JP2004160510A - Welding quality discriminating method and device for the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a welding quality discriminating method for enhancing the welding quality determination accuracy, and a device for the same. <P>SOLUTION: The voltage between electrodes and the current are measured while the welding current is made to flow, and the change in the resistance between the electrodes is obtained. Characteristic parameters (resistance Rst immediately after the start of energization, the inverse Rst<SP>-1</SP>thereof, the maximum resistance Rpeak, the elapsed time thereof Tpeak, the average resistance Rave from the predetermined time T1 to the energization completion time WeldEnd, the resistance Rfin immediately before the finish of energization, and the inverse Rfin<SP>-1</SP>thereof) are generated from the change of the resistance between electrodes. Acceptance/rejection of the welding is discriminated from the characteristic parameters through at least one of the discrimination analysis and the multiple regression analysis. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method and apparatus for determining welding quality.
[0002]
[Prior art]
In normal mild steel plate spot welding, when normal welding is performed, the base material undergoes thermal expansion due to heat generation during the passage of welding current (nugget generation process). One of the conventional methods for judging the quality of welding using the thermal expansion of the base material is a change or absolute change in the relative position of the electrode tip from a reference position independent of the thickness of the base material after pressing. There is a method in which the amount of change in position is measured, the displacement between electrode tips during welding current application is detected based on the measured value, and the quality of welding is determined based on the displacement between electrode tips (Patent Document 1). ).
[0003]
[Patent Document]
JP 2000-5882 A
[0004]
[Problems to be solved by the invention]
The conventional method is based on the premise that the position of the tip of the electrode tip is not affected by other than the thermal expansion and the plate gap of the welding member. However, in actual welding, large vibrations and shocks are often applied to the welding gun due to electrode replacement and dust generation, which changes the amount of electrode tip fitted into the shank. . In some cases, the displacement between the electrodes cannot be measured accurately.
[0005]
In this regard, when replacing the electrode, it is possible to correct it by taking a method such as measuring the reference position between the electrode tips again after the replacement, but the generation of dust and vibration during welding may cause the electrode tip to be replaced. It is impossible to cope with a change in the fitting amount, and it is impossible to obtain an accurate displacement amount between the electrodes. For this reason, there has been a problem that the accuracy of estimating the displacement between the electrode tips is reduced, and accurate welding quality determination may not be performed.
[0006]
Therefore, an object of the present invention is to provide a welding quality determination method and apparatus that enable highly accurate welding quality determination even when the tip position of an electrode tip is affected by factors other than the thermal expansion of the base material and the gap of the base material. To provide.
[0007]
[Means for Solving the Problems]
The above object is to measure a voltage and a current between the welding electrodes at predetermined time intervals during welding current conduction, and to calculate a resistance value at each predetermined time interval from the obtained voltage and current values, Generating a resistance value characteristic parameter indicating a nugget generation process from the resistance value, and performing a discriminant analysis or at least one of multiple regression analysis on the resistance value characteristic parameter to determine whether welding is good. Determining whether or not the welding is defective.
[0008]
Further, the object is to measure the voltage between the welding electrodes during welding current conduction, the inter-electrode voltage measurement means, to measure the current between welding electrodes during welding current conduction, between the electrodes, Resistance value calculation means for calculating a resistance value at predetermined time intervals during the energization from a voltage value measured by the inter-electrode voltage measurement means and a current value measured by the inter-electrode current measurement means; and Resistance value parameter generation means for generating a resistance value characteristic parameter indicating a nugget generation process from the resistance value calculated by the calculation means, and at least one of discriminant analysis and multiple regression analysis on the resistance value characteristic parameter. The quality of the welding is determined by the welding quality judging device, which comprises a good or bad judgment means for judging whether the welding is good or bad.
[0009]
【The invention's effect】
According to the present invention, the quality of welding is determined based on the resistance value calculated from the voltage and current between the electrodes measured during energization, so that the tip position of the electrode tip fluctuates due to factors other than thermal expansion. Even in such a case, the welding quality can be determined more accurately.
[0010]
BEST MODE FOR CARRYING OUT THE INVENTION
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings. In this embodiment, a case where a general mild steel plate is spot-welded using a single-phase AC spot welding apparatus will be exemplified.
[0011]
FIG. 1 is a schematic diagram showing a configuration of an entire system including a welding gun for performing spot welding according to the present invention, a control device thereof, and a welding quality determination device, and FIG. 2 is a schematic configuration diagram of the welding quality determination device. 3 and 4 are flowcharts showing the procedure for determining the welding quality.
[0012]
First, the system comprises a welding gun 1, a welding gun control device 2 for controlling the operation of the welding gun 1, and a welding quality judging device 3 connected to the welding gun control device 2 for judging welding quality. Is done.
[0013]
As is well known, the welding gun 1 has a fixed electrode 12 and a movable electrode 13 attached to a welding gun arm 11.
[0014]
The fixed electrode 12 is fixed to a lower part of the welding gun arm 11. On the other hand, the movable electrode 13 is mounted on the welding gun arm 11 so as to be able to move up and down by a ball screw mechanism 14, and a servo motor 15 provided above the welding gun arm 11 rotates the ball screw to move up and down. I do. In addition, the servomotor 15 is provided with an encoder 18 and outputs the rotation amount.
[0015]
The tip of each of the fixed electrode 12 and the movable electrode 13 has an electrode tip 17 fitted into a shank 16 provided on the welding gun arm 11.
[0016]
A voltage measuring probe 51 for measuring a voltage is attached to each of the fixed electrode 12 and the movable electrode 13.
[0017]
A current measuring coil 52 for measuring a welding current is wound around the outer periphery of the welding gun arm 11 on the fixed electrode 12 side. Thereby, the current flowing through a power line cable (not shown) for flowing the welding current flowing inside the welding gun arm 11 is measured. When a power line cable for supplying a welding current is passed through a portion other than the welding gun arm 11, it is necessary to wind a current measuring coil 52 around the power line cable. It is good to change the installation position of the use coil 52.
[0018]
The welding gun arm 11 itself is adapted to be attached to, for example, the robot arm 101 or the like.
[0019]
The welding gun control device 2 includes a data processing device 21, an arithmetic device 22, and a storage device 25.
[0020]
The data processing device 21 receives movable electrode position data, electrode voltage data, and welding current data from the encoder 18, the voltage measurement probe 51, and the current measurement coil 52. The data processing device 21 outputs these data to the arithmetic device 22 and the storage device 25, and also outputs the data to the welding quality determination device 3.
[0021]
In the movable electrode position data, the motor rotation amount is obtained from the number of pulse output signals of the encoder 18 and the movement amount of the movable electrode 13 is calculated from the screw pitch of the ball screw mechanism 14 according to the motor rotation amount. The electrode voltage data is a voltage value between the fixed electrode 12 and the movable electrode 13 measured by the voltage measuring probe 51. At this point, the welding current data is simply the voltage and current values induced in the current measuring coil 52.
[0022]
The arithmetic unit 22 controls the opening / closing operation of the welding gun 1 based on the data from the data processing unit 21, the welding conditions stored in the storage unit 25, and the welding sequence of the welding quality determination unit 3 described later.
[0023]
The storage device 25 stores each data from the data processing device 21 and also stores welding conditions such as a pressing force and a voltage value to be applied during welding.
[0024]
The welding quality determination device 3 includes a current value measurement unit 31, a voltage value measurement unit 32, an expansion amount measurement unit 33, a welding sequence (time management) unit 34, a resistance value calculation unit 35, a feature parameter generation unit 36, and a discriminant function calculation unit. 37, a discriminant analysis threshold calculator 38, a multiple regression equation calculator 39, a discriminant analyzer 40, a multiple regression analyzer 41, a memory 42, and a clock generator 43.
[0025]
The current value measurement unit 31 changes from moment to moment during welding from welding current data (the voltage value and current value induced in the current measurement coil 52, and the like in the following) from the data processing device 21, and the current during welding actually flows. Calculate the value of the inter-electrode current. That is, the voltage and current induced in the current measurement coil 52, which is welding current data, are used as the secondary current of the transformer, and the power cable of the welding gun 1 is used as the primary side of the transformer. The value of the current flowing through the power cable on the primary side is calculated from the values of the voltage and the current induced in 52. The current value measurement unit 31 and the current measurement coil 52 constitute an inter-electrode current measurement unit.
[0026]
The voltage value measuring unit 32 calculates the value of the inter-electrode voltage that actually flows during welding and changes every moment during welding from the welding voltage data from the data processing device 21. The voltage value measuring section 32 and the voltage measuring probe constitute an inter-electrode voltage measuring means.
[0027]
The expansion amount measurement unit 33 calculates how the electrodes of the welding gun 1 move during the nugget formation process from the movable electrode position data (movement amount of the movable electrode 13, the same applies hereinafter) from the data processing device 21. I have. In other words, when the supply of the welding current starts, the base material in the portion sandwiched between the electrodes expands, and when the supply of the welding current ends, the base material in that portion contracts.・ Move up and down according to contraction. The expansion amount measuring section 33 calculates this per unit time from the movable electrode position data and stores it in the memory 42. The expansion amount measuring means and the encoder 18 constitute an expansion amount measuring unit 33.
[0028]
The welding sequence unit (time management) 34 performs time management based on the clock signal output from the clock generation unit 43. Specifically, the time from the start of the spot welding by the clock signal from the clock generation unit 43 (after the start of the welding current supply) is counted. The amount of thermal expansion is measured by the expansion amount measuring unit 33, and the time after the start of spot welding is added to the measured amount of thermal expansion. Therefore, the thermal expansion data stored in the memory 42 is data obtained by measuring the time based on the start of spot welding and the thermal expansion of the base material at that time for each unit time.
[0029]
The resistance value calculation unit 35 calculates a resistance value that changes momentarily during welding from the current value during welding obtained by the current value measurement unit 31 and the voltage value during welding obtained by the voltage slow bottom. . The calculated resistance value is stored in the memory 42 as resistance value data with a time after the spot welding is started by the clock signal from the clock generator 43 (after the welding current is started). This resistance value calculation unit 35 serves as resistance value calculation means.
[0030]
The feature parameter generation unit 36 calculates a heat approximation in the nugget generation process based on a linear approximation having a maximum slope, a linear approximation having a minimum slope, a quadratic curve approximation, and a time at which these approximations were obtained. An expansion amount characteristic parameter is generated, and a resistance value characteristic parameter is generated from the resistance value data. The characteristic parameter generation unit 36 functions as a resistance value characteristic parameter generation unit and a thermal expansion characteristic parameter generation unit.
[0031]
The discriminant function calculator 37 calculates a linear approximation for expansion change and a linear approximation for contraction change based on the thermal expansion measured by the expansion measuring unit 33 in real time or the thermal expansion data stored in the memory 42. , And a quadratic curve approximation formula is calculated. It is determined whether the electrode is moving in the up or down direction based on the moving amount of the electrode based on the movable electrode position data based on the movable electrode position data. The time at which the approximate expression was obtained is added to the calculated approximate expression. For example, when a certain approximate expression is obtained from expansion amount data centered on time t with respect to the start of spot welding, the time at which the approximate expression was obtained is t. The discriminant function calculator 37 functions as an approximate expression calculator.
[0032]
The discriminant analysis threshold calculator 38 calculates a discriminant analysis threshold for the resistance characteristic parameter and the thermal expansion characteristic parameter generated by the characteristic parameter generator 36. The discriminant analysis threshold value is a value at which the actually measured nugget diameter is 0 and the discriminant analysis score is the maximum, or the discriminant analysis threshold value is at least the non-defective product standard value and This is the value that minimizes the score. The discriminant analysis threshold calculator 38 functions as discriminant analysis threshold calculator.
[0033]
The multiple regression equation calculation unit 39 calculates a multiple regression equation relating to the expansion change based on the thermal expansion amount measured in real time by the expansion amount measuring unit 33 or the thermal expansion amount data stored in the memory 42.
[0034]
The discriminant analysis unit 40 performs discriminant analysis on the resistance value characteristic parameter and the thermal expansion characteristic parameter based on the discriminant analysis threshold value calculated by the discriminant analysis threshold value calculation unit 38, and finally outputs the result of the welding quality determination. I do. The discriminant analysis unit 40 functions as discriminant analysis means.
[0035]
The multiple regression analysis unit 41 performs multiple regression analysis on only good data among the data obtained as a result of the discriminant analysis. The multiple regression analysis unit 41 functions as multiple regression analysis means.
[0036]
The memory 42 is generated by the thermal expansion amount data measured by the expansion amount measuring unit 33, the resistance value data calculated by the resistance value calculating unit 35, the time obtained by the welding sequence unit 34, and the characteristic parameter generating unit 36. The expansion amount characteristic parameter and the resistance value characteristic parameter, the discriminant analysis threshold calculated by the discriminant analysis threshold calculator 38, the approximate expression calculated by the discriminant function calculator 37, and the welding quality discrimination result by the discriminant analyzer 40 are stored. I do. This memory 42 functions as storage means.
[0037]
The discrimination analysis unit 40, the multiple regression analysis unit 41, and the discrimination analysis threshold value calculation unit 38 constitute a welding quality determination unit. The welding quality determination unit determines that the welding quality is good based on the result of the finally performed multiple regression analysis. Or bad. The welding quality determining means performs good or bad welding quality by performing at least one of discriminant analysis and multiple regression analysis on the generated resistance value characteristic parameter and thermal expansion characteristic parameter. May be determined.
[0038]
The clock generation unit 43 supplies a clock signal to all components constituting the welding quality determination device 3. All the components including the above-mentioned expansion measuring section 33 operate based on this clock signal.
[0039]
Next, the welding quality judgment method according to the present invention will be described with reference to the flowcharts of the welding quality judgment procedures shown in FIGS.
[0040]
First, a predetermined portion of the base material is pressed by the upper and lower electrodes of the welding gun 1. When energization of the welding current is started in this state, the welding sequence unit 34 sets the time n to 0 in order to measure the time based on the welding start time, and sets a predetermined time St for sampling (for example, half an AC). (Wave time) is inserted (S1, S2). The predetermined waiting time is for sampling the thermal expansion amount and the resistance value, which will be described later, at regular intervals.
[0041]
Simultaneously with the start of energization, the voltage measurement unit measures and stores the voltage between the electrodes (S3), the current measurement unit measures and stores the current between the electrodes (welding current) (S4), and the expansion amount measurement unit 33 determines the base material. Is measured and stored (S5).
[0042]
Subsequently, the discriminant function calculation unit 37 calculates a linear approximation equation based on the thermal expansion amount data accumulated in the memory 42 since the start of welding (S6).
[0043]
Subsequently, in the discriminant function calculating section 37, the slopes of the linear approximation formulas obtained so far from the start of the welding were compared, and the linear approximation formulas having the maximum and minimum gradients were obtained. It is stored in the memory 42 together with the time (S7 to S9). The above processing is performed for each unit time St until the energization of the welding current is completed (S10). Through the above processing, the thermal expansion amount data and the linear approximation formula when the base material expands are obtained.
[0044]
The thermal expansion of the base material in the processing so far will be described. FIG. 5 is a thermal expansion waveform diagram showing a change in the amount of thermal expansion of the base material.
[0045]
When the thermal expansion amount data stored in the memory 42 is arranged in time series, a thermal expansion waveform (from t = 0 to WeldEnd) as shown in FIG. 5 is generated. For this thermal expansion waveform, a linear approximation formula is obtained with an arbitrary time width Tw starting from each time Tn. Therefore, the linear approximation formula obtained during welding current conduction is obtained by subtracting the value Wn obtained by dividing the arbitrary time width Tw by the sampling unit time St from the total number N in which the thermal expansion amount is stored for each unit time St. n = N−Wn). The slopes of the n linear approximations are compared, and the time at which the maximum slope is observed during expansion and the approximate linear equation (H = a + bx in FIG. 5), and the minimum slope at the inflection point where the transition from expansion to contraction is observed. Time and its approximate linear expression (H = a in FIG. 5) ₀ + B ₀ x) is determined.
[0046]
After the end of the energization, the welding sequence unit 34 again inserts a waiting time of a predetermined time St for sampling (S11).
[0047]
Subsequently, the expansion amount measuring unit 33 continuously measures and stores the thermal expansion amount (negative thermal expansion amount) of the base material (S12). The measured amount of thermal expansion is stored in the memory 42 together with the time for each unit time St.
[0048]
Subsequently, the discriminant function calculating unit 37 calculates a linear approximation formula based on the thermal expansion data accumulated in the memory 42 up to the end of the current supply of the welding current (S13).
[0049]
Subsequently, the discriminant function calculation unit 37 compares the slopes of the linear approximation formulas obtained so far after the welding current has been supplied, and obtains a linear approximation formula having the largest gradient. It is stored in the memory 42 together with the time (S14 to S16).
[0050]
The above process is performed until the time (sampling) set as the time from the end of the supply of the welding current to the end of the contraction of the base material ends (S17).
[0051]
Through the above processing, the data of the amount of thermal expansion when the base material contracts and the linear approximation formula are obtained.
[0052]
The processing up to this point will be described with reference to the thermal expansion waveform diagram shown in FIG.
[0053]
When the thermal expansion amount data stored in the memory 42 is arranged in chronological order, a thermal expansion waveform (from t = WeldEnd to Hend) as shown in FIG. 5 is generated. For this thermal expansion waveform, a linear approximation formula is obtained with an arbitrary time width Tw starting from each time Tn. Similarly to the above, the linear approximation formula obtained during welding current conduction is obtained by subtracting the value Wn obtained by dividing the arbitrary time width Tw by the sampling unit time St from the total number N in which the thermal expansion amount is stored for each unit time St. The number is n (n = N−Wn). By comparing the slopes of the n linear approximation formulas, a time at which the maximum gradient is observed at the time of expansion and the approximate straight-line formula (H = a '+ b'x in FIG. 5) are obtained.
[0054]
Thereafter, the discriminant function calculating unit 37 calculates the quadratic curve approximation formula (H = a "+ b" in FIG. 5) based on the thermal expansion amount data accumulated in the memory 42 since the welding current was supplied. x + c ″ x2) is calculated and stored in the memory 42 (S18).
[0055]
Subsequently, the resistance value calculation unit 35 calculates the resistance value from the start and end of energization (from t = 0 to WeldEnd) from the voltage value and the current value stored in the previous steps S3 and S4 (S19). ).
[0056]
Here, the resistance change during energization will be described. FIG. 6 is a resistance waveform diagram showing a change in resistance value during energization measured as described above.
[0057]
Here, the resistance value measurement interval, that is, the unit time St, calculates the interelectrode resistance value for each AC half-wave of the welding current.
[0058]
FIG. 6 is a resistance waveform in which the calculated resistance value data is arranged in time series.
[0059]
First, the resistance immediately after the start of energization is Rst, and the reciprocal thereof is Rst. ^-1 And The maximum resistance value during energization is Rpeak, and the elapsed time from the start of energization (t = 0) to Rpeak is Tpeak. Further, a predetermined period of time from the start of energization is T1, the end time of energization is WeldEnd, and the average value of resistance between T1 and WeldEnd is Rave. The resistance immediately before the end of energization is Rfin, and the reciprocal thereof is Rfin. ^-1 And
[0060]
Generally, the following matters are known in spot welding. The calorific value of the melted portion causing the nugget formation is largely affected by the current density in the base material. Since the welding current is kept substantially constant in the constant current control, the current density is determined by the area of the conducting path between the plates as the members to be welded. Further, the area of the current path has a strong correlation with the contact area between the plates.
[0061]
For example, when there is a gap between the base materials, when the welding current is started, the welded portion is softened by heat generation, and a so-called “fit-in” occurs, in which the joint of the plates is improved by plastic deformation. In this familiarization, the contact area between the plates and the amount of thermal expansion behave differently as compared to the case where there is no gap between the base materials.
[0062]
FIG. 7 is a graph showing the pressurized state during welding when there is no plate gap and the amount of thermal expansion due to a signal from the encoder 18 at that time. FIG. 6 is a graph showing a thermal expansion amount based on a signal from the encoder 18 at that time. 7 and 8 show the case where the same nugget diameter is formed.
[0063]
In the case where there is no gap as shown in FIG. 7A, almost all of the measured thermal expansion is due to the thermal expansion of the base material as shown in FIG. 7B. On the other hand, when there is a gap as shown in FIG. 8A, as shown in FIG. 8B, the amount of thermal expansion measured by deformation due to familiarity of the gap decreases, and accurate heat The amount of expansion cannot be accurately estimated only by the signal from the encoder 18. Further, in the measurement of the amount of thermal expansion, even when the electrode tip 17 is sunk into the shank 16 due to the impact during welding or during welding, the same as shown in FIG. In addition, the measurement tail of the amount of thermal expansion corresponding to the dive into the shank 16 differs.
[0064]
Therefore, in the welding quality judgment / estimation using only the thermal expansion amount, when there is a gap between the base materials, the information on the difference due to the familiarity is missing, and the correlation between the thermal expansion amount waveform and the nugget formation situation is reduced. Also, the correlation between the waveform of the thermal expansion amount and the state of the nugget formation is reduced even when the electrode tip 17 enters the shank 16. Therefore, the correlation between the explanatory variable based on the thermal expansion amount characteristic parameter obtained from the thermal expansion amount and the nugget quality or the nugget diameter, which is the objective variable, also decreases, and the judgment and estimation accuracy of discriminant analysis and multiple regression analysis deteriorate.
[0065]
In the present embodiment, by further using the resistance value characteristic parameter obtained from the resistance waveform, the correlation between the explanatory variable based on the thermal expansion amount characteristic parameter obtained from the thermal expansion amount and the nugget pass / fail or the nugget diameter as the objective variable. It is to compensate for the decline.
[0066]
That is, by using the resistance value characteristic parameter obtained from the resistance waveform, the information on the contact area and the information on the direct nugget formation are included, whereby the information on the familiarity and the difference due to the electrode tip 17 slipping into the shank 16 is included. Is supplemented.
[0067]
In spot welding, resistance R between electrodes = volume resistance + contact resistance. However, the contact resistance almost disappears at 0.5 CYC (10 msec) immediately after the start of energization. Therefore, the resistance between the electrodes R ≒ ρ × L / S (where ρ: specific resistance, L: length of the current path, S: area of the current path).
[0068]
Here, since the specific resistance ρ has temperature dependency, ρ also changes in accordance with a change in the temperature of the base material due to energization. The length L of the current path is substantially equal to the plate thickness, and the area S of the current path increases with the softening and melting of the base material.
[0069]
At the beginning of energization, the temperature of the energizing path has not yet risen and is close to room temperature, so that the specific resistance ρ is substantially constant, and the area of the energizing path and the contact area have a strong correlation as described above. Therefore Rst, Rst ^-1 Is correlated with the initial contact area.
[0070]
The energization path area S gradually increases with the start of energization, but the effect of an increase in the specific resistance ρ due to a temperature rise due to a decrease in 1 / S is large, and as a result, the interelectrode resistance R increases. Then, when the temperature of the contact portion exceeds the melting temperature of the base material, the formation of the molten portion starts and the area S of the current path rapidly increases. However, in this region, the specific resistance ρ is almost constant. To decrease.
[0071]
Thus, the peak timing of the inter-electrode resistance R is close to the nugget formation start timing. Therefore, Tpeak is considered to have a direct correlation with the nugget formation speed irrespective of the presence or absence of a plate gap or the penetration of the electrode tip 17 into the shank 16.
[0072]
On the other hand, after the inter-electrode resistance R has passed the peak Rpeak, the melted portion becomes liquid and the temperature change becomes small, so that the specific resistance ρ and the length L of the current path become almost constant. Therefore, the current path area S can be estimated from the interelectrode resistance R. Further, there is a strong correlation between the current path area S and the area of the fusion zone, that is, the nugget diameter. Therefore, Rave, Rfin ^-1 Has a correlation with the nugget diameter at the end of energization.
[0073]
Therefore, after the resistance value is calculated, the characteristic parameter generation unit 36 firstly obtains the time (T in FIG. 5) at which the maximum slope was obtained, obtained from the thermal expansion amount data at the time of expansion, and the approximate linear equation (FIG. 5, H = a + bx), the time at which the minimum slope was observed (T in FIG. 5), the approximate linear equation (H = a0 + b0x in FIG. 5), and the maximum slope obtained from the thermal expansion amount data during contraction. Observed time (T in FIG. 5), approximate linear equation (H = a ′ + b′x in FIG. 5), quadratic curve approximate equation (H = a ″ + b ″ x + c ″ x in FIG. 5) ² ), Thermal expansion characteristic parameters due to thermal expansion are generated (S20 and S21), and the resistance value (Rst in FIG. 6) immediately after the start of electricity generation and its reciprocal (Rst in FIG. 6) ^-1 ), The maximum value of the resistance during energization (Rpeak in FIG. 6), the elapsed time from the start of energization to this Rpeak (Tpeak in FIG. 6), and a predetermined fixed time from the start of energization as T1, and the resistance between T1 and WeldEnd. The average value (Rave in FIG. 6), the resistance value immediately before the end of energization (Rfin in FIG. 6), and the reciprocal thereof (Rfin in FIG. 6) ^-1 ) Is generated as a resistance value characteristic parameter obtained from the resistance measurement (S20). Note that T1 may be an arbitrary time as a predetermined time from the start of energization, but is a predetermined time from Rpeak to the end of energization.
[0074]
Subsequently, the discriminant analysis threshold calculator 38 calculates a discriminant analysis threshold for the resistance characteristic parameter and the thermal expansion characteristic parameter generated by the characteristic parameter generator 36 (S22). The discriminant analysis threshold value is a value at which the actually measured nugget diameter is 0 and the discriminant analysis score is the maximum, or the discriminant analysis threshold value is at least the non-defective product standard value and This is the value that minimizes the score.
[0075]
Subsequently, the discriminant analysis unit 40 performs discriminant analysis on the resistance value characteristic parameter and the thermal expansion characteristic parameter based on the discriminant analysis threshold calculated by the discriminant analysis threshold calculator 38. By performing this discriminant analysis, defective data is excluded from the thermal expansion data and the resistance value data of the non-linear portion (S23). Therefore, only good data necessary for accurate determination of welding quality is extracted (S24).
[0076]
Subsequently, the multiple regression equation calculation unit 39 calculates a multiple regression equation only for the good data obtained as a result of the discriminant analysis (S25).
[0077]
Finally, the multiple regression analysis unit 41 performs multiple regression analysis that is finally performed using good data (S26), and estimates the nugget diameter from the result (S27), and determines whether the welding quality is good or poor. Is determined (S28).
[0078]
Here, there are three methods for determining the pass / fail section.
First, a thermal expansion amount characteristic parameter obtained from a thermal expansion waveform is generated, and Rst, Rst ^-1 , Tpeak, Rave, Rfin ^-1 Is used as an explanatory variable to perform discriminant analysis to determine the quality of the nugget diameter.
[0079]
Second, a nugget diameter is estimated by multiple regression analysis using characteristic parameters obtained from thermal expansion and resistance waveforms as explanatory variables.
[0080]
Third, discriminant analysis is performed using the characteristic parameters obtained from the thermal expansion and the resistance waveform as explanatory variables. At this time, the value at which the actual nugget diameter is equal to or greater than the non-defective product standard value and the discriminant analysis score is minimized is determined by discriminant analysis It is a threshold. As a result, the nugget diameter is estimated by multiple regression analysis using the characteristic parameters obtained from the thermal expansion and resistance waveforms as explanatory variables for only the samples determined to be non-defective. In this way, by performing a discriminant analysis and then calculating an expression for estimating the nugget diameter based on a statistical method such as multiple regression analysis for only good data, the actual welding nugget diameter and the estimated nugget diameter have a linear relationship. Since the feature for judging the nugget diameter can be grasped with high accuracy, a high correlation is exhibited, and the accuracy of judging the quality of welding quality is greatly improved.
[0081]
In addition, by using only the extracted good data and using the feature parameter as the target variable for multiple regression analysis using the nugget diameter as the target variable, it is possible to estimate the nugget diameter from the results of multiple regression analysis. It becomes.
[0082]
As described above, in the present embodiment, the resistance value characteristic parameter obtained from the resistance waveform is used as an explanatory variable in addition to the thermal expansion amount characteristic parameter obtained from thermal expansion, so that there is a gap between the base materials. Even if the electrode tip 17 is sunk into the shank 16 in the case or the shank 16, the method using only the thermal expansion characteristic parameter by thermal expansion is compensated and determined by the resistance characteristic parameter based on the resistance value with respect to the nugget formation state. The correlation between the explanatory variable and the objective variable in the analysis and the multiple regression analysis is improved, and the accuracy of nugget pass / fail determination and the accuracy of nugget diameter estimation are improved.
[0083]
Next, the results of actual welding of a base material with and without a plate gap and an experiment of estimating a nugget diameter and determining whether welding is good or bad according to the present embodiment will be described.
[0084]
FIG. 9 is a resistance waveform diagram obtained by actually welding a base material with and without a plate gap.
[0085]
As shown in FIG. 9, irrespective of the presence or absence of a plate gap, those having a good nugget (OK in the drawing) and those having a bad nugget (NG in the drawing) are among the resistance value characteristic parameters, in particular, after a predetermined time T1. Rave of the resistance value and the reciprocal Rfin of the resistance value immediately before the end of energization ^-1 Is at a high position on the graph, and it can be determined that the non-defective product is defective.
[0086]
In FIG. 9, although there are only four data, non-defective data are extracted from a plurality of actually welded results according to the above-described processing procedure, and the quality of the nugget is determined based on the feature parameters by using a statistical method. Will be able to do it.
[0087]
As described above, according to the present embodiment, not only the amount of thermal expansion of the base material based on the amount of displacement between the electrodes during welding, but also whether or not the nugget is good or bad based on the resistance value during welding energization is determined. Therefore, in addition to the effect of the gap at the welded part, the displacement of the electrode tip is caused by factors other than thermal expansion, such as the electrode tip sunk into the shank, and the impact applied to the welding gun or electrode tip during other welding. Even in the case of fluctuation, it is possible to compensate for these effects and more accurately determine the welding quality.
[0088]
This embodiment described above does not limit the present invention. For example, as is clear from the above-described embodiment, welding is performed only by measuring the resistance value without measuring the thermal expansion amount. The quality may be determined. Further, the method may be combined with various nugget determination and estimation methods other than the measurement of the thermal expansion amount.
[0089]
Further, in the above-described embodiment, the control device for welding control and the welding quality determination device for welding determination have been described as separate devices, but these may be integrated.
[0090]
Furthermore, it goes without saying that various modifications can be made by those skilled in the art.
[Brief description of the drawings]
FIG. 1 is a schematic diagram showing a configuration of an entire system including a welding gun for performing spot welding according to the present invention, a control device thereof, and a welding quality determination device.
FIG. 2 is a schematic configuration diagram of a welding quality determination device.
FIG. 3 is a flowchart showing a welding quality determination procedure.
FIG. 4 is a flowchart showing a welding quality determination procedure.
FIG. 5 is a thermal expansion waveform diagram showing a change in a thermal expansion amount of a base material.
FIG. 6 is a resistance waveform diagram showing a change in resistance value during energization.
FIG. 7 is a graph showing a pressurized state during welding when there is no plate gap and a thermal expansion amount based on a signal from the encoder 18 at that time.
FIG. 8 is a graph showing a pressurized state during welding when there is a plate gap and a thermal expansion amount based on a signal from the encoder 18 at that time.
FIG. 9 is a resistance waveform diagram in which a base material with and without a plate gap is actually welded.
[Explanation of symbols]
1 welding gun
2 Welding gun control device
3 Welding quality judgment device
11 Welding gun arm
12 fixed electrodes
13 movable electrode
14 Ball screw mechanism
15 Servo motor
15 Shank
17 electrode tip
21 Data processing device
22 Arithmetic unit
25 Storage device
31 Current value measurement section
32 Voltage measurement section
33 Expansion measurement section
34 welding sequence part
35 Resistance calculator
36 Feature Parameter Generator
37 Discriminant function calculator
38 discriminant analysis threshold calculator
39 Multiple regression equation calculator
40 Discriminant analysis unit
41 Multiple regression analysis
42 memory
43 Clock generator
51 Probe for voltage measurement
52 Current measurement coil

Claims (9)

Measuring the voltage and current between the welding electrodes at predetermined time intervals during welding current conduction,
Calculating a resistance value for each predetermined time interval from the obtained voltage and current values;
Generating a resistance value feature parameter indicating a nugget generation process from the resistance value;
For the resistance value feature parameter, discriminant analysis, or a step of determining whether the welding is good or bad by performing at least one analysis of multiple regression analysis,
A method for determining welding quality, comprising: Determining whether the welding is good or bad,
Performing a discriminant analysis on the resistance characteristic parameter;
Extracting only good data from the data obtained as a result of the discriminant analysis;
Only for the extracted good data, a step of obtaining quantity data for performing multiple regression analysis using the nugget diameter as an objective variable,
Estimating the nugget diameter from the result of the multiple regression analysis,
The welding quality determination method according to claim 1, further comprising: A first step of obtaining a linear approximation formula for a change in expansion of the base material from a thermal expansion amount of the base material per unit time obtained during welding current conduction;
A second step of obtaining a straight line approximation expression having a maximum and a minimum slope from the linear approximation expressions obtained in the first step;
A third step of obtaining a linear approximation formula for a change in shrinkage of the base material from the amount of thermal expansion of the base material per unit time obtained after the welding current is applied;
A fourth step of obtaining a straight-line approximation formula having a maximum slope among the straight-line approximation formulas obtained in the third step;
A fifth step of calculating a quadratic curve approximation formula for a change in shrinkage of the base material from the amount of thermal expansion of the base material after the welding current is applied;
A sixth step of determining the time at which each of the linear approximation formula having the largest slope, the linear approximation formula having the smallest slope, and the quadratic curve approximation formula is obtained;
A linear approximation formula having the largest slope, a linear approximation formula having the smallest slope, the quadratic curve approximation formula, and a thermal expansion characteristic parameter indicating a nugget generation process based on the time when these approximation formulas were obtained. And a seventh step of generating
Determining whether the welding is good or bad, using the thermal expansion characteristic parameter together with the resistance characteristic parameter, by performing at least one analysis of discriminant analysis, or multiple regression analysis, The method according to claim 1, wherein it is determined whether the welding is good or bad. The resistance value characteristic parameter is:
The resistance value immediately after the start of energization, the reciprocal of the resistance value immediately after the start of energization, the maximum resistance during energization, the elapsed time from the start of energization to the maximum value of the resistance, and the elapse of a predetermined time from the start of energization to the end of energization The welding quality determination method according to any one of claims 1 to 3, wherein the average value of the resistance values, the resistance value immediately before the end of the energization, and the reciprocal of the resistance value immediately before the end of the energization. Inter-electrode voltage measuring means for measuring the voltage between the welding electrodes during welding current flow,
An inter-electrode current measuring means for measuring a current between welding electrodes during welding current conduction, and a voltage value measured by the inter-electrode voltage measuring means and a current value measured by the inter-electrode current measuring means, Resistance value calculating means for calculating a resistance value at predetermined time intervals during energization,
Resistance value characteristic parameter generation means for generating a resistance value characteristic parameter indicating a nugget generation process from the resistance value calculated by the resistance value calculation means,
For the resistance value feature parameter, discriminant analysis, or good or bad judgment means to judge whether welding is good or bad by performing at least one analysis of multiple regression analysis,
A welding quality judging device characterized by having: The welding quality determination means,
Discriminant analysis means for performing discriminant analysis on the resistance value feature parameter,
Having multiple regression analysis means for performing multiple regression analysis only on good data among data obtained as a result of the discriminant analysis by the discriminant analysis means,
The welding quality determination device according to claim 5, wherein it is determined whether the welding quality is good or bad from the result of the multiple regression analysis. The welding quality determination means,
Further comprising a discriminant analysis threshold calculating means for calculating a discriminant analysis threshold for the resistance value feature parameter,
The welding quality judging device according to claim 6, wherein the discriminant analyzing means performs discriminant analysis on the resistance value characteristic parameter generated based on the calculated discriminant analysis threshold. Expansion amount measuring means for measuring the amount of thermal expansion of the base material during the nugget formation process,
Storage means for storing the measured amount of thermal expansion;
From the thermal expansion amount measured by the expansion amount measuring means or the thermal expansion amount stored in the storage means, a linear approximation formula and a quadratic curve approximation formula for a change in expansion and contraction of the base material per unit time are calculated. Approximation formula calculating means,
An approximate expression inclination determining means for determining an approximate expression having a maximum and a minimum inclination in the calculated linear approximate expression,
Based on the linear approximation formula having the largest slope, the linear approximation formula having the smallest slope, the quadratic curve approximation formula, and the time at which these approximation formulas were obtained, the thermal expansion characteristic parameter in the nugget generation process is generated. Thermal expansion characteristic parameter generating means,
6. The welding quality judgment according to claim 5, wherein the welding quality judgment means judges whether the welding quality is good or bad by using the thermal expansion characteristic parameter together with the resistance value characteristic parameter. apparatus. The resistance value characteristic parameter is:
The resistance value immediately after the start of energization, the reciprocal of the resistance value immediately after the start of energization, the maximum resistance during energization, the elapsed time from the start of energization to the maximum value of the resistance, and the elapse of a predetermined time from the start of energization to the end of energization The welding quality determination device according to any one of claims 5 to 8, wherein the average value of the resistance values, the resistance value immediately before the end of the energization, and the reciprocal of the resistance value immediately before the end of the energization.

Images