+1 ... c n ], the welding electrode position sequence L 1×n =[l0... l i l i+1 ... l n ] and the upsetting force sequence P 1×n =[p0... p i p i+1 ... p n ]; wherein t i , c i , l i and p i represent the welding time, welding current, welding electrode position and upsetting force corresponding to the i-th signal;

[0013] (1.2) The welding state sequence S 1×n =[s0... s i s i+1 ... s nThe welding state corresponding to the i-th welding signal value is denoted as s. i s i ∈{0,1,2,...,7}, where 0 is the flash leveling stage, 1 is the preheating advance stage, 2 is the preheating stop stage, 3 is the preheating retreat stage, 4 is the flash advance stage, 5 is the flash retreat stage, 6 is the upsetting advance stage, and 7 is the upsetting holding stage.

[0014] (1.3) The signal values ​​in steps (1.1) and (1.2) are generated according to the chronological order of the acquisition time, that is, for t i At time t, the acquired signal value of the weldment is represented as Singnal = [t] i ,s i ,c i ,l i ,p i ] T .

[0015] Furthermore, step (2) includes the following steps:

[0016] (2.1) Extract the number of preheating cycles N. For each preheating process, the welding state sequence starts from s. i To s j It should be [1,...,1,2,...,2,3,...,3], where s i As the starting point for preheating, s j This marks the end of the preheating process.

[0017] (2.1.1) Initialize the parameters, let N = 0, k = 0;

[0018] (2.1.2) Increment k by one;

[0019] (2.1.3) If s k =2 and s k-1 =1, then we are in the transition phase from preheating to preheating stopping, and proceed to step (2.1.4); otherwise, return to step (2.1.2);

[0020] (2.1.4) Add one more to k based on step (2.1.2);

[0021] (2.1.5) If s k =3 and s k-1 If N = 2, then increment N by one and return to step (2.1.2); otherwise, return to step (2.1.4).

[0022] (2.1.6) Repeat steps (2.1.2) to (2.1.5) until k > n, then the welding state sequence is traversed and the number of preheating times N is output.

[0023] (2.2) Extract the overall preheating intensity Q q traverse the welding state sequence to find s k =1 and c k For welding stages with currents exceeding the threshold, calculate the overall preheating intensity;

[0024] (2.3) Extract the duration t of continuous flash before upsetting;

[0025] (2.3.1) Initialize the parameters, let t = 0, k = n + 1;

[0026] (2.3.2) Decrease k by one;

[0027] (2.3.3) If s k If the value is ≥5, proceed to step (2.3.4); otherwise, proceed to step (2.3.6).

[0028] (2.3.4) If s k =6 and s k-1 =5, then record T 1×n The signal value t in column k k If the condition is not met, proceed to step (2.3.5); otherwise, return to step (2.3.2).

[0029] (2.3.5) The time constant t e1 Let t be the time when the continuous flash ends. e1 =t k Then return to step (2.3.2);

[0030] (2.3.6) If s k =4 and s k-1 If ≠4, then record T. 1×n The signal value t in column k k If the condition is met, proceed to step (2.3.7); otherwise, return to step (2.3.2).

[0031] (2.3.7) The time constant t s1 Let t be the starting time of the continuous flashes. s1 =t k Then proceed to step (2.3.8);

[0032] (2.3.8) Duration t of continuous flashes:

[0033] t = t e1 -t s1

[0034] (2.4) Extract the upsetting speed v;

[0035] (2.4.1) Initialize the parameters, let v = 0, k = 0;

[0036] (2.4.2) let k = k + 1;

[0037] (2.4.3) if s k ≤ 6, go to step (2.4.4); if s k > 6, go to step (2.4.6);

[0038] (2.4.4) if s k = 6 and s k-1 ≠ 6, query the record T 1×n , L 1×n the kth column signal value t k and l k , and go to step (2.4.5); otherwise, return to step (2.4.2);

[0039] (2.4.5) let t s2 and l s2 be the initial time and position of the upsetting advancing stage, let t s2 = t k , l s2 = l k , and return to step (2.4.2);

[0040] (2.4.6) if s k = 7 and s k-1 ≠ 7, query the record T 1×n , L 1×n the kth column signal value t k and l k , and go to step (2.4.7); otherwise, return to step (2.4.2);

[0041] (2.4.7) let t e2 and l e2 be the final time and position of the upsetting advancing stage, let t e2 = t k , l e2 = l k , and go to the next step;

[0042] (2.4.8) calculate the upsetting speed v of the upsetting advancing stage:

[0043]

[0044] (2.5) extract the parameter x of the overall position change of the workpiece in the upsetting holding stage;

[0045] (2.5.1) initialize the parameter, let x = 0, k = n + 1, l max = 0, l min = 100e, e is a constant, and l max is the farthest position of the electrode;min is the farthest position;

[0046] (2.5.2) Let k = k - 1;

[0047] (2.5.3) If s k ≠ 7, repeat step (2.5.2) until s k = 7, then go to next step;

[0048] (2.5.4) If s k = 7, then query record L 1×n , k column signal value l k , and go to next step;

[0049] (2.5.5) If l max < l k , then let l max = l k , otherwise do not assign value, and go to next step;

[0050] (2.5.6) If l min > l k , then let l min = l k , otherwise do not assign value, and return to step (2.5.2);

[0051] (2.5.7) Repeat steps (2.5.2) to (2.5.6) until k = 0, the traversal of the welding state sequence is completed, and the farthest position of the electrode in the top forging holding stage is l max , and the nearest position is l min .

[0052] (2.5.8) Calculate the total position change of the welding piece, and the calculation formula is:

[0053] x = l max - l min

[0054] (2.6) Extract the total position change rate x' of the top forging holding stage;

[0055] (2.6.1) Initialize parameters. x' = 0, k = n + 1;

[0056] (2.6.2) Let k decrease by one;

[0057] (2.6.3) If s k ≠ 7, return to (2.6.2) until s k = 7, then go to step (2.6.4);

[0058] (2.6.4) Query record T 1×n , L 1×nThe signal values ​​t in column k and column (k-1) k t k-1 With l k l k-1 Then proceed to step (2.6.5);

[0059] (2.6.5) Calculate the rate of change of position x' k :

[0060]

[0061] (2.6.6) If x′<x′ k Let x′=x′ k ;

[0062] (2.6.7) If x′≥x′ k If the result is not specified, no assignment operation will be performed, and the process will return to step (2.6.2).

[0063] (2.6.8) Repeat steps (2.6.2) to (2.6.7) until k = 0, and the welding state sequence is completed, and the maximum position change rate x′ of the electrode during the upsetting holding stage is obtained;

[0064] (2.7) Extract the upsetting force parameter P f ;

[0065] (2.7.1) Initialize parameters, let the average upsetting force be . Minimum upsetting force p min =8e, where e is a constant, f=0, k=0, count=0;

[0066] (2.7.2) Increment k by one;

[0067] (2.7.3) If s k If the value is less than 6, repeat step (2.7.2) until s. k After reaching ≥6, proceed to step (2.7.4);

[0068] (2.7.4) If s k =6 or 7, then query record P 1×n The signal value p in column k k Proceed to step (2.7.5);

[0069] (2.7.5) Let P f+1 =P f +p k ;

[0070] (2.7.6) Increment count by one;

[0071] (2.7.7) If p min >p k Then let pmin =p k ;

[0072] (2.7.8) If p min ≤p k If the assignment operation fails, the process returns to step (2.7.2).

[0073] (2.7.9) Repeat steps (2.7.2) to (2.7.8) until k > n, the welding process ends, and the sum of the upsetting forces P and the minimum upsetting force p are obtained. min ;

[0074] (2.7.10) Calculate the average upsetting force p during the upsetting process:

[0075]

[0076] Furthermore, step (2.2) extracts the overall preheating intensity Q. q This includes the following steps:

[0077] (2.2.1) Initialize the parameters, let Q0 = 0, k = 0;

[0078] (2.2.2) Increment k by one;

[0079] (2.2.3) If s k =1, then query C 1×n The signal value c in column k k If the condition is not met, proceed to step (2.2.4); otherwise, return to step (2.2.2).

[0080] (2.2.4) If c k >I, where I is the threshold current, then query record c k t k-1 With t k Let q increment by one to proceed to step (2.2.5);

[0081] (2.2.5) Overall preheating intensity Q q =Q q-1 +c k ×(t k -t k-1 ), and return to step (2.2.2);

[0082] (2.2.6) If c k If the value is less than or equal to I, then no recording operation is performed, and the process returns to step (2.2.2).

[0083] (2.2.7) Repeat steps (2.2.2) to (2.2.6) until k > n, traversing the welding state sequence to end, and output the overall preheating intensity Q. q .

[0084] Further, the step (3) classifying the collected features is to analyze the causes of the defects of the welded joint and the influence of the defects on the quality of the welded part, and classify the extracted features; the overall position change x, the position change rate x', the continuous flash duration t, the minimum upset force p min The upset speed v is taken as the T1 level index, the preheating times N, the overall preheating intensity Q q The average upset force p is taken as the T2 level index.

[0085] Further, the step (4) includes the following steps:

[0086] (4.1) For the T1 level index, determine the qualified interval of the parameter value: the overall position change x∈[0,b1], the position change rate x'∈[0,b2], the continuous flash duration t∈[a3,+∞), the minimum upset force p min ∈[a4,+∞], the upset speed v∈[a5,+∞), wherein b1, b2, a3, a4 and a5 are constants greater than 0;

[0087] (4.2) Introduce the indicator function to quantitatively analyze the influence of the T1 level index on the evaluation of the welding quality, and the indicator function includes:

[0088] The indicator function f1(x) of the overall position change x:

[0089]

[0090] The indicator function f2(x) of the position change rate x':

[0091]

[0092] The indicator function f3(x) of the continuous flash duration t:

[0093]

[0094] The indicator function f4(x) of the minimum upset force p min

[0095]

[0096] The indicator function f4(x) of the upset speed v:

[0097]

[0098] (4.3) For the T2 level index, calculate the deviation of the T2 level feature parameter from the ideal value, and normalize the processing;

[0099] ​

[0100]

[0101]

[0102] wherein, ΔN * , ΔQ * and is the deviation value after standardization, N is the preheating number, N * is the ideal value of the preheating number, Q q is the total preheating intensity, Q * is the ideal value of the preheating intensity, is the average upsetting force, is the ideal value of the average upsetting force;

[0103] (4.5) determining the weight of the T2 level index;

[0104] (4.5.1) setting the total preheating intensity weight Weight(Q) = λ1, the preheating number weight Weight(N) = λ2, and the average upsetting force weight Weight(F) = λ3 , and λ1 + λ2 + λ3 = 1;

[0105] (4.5.2) obtaining the weight vector w = [λ1, λ2, λ3] of the T2 level index T ;

[0106] (4.6) establishing a welding quality evaluation model;

[0107] (4.6.1) quality evaluation function W:

[0108]

[0109] if W > σ, the welding quality is judged to be qualified; if W ≤ σ, the welding quality is judged to be unqualified; wherein σ is a standard quality index.

[0110] Advantages: compared with the prior art, the present application has the following advantages: 1. the thought that different welding stages have different influences on the welding quality is embodied in the feature extraction process, the original feature parameters are reserved when the quality is evaluated, and the stage where the defect occurs can be tracked when the welding defect occurs; 2. eight typical features in different welding stages are extracted, the welding quality is evaluated from multiple aspects, and the accuracy and scientificity of the evaluation are ensured; 3. the two-level index evaluation model solves the problem that the flash butt welding process is affected by the mutual coupling of multiple factors to a certain extent, the decisive influence of the primary index on the quality evaluation makes the evaluation process more simple, and the secondary index makes the evaluation result more accurate. BRIEF DESCRIPTION OF DRAWINGS

[0111] Figure 1 is the flow chart of the present application;

[0112] Figure 2 Flow chart for extracting preheating times;

[0113] Figure 3 Flow chart for extracting overall preheating intensity;

[0114] Figure 4 Flow chart for extracting continuous flash duration;

[0115] Figure 5 Flow chart for extracting upset speed in upset stage;

[0116] Figure 6 Flow chart for extracting overall position change of weldment in upset holding stage;

[0117] Figure 7 Flow chart for extracting position change rate of weldment in upset holding stage;

[0118] Figure 8 Flow chart for extracting average upset force and minimum upset force in upset stage. DETAILED DESCRIPTION

[0119] The application will be further described below with reference to the accompanying drawings.

[0120] The flash butt welding quality evaluation method based on welding process feature extraction according to the application has the process as shown in FIG. 1 and is specifically as follows: Figure 1

[0121] (1) Welding signal parameter sequences are obtained by signal acquisition of the welding process through current sensors, displacement sensors and upset force sensors, and welding state sequences are acquired through a welding control system;

[0122] (1.1) 200 signal values are acquired for one complete welding process, and the welding time sequence is represented as T 1×200 =[t0... t 100 t 101 ... t 200 ], the welding current sequence is represented as C 1×200 =[c0... c 100 c 101 ...c 200 ], the welding electrode position sequence is represented as L 1×200 =[l0... l 100 l 101 ... l 200 ], and the upset force sequence is represented as P 1×200 =[p0... p 100 p 101 ... p 200 ​], the welding state sequence is represented as S 1×200 = [s0...s 100 101 ...s 200 ], wherein s i ∈{0, 1, 2,..., 7}.

[0123] (1.2) The signal value in step (1.1) is generated according to the chronological order of the collection time, that is, for t 100 , the collected signal value of the welding part can be represented as Singnal = [t 100 , s 100 , c 100 , l 100 , p 100 ] T .

[0124] (2) Feature extraction is performed on the welding parameter sequence obtained in step (1) to obtain welding parameter feature values;

[0125] (2.1) Extract the preheating number N. As Figure 2 is a flowchart for extracting the preheating number, and the details are as follows:

[0126] (2.1.1) Initialize parameters. Let N = 0, k = 0;

[0127] (2.1.2) Let k increase by one;

[0128] (2.1.3) If s k = 2 and s k-1 = 1, at this time, it is in the transition stage from preheating forward s k = 1 to preheating stop s k = 2, and the next step (determine whether preheating retreat s k = 3 is performed) is executed; otherwise, return to step (2.1.2);

[0129] (2.1.4) Increase k by one on the basis of k increasing by one in step (2.1.2);

[0130] (2.1.5) If s k = 3 and s k-1 = 2, N = N + 1, and return to step (2.1.2); otherwise, return to step (2.1.4);

[0131] (2.1.6) Repeat steps (2.1.2) to (2.1.5) until k > 200, and the traversal of the welding state sequence ends, and the preheating number N is output.

[0132] (2.2) Extract the total preheating intensity Q q ​Parameter settings: Threshold current I = 450A. The process for extracting the overall preheating intensity is as follows: Figure 3 As shown, the details are as follows:

[0133] (2.2.1) Initialize the parameters, let Q0 = 0, k = 0;

[0134] (2.2.2) Increment k by one;

[0135] (2.2.3) If s k =1, then query C 1×n The signal value c in column k k If the condition is not met, proceed to step (2.2.4); otherwise, return to step (2.2.2).

[0136] (2.2.4) If c k >I(450A), where I is the threshold current, then query record c k t k-1 With t k Let q increment by one to proceed to step (2.2.5);

[0137] (2.2.5) Overall preheating intensity Q q =Q q-1 +c k ×(t k -t k-1 ), and return to step (2.2.2);

[0138] (2.2.6) If c k If the value is less than or equal to I(450A), then no recording operation is performed, and the process returns to step (2.2.2).

[0139] (2.2.7) Repeat steps (2.2.2) to (2.2.6) until k > 200, then the welding state sequence is completed, and the overall preheating intensity Q is output. q .

[0140] (2.3) Extract the duration t of continuous flashes before upsetting. The process for extracting the duration of continuous flashes is as follows: Figure 4 As shown, the specific steps are as follows:

[0141] (2.3.1) Initialize parameters. t = 0, k = 201;

[0142] (2.3.2) Decrease k by one;

[0143] (2.3.3) If s k If the value is ≥5, proceed to step (2.3.4); otherwise, proceed to step (2.3.6).

[0144] (2.3.4) If s k =6 and s k-1=5, then record T 1×200 The signal value t in column k k If the condition is met, proceed to the next step; otherwise, return to step (2.3.2).

[0145] (2.3.5) The time constant t e1 Let t be the time when the continuous flash ends. e1 =t k Then return to step (2.3.2);

[0146] (2.3.6) If s k =4 and s k-1 If ≠4, then record T. 1×200 The signal value t in column k k If the condition is met, proceed to the next step; otherwise, return to step (2.3.2).

[0147] (2.3.7) The time constant t s1 Let t be the starting time of the continuous flashes. s1 =t k Then proceed to the next step;

[0148] (2.3.8) The initial time t of the continuous flash phase obtained according to the above steps s1 With the final time t e1 The duration t of the continuous flashes is obtained:

[0149] t = t e1 -t s1

[0150] (2.4) Extracting the upsetting speed v. The process for extracting the upsetting speed is as follows: Figure 5 As shown, the specific steps are as follows:

[0151] (2.4.1) Initialize parameters. v = 0, k = 0;

[0152] (2.4.2) Let k increase by one;

[0153] (2.4.3) If s k If ≤6, then proceed to step (2.4.4); if s k If the value is greater than 6, then proceed to step (2.4.6).

[0154] (2.4.4) If s k =6 and s k-1 If the value is not equal to 6, then query record T. 1×200 L 1×200 The signal value t in column k k With l k If the condition is met, proceed to the next step; otherwise, return to step (2.4.2).

[0155] (2.4.5) The constant t s2 With l s2 Let t be the initial time and position of the upsetting forward stage. s2 =t k , l s2 =l k And return to step (2.4.2);

[0156] (2.4.6) If s k =7 and s k-1 If the value is not equal to 7, then query record T. 1×200 L 1×200 The signal value t in column k k With l k If the condition is met, proceed to the next step; otherwise, return to step (2.4.2).

[0157] (2.4.7) The constant t e2 With l e2 Let t be the final moment and position of the upsetting forward stage. e2 =t k , l e2 =l k Then proceed to the next step;

[0158] (2.4.8) Calculate the upsetting speed v during the upsetting process:

[0159]

[0160] (2.5) Extract the overall position change parameter x of the weldment during the upsetting holding stage. Parameter setting: the farthest position of the electrode is l. max The nearest position is l min The process for extracting overall location change parameters is as follows: Figure 6 As shown, the specific steps are as follows:

[0161] (2.5.1) Initialize parameters. x = 0, k = 201, l max =0,l min =100e (the natural constant e is taken as 2.718);

[0162] (2.5.2) Decrease k by one;

[0163] (2.5.3) If s k If ≠ 7, then repeat step (2.5.2) until s. k =7, then proceed to the next step;

[0164] (2.5.4) If s k =7, then query record L 1×200 The signal value in column k k Then proceed to the next step;

[0165] (2.5.5) If l max < l k , then let l max = l k , otherwise do not assign value and go to next step;

[0166] (2.5.6) If l min > l k , then let l min = l k , otherwise do not assign value and return to step (2.5.2);

[0167] (2.5.7) Repeat steps (2.5.2) to (2.5.6) until k = 0, the traversal of the sequence of welding states is completed, and the farthest position of the electrode in the top-punching holding stage is l max and the nearest position is l min .

[0168] (2.5.8) Calculate the total position change amount x of the welding piece:

[0169] x = l max - l min

[0170] (2.6) Extract the total position change rate x' of the top-punching holding stage. As Figure 7 is the flow chart for extracting the total position change rate, the specific steps are as follows:

[0171] (2.6.1) Initialize parameters. x' = 0, k = 201;

[0172] (2.6.2) Let k decrease by one;

[0173] (2.6.3) If s k ≠ 7, repeat step (2.6.2) until s k = 7, and then go to next step;

[0174] (2.6.4) If s k = 7, query the record T 1×200 , L 1×200 kth column, (k-1)th column signal values t k , t k-1 and l k , l k-1 , and go to next step;

[0175] (2.6.5) Calculate the position change rate x' k :

[0176]

[0177] (2.6.6) If x' < x'k Let x′=x′ k ;

[0178] (2.6.7) If x′≥x′ k If the result is not specified, no assignment operation will be performed, and the process will return to step (2.6.2).

[0179] (2.6.8) Repeat steps (2.6.2) to (2.6.7) until k = 0, and the welding state sequence is completed, and the maximum position change rate x′ of the electrode during the upsetting holding stage is obtained;

[0180] (2.7) Extract the upsetting force parameters. Parameter settings: average upsetting force p, minimum upsetting force p min The process for extracting upsetting force parameters is as follows: Figure 8 As shown, the details are as follows:

[0181] (2.7.1) Initialize parameters. P = 0, p min =16MPa, k=0, count=0;

[0182] (2.7.2) Let k increase by one;

[0183] (2.7.3) If s k If the value is less than 6, repeat step (2.7.2) until s. k If ≥6, proceed to the next step;

[0184] (2.7.4) If s k =6 or 7, then query record P 1×200 The signal value p in column k k Then proceed to the next step;

[0185] (2.7.5) Let P f+1 =P f +p k ;

[0186] (2.7.6) count = count + 1;

[0187] (2.7.7) If p min >p k Then let p min =p k ;

[0188] (2.7.8) If p min ≤p k If the assignment operation fails, the process returns to step (2.7.2).

[0189] (2.7.9) Repeat steps (2.7.2) to (2.7.8) until k > 200, the welding process is over, and the sum P of the upsetting forces and the minimum upsetting force p of the upsetting process are obtained min .

[0190] (2.7.10) Calculate the average upsetting force of the upsetting process

[0191]

[0192] (3) Classify the collected features

[0193] The overall position change x, the position change rate x', the continuous flash duration t, and the minimum upsetting force p of the upsetting stage min are taken as T1-level indicators, and the preheating number N and the overall preheating intensity Q q and the average upsetting force are taken as T2-level indicators.

[0194] (4) Establish a welding quality evaluation model to evaluate the welding quality.

[0195] (4.1) For the first-level indicators, first determine the qualified interval of the parameter values.

[0196] (4.1.1) The overall position change x ∈ [0, 0.5], unit: mm;

[0197] (4.1.2) The position change rate x' ∈ [0, 0.05];

[0198] (4.1.3) The continuous flash duration t ∈ [7.5, +∞), unit: s;

[0199] (4.1.4) The minimum upsetting force p min ∈ [14, +∞], unit: MPa;

[0200] (4.1.5) The upsetting speed v ∈ [27.5, +∞), unit: mm / s;

[0201] (4.2) Introduce an indicator function to quantitatively analyze the influence of the first-level indicators on the welding quality evaluation.

[0202] (4.2.1) The indicator function f1(x) of the overall position change x is:

[0203]

[0204] (4.2.2) The indicator function f2(x) of the position change rate x' is:

[0205]

[0206] (4.2.3) The characteristic function f3(x) of the continuous flash duration t is:

[0207]

[0208] (4.2.4) The minimum upset force p min The characteristic function f4(x) is:

[0209]

[0210] (4.2.5) The upset speed v The characteristic function f4(x) is:

[0211]

[0212] (4.3) For the secondary index, first determine the ideal value of the parameter value.

[0213] (4.3.1) The ideal value N of the preheating number * = 9 times;

[0214] (4.3.2) The ideal value Q of the preheating intensity * = 1000 A·s;

[0215] (4.3.3) The average upset force

[0216] (4.4) Calculate the deviation of the secondary characteristic parameters from the ideal value, and perform normalization processing;

[0217]

[0218]

[0219]

[0220] Where, ΔN * , ΔQ * and are the deviation values after standardization processing.

[0221] (4.5) Determine the weight of the secondary index.

[0222] (4.5.1) Let Weight(Q) = λ1(0.55), Weight(N) = λ2(0.2), and λ1+λ2+λ3=1;

[0223] (4.5.2) Get the weight vector w of the secondary index = [0.55, 0.2, 0.25] T ;

[0224] (4.6) Establishing a welding quality evaluation model.

[0225] (4.6.1) The quality evaluation function is:

[0226]

[0227] (4.7.2) If W > σ(0.8), the welding quality is judged to be qualified;

[0228] (4.7.3) If W ≤ σ(0.8), the welding quality is judged to be unqualified.

Claims

1. A method for evaluating the quality of flash butt welding based on welding process characteristics, characterized in that, Includes the following steps: (1) The welding process is signal acquired by current sensor, displacement sensor and upsetting force sensor to determine the welding signal parameter sequence, and the welding status sequence is acquired by welding control system; (2) Extract features from the welding parameter sequence obtained in step (1) to obtain the feature values ​​of the welding parameters; (3) Classify the collected features; (4) Establish a two-level index quality assessment model to evaluate welding quality; The classification of the collected features in step (3) is achieved by analyzing the causes of defects in the weld joint and the degree of impact of the defects on the quality of the weldment, and then classifying the extracted features. Changes in the overall position during the upsetting stage Rate of change of position Continuous flash duration Minimum forging force With upsetting speed As a T1 level indicator, the number of warm-ups Overall preheating intensity With average upsetting force As a T2 level indicator.

2. The flash butt welding quality assessment method based on welding process characteristics according to claim 1, characterized in that, Step (1) includes the following steps: (1.1) The welding process is monitored by sensors to obtain a welding parameter sequence, which includes a welding time sequence. Welding current sequence welding electrode position sequence and upsetting force sequence ;in, , , and Indicates the first The welding time, welding current, welding electrode position, and upsetting force corresponding to each signal; (1.2) Collect welding status sequence through welding control system , No. The welding state corresponding to each welding signal value is denoted as: , Among them, 0 is the flash leveling stage, 1 is the preheating forward stage, 2 is the preheating stop stage, 3 is the preheating retreat stage, 4 is the flash forward stage, 5 is the flash retreat stage, 6 is the upsetting forward stage, and 7 is the upsetting holding stage. (1.3) The signal values ​​in steps (1.1) and (1.2) are generated according to the chronological order of the acquisition time, that is, for At any given time, the signal value acquired by the weldment is represented as follows: .

3. The flash butt welding quality assessment method based on welding process characteristics according to claim 1, characterized in that, Step (2) includes the following steps: (2.1) Number of preheating cycles In each preheating process, the welding state sequence changes from... to Should be ,in This is the starting point for preheating. This marks the end of the preheating process; (2.1.1) Initialize the parameters, let , ; (2.1.2) Increment k by one; (2.1.3) If If the preheating progresses to the preheating stop phase, proceed to step (2.1.4); otherwise, return to step (2.1.2). (2.1.4) Add one more to k based on step (2.1.2); (2.1.5) If If N is positive, increment N by one and return to step (2.1.2); otherwise, return to step (2.1.4). (2.1.6) Repeat steps (2.1.2) to (2.1.5) until... After traversing the welding state sequence, output the number of preheating cycles. ; (2.2) Extracting the overall preheating intensity traversing the welding state sequence to find and For welding stages with currents exceeding the threshold, calculate the overall preheating intensity; (2.3) Extract the duration of continuous flash before upsetting ; (2.3.1) Initialize parameters, let , ; (2.3.2) Let Subtract one; (2.3.3) If If yes, proceed to step (2.3.4); otherwise, proceed to step (2.3.6). (2.3.4) If Then record No. Column signal value If the condition is not met, proceed to step (2.3.5); otherwise, return to step (2.3.2). (2.3.5) The time constant Let be the time when the continuous flash ends, and let , and return to step (2.3.2); (2.3.6) If Then record No. Column signal value If the condition is not met, proceed to step (2.3.7); otherwise, return to step (2.3.2). (2.3.7) The time constant Let be the starting time of the continuous flashes, and let And proceed to step (2.3.8); (2.3.8) Duration t of continuous flashes: ； (2.4) Extracting the upsetting speed ; (2.4.1) Initialize parameters, let , ; (2.4.2) Increment k by one; (2.4.3) If If so, proceed to step (2.4.4); if Then proceed to step (2.4.6). (2.4.4) If Then query records , No. Column signal value and If the condition is not met, proceed to step (2.4.5); otherwise, return to step (2.4.2). (2.4.5) constant and Let be the initial time and position of the upsetting forward stage, and let . , , and return to step (2.4.2); (2.4.6) If Then query records , No. Column signal value and If the condition is met, proceed to step (2.4.7); otherwise, return to step (2.4.2). (2.4.7) constant and Record the final moment and position of the upsetting advance stage, and let , Then proceed to the next step; (2.4.8) Calculate the upsetting speed v during the upsetting process: ； (2.5) Extracting the overall positional change parameters of the weldment during the upsetting holding stage ; (2.5.1) Initialize parameters, let , , , , It is a constant. The farthest position of the electrode. The nearest location; (2.5.2) Let ; (2.5.3) If Then repeat step (2.5.2) until... Then proceed to the next step; (2.5.4) If Then query records No. Column signal value Then proceed to the next step; (2.5.5) If Then let Otherwise, no assignment operation is performed, and the next step is executed; (2.5.6) If Then let Otherwise, no assignment operation is performed, and the process returns to step (2.5.2). (2.5.7) Repeat steps (2.5.2) through (2.5.6) until... After traversing the welding state sequence, the farthest position of the electrode during the upsetting holding stage is obtained. Recent location ; (2.5.8) Calculate the overall positional change of the weldment using the following formula: ； (2.6) Extract the overall position change rate during the upsetting holding stage ; (2.6.1) Initialize parameters; , ; (2.6.2) Let Subtract one; (2.6.3) If Then return to (2.6.2) until... Then proceed to step (2.6.4). (2.6.4) Query records , No. Column, No. Column signal value , and , And proceed to step (2.6.5); (2.6.5) Calculate the rate of change of position : ； (2.6.6) If Then let ; (2.6.7) If If the assignment operation fails, return to step (2.6.2). (2.6.8) Repeat steps (2.6.2) through (2.6.7) until... After traversing the welding state sequence, the maximum position change rate of the electrode during the upsetting holding stage is obtained. ; (2.7) Extracting upsetting force parameters ; (2.7.1) Initialize parameters, let Average upsetting force Minimum upsetting force , It is a constant. , , ; (2.7.2) Let Add one; (2.7.3) If Then repeat step (2.7.2) until... Then proceed to step (2.7.4). (2.7.4) If Then query records No. Column signal value Proceed to step (2.7.5); (2.7.5) Let ; (2.7.6) Let Add one; (2.7.7) If Then let ; (2.7.8) If If the assignment operation fails, the process returns to step (2.7.2). (2.7.9) Repeat steps (2.7.2) through (2.7.8) until... After the welding process is completed, the sum of the forging forces during the upsetting process is obtained. With minimum upsetting force ; (2.7.10) Calculate the average upsetting force during the upsetting process. : 。 4. The flash butt welding quality assessment method based on welding process characteristics according to claim 3, characterized in that, Step (2.2) Extract the overall preheating intensity This includes the following steps: (2.2.1) Initialize parameters, let , ; (2.2.2) Let Add one; (2.2.3) If Then query No. Column signal value Proceed to step (2.2.4); otherwise, return to step (2.2.2). (2.2.4) If , If it is the threshold current, then query the record. , and ,make Add one to proceed to step (2.2.5); (2.2.5) Overall preheating intensity , and return to step (2.2.2); (2.2.6) If If not, no recording operation will be performed, and the process will return to step (2.2.2). (2.2.7) Repeat steps (2.2.2) to (2.2.6) until... After traversing the welding state sequence, the overall preheating intensity is output. .

5. The flash butt welding quality assessment method based on welding process characteristics according to claim 1, characterized in that, Step (4) includes the following steps: (4.1) For T1 level indicators, determine the acceptable range of parameter values: overall positional change Rate of change of position Continuous flash duration Minimum upsetting force Upsetting speed ,in, , , , and A constant greater than 0; (4.2) Indicator functions are introduced to quantitatively analyze the impact of T1 grade indicators on welding quality assessment. The indicator functions include: Overall location change Indicator function : ； Rate of change of position Indicator function : ； Continuous flash duration Indicator function : ； Minimum upsetting force Indicator function : ； Upsetting speed Indicator function : ； (4.3) For T2 level indicators, calculate the deviation between the T2 level characteristic parameters and the ideal values, and normalize them; ； ； ； in, , and These are the deviation values ​​after standardization. It refers to the number of preheating cycles. This is the ideal number of preheating cycles. It is the overall preheating intensity. This is the ideal value for preheating intensity. It is the average upsetting force. It is the ideal value of average upsetting force; (4.5) Determine the weights of the T2 level indicators; (4.5.1) Let the overall preheating intensity weight be... Preheating times weight Average upsetting force weight ,and ; (4.5.2) Obtain the weight vector of the T2 level index. ; (4.6) Establish a welding quality assessment model; (4.6.1) Quality assessment function W: ； like If the welding quality is deemed acceptable, then the welding quality is considered acceptable; if If the welding quality is not up to standard, then the welding quality is deemed unqualified; among which This is the standard quality index.

Images