Electrode misplacement detection method and system based on double-pole hall effect sensor

Abstract

The application discloses an electrode misplacement detection method and system based on a double-magnetic-pole Hall effect sensor, and the method comprises the following steps: eliminating the misadjustment voltage and noise of the Hall effect sensor by a rotating current misadjustment elimination method and a chopping stabilization technology elimination method, so as to obtain a double-magnetic-pole Hall effect sensor; collecting and processing the Hall voltage of the electrode based on the double-magnetic-pole Hall effect sensor, so as to obtain the contact area change amount of the electrode tip; extracting and processing the contact area change amount of the electrode tip by a neural network model, determining the misplacement amount of the electrode, comparing the misplacement amount with a standard threshold misplacement amount, and obtaining an electrode misplacement detection result. The application can eliminate the misadjustment voltage and noise, improve the precision of the magnetic field measurement, and thus improve the detection precision of the electrode misplacement. The application can be widely applied to the technical field of electrode misplacement detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electrode misalignment detection, and particularly relates to an electrode misalignment detection method and system based on a double-magnetic-pole Hall effect sensor. BACKGROUND

[0002] Resistance spot welding is a metal joining technique widely used in the automotive industry, favored for its high joint quality, high production efficiency, few auxiliary processes, and ease of automation. In the resistance spot welding process, the centering state of the electrode head has a significant impact on the mechanical properties of the welded joint. If the two electrode heads are not on the same straight line, the area through which the current passes during welding becomes smaller, and the size of the nugget formed also decreases, resulting in a decrease in the mechanical properties of the welded joint. When there is a certain angular offset between the axes of the two electrode heads, the contact between the electrode head and the material to be welded has a sequence, and the part that first contacts the electrode head melts and causes the indentation depth of that part to be deeper. When the welding current is too large, this part is more prone to spatter, resulting in less metal inside the final nugget and the formation of defects such as porosity and shrinkage. The electrode indentation of the welded joint is too deep, which will cause stress concentration in that part and reduce the strength and toughness of the welded joint. When the electrode head axes are on the same straight line, i.e., centered, the plastic ring is well sealed during the welding process, and the molten pool metal is not easily ejected, avoiding the generation of welding spatter, resulting in a good appearance of the welded joint and a perfect symmetry of the nugget cross-section, and the mechanical properties of the welded joint are better, with improved strength and toughness. Therefore, the centering accuracy of the resistance spot welding electrode head directly determines the welding quality of the welded joint.

[0003] In actual production environments, due to mechanical fatigue or poor adjustment of the welding gun, the electrode head may exhibit lateral deviation, which can lead to instability during welding and a decrease in welding quality. First, traditional detection methods are mostly destructive experiments after welding, which cannot monitor the lateral deviation of the electrode head in real time during welding. This means that electrode deviation problems cannot be discovered and adjusted in a timely manner before welding is completed, resulting in defects that may occur during welding that cannot be corrected in a timely manner. Second, due to the lack of real-time monitoring means, even if electrode deviation problems are detected, they cannot be adjusted in a timely manner during welding. This can lead to the generation of welding defects, affecting the mechanical properties of the welded joint, and increasing the workload and cost of rework and maintenance. Traditional detection methods may not accurately detect minor electrode deviations, resulting in some potential quality issues being overlooked. These problems not only affect the quality of welding, but also can have a negative impact on subsequent production processes and product quality. SUMMARY

[0004] To solve the above technical problems, the purpose of the present application is to provide an electrode misalignment detection method and system based on a double-magnetic-pole Hall effect sensor, which can improve the accuracy of magnetic field measurement by eliminating the offset voltage and noise, thereby improving the detection accuracy of electrode misalignment.

[0005] The first technical solution adopted by the application is an electrode misplacement detection method based on a double-magnetic-pole Hall effect sensor, including the following steps:

[0006] The Hall effect sensor is subjected to a rotating current offset elimination method and a chopping stabilization technology elimination method to eliminate offset voltage and noise, and a double-magnetic-pole Hall effect sensor is obtained.

[0007] The electrode is subjected to Hall voltage collection and processing based on the double-magnetic-pole Hall effect sensor, and the contact area change amount of the electrode tip is obtained.

[0008] The contact area change amount of the electrode tip is subjected to feature extraction processing through a neural network model, the misplacement amount of the electrode is determined and compared with a standard threshold misplacement amount, and an electrode misplacement detection result is obtained.

[0009] Further, the double-magnetic-pole Hall effect sensor includes a rotating current source, a Hall element, a modulation circuit, a chopping amplification circuit, a demodulation circuit, a comparison circuit, and an output control circuit, which are connected in sequence.

[0010] Further, the step of obtaining the double-magnetic-pole Hall effect sensor by subjecting the Hall effect sensor to a rotating current offset elimination method and a chopping stabilization technology elimination method to eliminate offset voltage and noise specifically includes:

[0011] Based on the rotating current source, the Hall effect sensor is subjected to offset voltage elimination processing through the rotating current offset elimination method to obtain a Hall effect sensor after offset voltage elimination.

[0012] Based on the chopping amplification circuit, the Hall effect sensor is subjected to offset voltage noise elimination processing through the chopping stabilization technology elimination method to obtain a Hall effect sensor after offset voltage noise elimination.

[0013] The Hall effect sensor after offset voltage elimination and the Hall effect sensor after offset voltage noise elimination are combined to obtain the double-magnetic-pole Hall effect sensor.

[0014] Further, the step of obtaining the Hall effect sensor after offset voltage elimination by subjecting the Hall effect sensor to offset voltage elimination processing through the rotating current offset elimination method based on the rotating current source specifically includes:

[0015] The Hall element bias voltage of the vertical phase and the Hall element bias voltage of the horizontal phase are generated through the rotating current source.

[0016] Based on the vertical phase Hall element bias voltage, the power supply current of the Hall element flows from top to bottom, and the vertical phase output voltage of the Hall element is obtained;

[0017] Based on the horizontal phase Hall element bias voltage, the power supply current of the Hall element flows from left to right, and the horizontal phase output voltage of the Hall element is obtained;

[0018] Superimpose the vertical phase output voltage of the Hall element and the horizontal phase output voltage of the Hall element and take the average to obtain the Hall effect sensor after eliminating the offset voltage.

[0019] Further, the chopper amplification circuit based on the chopper stabilization technique eliminates the offset voltage noise of the Hall effect sensor, and the step of obtaining the Hall effect sensor after eliminating the offset voltage noise, which specifically includes:

[0020] The Hall effect sensor is powered on to obtain the Hall voltage to be processed;

[0021] Based on the chopper amplification circuit, generate a high-frequency square wave modulation signal to modulate the Hall voltage to be processed in the high-frequency band to obtain a high-frequency modulation signal;

[0022] Superimpose the high-frequency modulation signal and the offset voltage and noise of the Hall effect sensor to obtain a mixed signal;

[0023] After amplifying the mixed signal, demodulate it through the high-frequency square wave modulation signal to obtain a baseband signal;

[0024] Filter the high-frequency noise and offset voltage of the baseband signal to obtain the Hall effect sensor after eliminating the offset voltage noise.

[0025] Further, the step of the double-magnetic-pole Hall effect sensor collecting and processing the Hall voltage of the electrode to obtain the contact area change of the electrode tip specifically includes:

[0026] The electrode is powered on to obtain the current density of the electrode;

[0027] According to the Ampere loop law, determine the relationship between the electrode magnetic field strength and the current density of the electrode;

[0028] Collect and process the Hall voltage of the electrode through the double-magnetic-pole Hall effect sensor to obtain the Hall voltage of the electrode;

[0029] Combine the Hall voltage of the electrode and the relationship between the electrode magnetic field strength and the current density of the electrode to calculate the contact area of the electrode to obtain the contact area change of the electrode tip.

[0030] Further, the expression of the electrode contact area calculation is specifically as follows:

[0031]

[0032] In the above formula, V H represents the Hall voltage of the electrode, K represents the sensitivity coefficient, k represents the proportional coefficient, C a represents the contact area of the electrode tip, and I represents the current size passing through the electrode.

[0033] Further, the step of extracting features of the contact area change of the electrode tip by the neural network model, determining the misalignment amount of the electrode and comparing it with the standard threshold misalignment amount to obtain the electrode misalignment detection result, specifically includes:

[0034] extracting features of the contact area change of the electrode tip by the neural network model to obtain the mean and variance of the magnetic field strength;

[0035] determining the misalignment amount of the electrode according to the mean and variance of the magnetic field strength;

[0036] collecting and processing the Hall voltage of the standard electrode based on the double-magnetic-pole Hall effect sensor to obtain the misalignment amount of the standard electrode as the standard threshold misalignment amount;

[0037] If the misalignment amount of the electrode is greater than the standard threshold misalignment amount, the electrode currently has an axial misalignment problem, and an audible and light alarm is triggered;

[0038] If the misalignment amount of the electrode meets the range of the standard threshold misalignment amount, the electrode currently does not have an axial misalignment problem, and the electrode misalignment detection result is output.

[0039] The second technical solution adopted by the present application is: an electrode misalignment detection system based on a double-magnetic-pole Hall effect sensor, comprising:

[0040] A first module is used to eliminate the offset voltage and noise of the Hall effect sensor by a rotating current offset elimination method and a chopper stabilization technique elimination method to obtain a double-magnetic-pole Hall effect sensor;

[0041] A second module is used to collect and process the Hall voltage of the electrode based on the double-magnetic-pole Hall effect sensor to obtain the contact area change of the electrode tip;

[0042] A third module is used to extract features of the contact area change of the electrode tip by the neural network model, determine the misalignment amount of the electrode, and compare it with the standard threshold misalignment amount to obtain the electrode misalignment detection result.

[0043] The method and system have the advantages that the present application eliminates the offset voltage and noise of the Hall effect sensor through the rotating current offset elimination method and the chopping stabilization technology elimination method, obtains a double-magnetic-pole Hall effect sensor, effectively eliminates the offset voltage and noise by combining the rotating current offset elimination technology and the chopping stabilization technology, ensures more accurate output signals when the magnetic field changes, effectively reduces the offset and noise of the Hall element and the amplifier circuit by using the rotating current offset elimination technology and the chopping stabilization technology, thereby improving the precision of the magnetic field measurement, further collects and processes the Hall voltage of the electrode based on the double-magnetic-pole Hall effect sensor, obtains the contact area change amount of the electrode tip, finally extracts the features of the contact area change amount of the electrode tip through the neural network model, determines the misplacement amount of the electrode and compares it with the standard threshold misplacement amount, so that the electrode misplacement in the welding process can be monitored in real time, and timely and accurate data support is provided for the evaluation of the welding quality. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 is a step flow chart of the electrode misplacement detection method based on the double-magnetic-pole Hall effect sensor of the present application;

[0045] Figure 2 is a structural block diagram of the electrode misplacement detection system based on the double-magnetic-pole Hall effect sensor of the present application;

[0046] Figure 3 is a schematic diagram of the electrode alignment state provided by the specific embodiment of the present application;

[0047] Figure 4 is a schematic diagram of the reduction of the contact area due to the axial misalignment provided by the specific embodiment of the present application;

[0048] Figure 5 is a schematic diagram of the magnetic field generation principle provided by the specific embodiment of the present application;

[0049] Figure 6 is a schematic diagram of the Wen's bridge model provided by the specific embodiment of the present application;

[0050] Figure 7 is a schematic diagram of the principle of the rotating current offset elimination technology provided by the specific embodiment of the present application;

[0051] Figure 8 is a schematic diagram of the working process of the chopping stabilization technology provided by the specific embodiment of the present application;

[0052] Figure 9 is a schematic diagram of the principle of the chopping stabilization technology provided by the specific embodiment of the present application;

[0053] Figure 10 is a schematic diagram of the electrode misplacement detection process provided by the specific embodiment of the present application;

[0054] Figure 11 is a schematic diagram of a dual-magnetic-pole Hall effect sensor circuit framework provided by an embodiment of the present application;

[0055] Figure 12 is a schematic diagram of a dual-magnetic-pole Hall effect sensor package provided by an embodiment of the present application;

[0056] Figure 13 is a schematic diagram of an electrode centering and misalignment monitoring system device provided by an embodiment of the present application;

[0057] Figure 14 is a schematic diagram of an electrode centering and misalignment monitoring process provided by an embodiment of the present application;

[0058] Figure 15 is a schematic diagram of data collected during a transmission check provided by an embodiment of the present application. DETAILED DESCRIPTION

[0059] The present application will be further described below in conjunction with the accompanying drawings and specific embodiments. For the step numbers in the following embodiments, they are only set for the convenience of description, and the order between the steps is not limited in any way, and the execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0060] First of all, it needs to be pointed out that the Hall effect sensor is a magnetic field sensor that works on the principle of the Hall effect, which can convert magnetic field signals into electrical signals. With the advancement of semiconductor technology, its performance is constantly improving, and its application range is increasingly wide. The dual-magnetic-pole Hall effect sensor can detect the magnetic field of S and N two magnetic poles, has the characteristics of high sensitivity and high precision, and is suitable for application in occasions with high requirements for magnetic field detection. In the Hall sensor, the offset voltage and noise will affect the measurement accuracy. The development of rotating current offset elimination technology and chopping stabilization technology effectively reduces the offset voltage and noise of the sensor and improves the measurement accuracy. At the same time, with the continuous improvement of the power consumption requirements of electronic equipment, low-power design technology has also been widely applied in the field of sensors. Through optimizing the circuit design and working mode, the power consumption can be significantly reduced on the premise of ensuring the performance of the sensor, and used for detecting the electrode centering or misalignment state of resistance welding.

[0061] Based on this, the embodiment of the present application aims to provide a method and system for detecting electrode centering and misalignment in the resistance spot welding process. First, a double-magnetic-pole Hall effect sensor manufactured by COMS technology is designed, and a rotating current imbalance elimination technology and a chopper stabilization technology are introduced to eliminate imbalance voltage noise, so as to obtain a more accurate output signal when the magnetic field changes. Four Hall effect sensors are arranged in an anti-parallel layout to realize synchronous collection of magnetic field data on four Cartesian coordinate axes to comprehensively capture the change of the magnetic field. Secondly, a data acquisition and communication protocol is designed, and data is transmitted from the sensor to the PLC through the LoRaWAN communication protocol and then stored in the database, so as to realize efficient processing and analysis of the data. Finally, a decision tree model is introduced to realize real-time and non-destructive detection of electrode misalignment, timely evaluation of welding quality and setting of a threshold value through the decision tree model result, and an alarm is issued when the threshold value is exceeded, so as to timely adjust the welding parameters, improve the welding efficiency and reduce the production cost, promote the intelligent development of the welding technology and improve the stability of the welding process and the welding quality.

[0062] Reference Figure 1 The present application provides an electrode misalignment detection method based on a double-magnetic-pole Hall effect sensor, which comprises the following steps:

[0063] S100, imbalance voltage and noise of the Hall effect sensor are eliminated by a rotating current imbalance elimination method and a chopper stabilization technology, and a double-magnetic-pole Hall effect sensor is obtained.

[0064] First of all, it should be pointed out that the embodiment of the present application is based on the detection of the magnetic field change caused by the electrode offset, and a double-magnetic-pole Hall effect sensor manufactured by COMS technology is designed. The sensor eliminates imbalance voltage noise by introducing a rotating current imbalance elimination technology and a chopper stabilization technology to obtain a more accurate output signal when the magnetic field changes.

[0065] Further, it should be pointed out that the design adopts a double-magnetic-pole Hall effect sensor manufactured based on CMOS technology, and a method for detecting the electrode alignment state by measuring the magnetic field of the electrode contact plane is proposed. The specific measurement process is as follows: magnetic field measurement is performed on two perpendicular Cartesian coordinate axes (positive and negative axes), and the alignment of the electrode is accurately determined by comparing the magnetic field difference generated by the misaligned electrode and the perfectly aligned electrode.

[0066] The embodiment of the present application adopts a double-magnetic-pole Hall effect sensor manufactured by COMS technology, as shown in Figure 11 A part of it is the main path of the signal, including a rotating current source, a Hall element, a modulation circuit, a chopper amplification circuit, a demodulation circuit, a comparison circuit and an output control circuit.

[0067] First, the rotating current source rotates the current to change the current direction of the Hall element, so that the offset voltage is opposite in different phases to offset the offset voltage, and the Hall induction voltage remains unchanged, so as to eliminate the offset voltage by superimposing and averaging the output voltages of the two phases, and obtain a more accurate Hall induction signal. After the Hall induction signal is modulated, amplified, demodulated and compared, the best threshold value is obtained by comparison and judgment, and finally the digital high and low levels are output after the threshold value is judged.

[0068] Another part is an auxiliary module of the circuit, and the main component modules include an offset voltage compensation circuit, a voltage stabilizing circuit, a Hall element power supply current bias, a sleep clock control circuit and the like, to assist in processing the main path signal. The functions of each circuit are explained as follows:

[0069] 1) Offset voltage compensation circuit: used for compensating the offset voltage of the Hall element when the rotating current is rotated, to ensure the accuracy of the output signal

[0070] 2) Voltage stabilizing circuit: the voltage stabilizing circuit is used for converting an input wide input range power supply voltage into a precise constant standard voltage, to provide a stable power supply for other modules inside the chip, to suppress power supply noise and reduce the influence of power supply fluctuation on the operation of the internal circuit of the chip.

[0071] 3) Hall element power supply current bias: the circuit block uses a current with a certain temperature coefficient as a constant current power supply bias current source for the Hall element, to realize voltage induction of the Hall element.

[0072] 4) Sleep clock control circuit: the integrated oscillator module uses a constant current to charge and discharge a capacitor to generate a clock signal and provide it to the remaining modules, to control the sleep and working state switching of the overall circuit, to realize a low-power working mode.

[0073] The digital switch Hall sensor chip of the embodiment of the application has only three pins of power supply, reference ground and output, and the packaging diagram thereof is as shown in Figure 12 .

[0074] S110, based on the rotating current source, the Hall effect sensor is processed by the rotating current offset elimination method to eliminate the offset voltage, and the Hall effect sensor after the offset voltage is eliminated is obtained.

[0075] Specifically, the vertical-phase Hall element bias voltage and the horizontal-phase Hall element bias voltage are generated by rotating the current source; based on the vertical-phase Hall element bias voltage, the power supply current of the Hall element flows from top to bottom, and the vertical-phase output voltage of the Hall element is obtained; based on the horizontal-phase Hall element bias voltage, the power supply current of the Hall element flows from left to right, and the horizontal-phase output voltage of the Hall element is obtained; the vertical-phase output voltage of the Hall element and the horizontal-phase output voltage of the Hall element are superimposed and averaged to obtain the Hall effect sensor after the offset voltage is eliminated.

[0076] First, the principle of the rotating current offset elimination technology is described. The Hall element can be equivalent to a Wheatstone bridge model, and the distributed resistors of the four ports of the Hall element can be equivalent to the resistor elements R1-R4 connected between adjacent ports, as shown in Figure 6 In the ideal Hall element equivalent Wheatstone bridge model, the four terminals A, B, C, and D correspond to the four ports of the Hall element, and the distributed resistors R1=R2=R3=R4. Assuming that a bias voltage of V b is applied between the A and C ports, in the ideal case, the potentials at the B and D terminals are equal, both being V b , and there is no voltage difference, so the offset voltage is zero, that is, V os =0. However, in the actual production of the Hall element, due to the pressure resistance effect and manufacturing process differences and other external factors, the resistance values of R1-R4 are often not equal. Assuming that R2=R3=R4=R, the resistance R1 has a deviation R1=R+ΔR, and in the state of zero magnetic field, there will be an offset voltage V os between the D and B terminals, and the expression is:

[0077]

[0078] In order to avoid the influence of the offset voltage on the result, the embodiment of the present application proposes a rotating current offset elimination technology. Only one Hall element needs to be used to eliminate the influence of the offset voltage.

[0079] As shown in Figure 7 , a set of complementary clock signals CLK and CLKB are used to "rotate" the Hall element bias voltage V b of 90° vertical phase (as shown in Figure 7 (a)) and 0° horizontal phase (as shown in Figure 7 (b)). In the case of 90° vertical phase, the power supply current of the Hall element flows from top to bottom, and the output voltage V O can be represented as:

[0080]

[0081] Further, in the case of 0° horizontal phase, the supply current of the Hall element flows from left to right, and the output voltage V o may be expressed as:

[0082]

[0083] It can be seen that the polarities of the offset voltages of the Hall element in the two phase cases are opposite, and the Hall-induced voltage remains unchanged, so as long as the frequency of the current rotation is far higher than the frequency of the magnetic field change, the output voltages of the two phases are superimposed and averaged, the offset voltage can be eliminated, and the output voltage is:

[0084] V O = V H | 0° = V H | 90°

[0085] The rotation current offset elimination technology can at most reduce the equivalent offset magnetic field to within ±30μT, and can better achieve the effect of offset elimination.

[0086] S120, based on the chopper amplification circuit, the offset voltage noise of the Hall effect sensor is eliminated by the chopper stabilization technology, and the Hall effect sensor after eliminating the offset voltage noise is obtained;

[0087] Specifically, the Hall effect sensor is powered on, and the Hall voltage to be processed is obtained; based on the chopper amplification circuit, a high-frequency square wave modulation signal is generated to modulate the Hall voltage to be processed to a high-frequency band, and a high-frequency modulation signal is obtained; the high-frequency modulation signal and the offset voltage and noise of the Hall effect sensor are superimposed to obtain a mixed signal; the mixed signal is amplified and demodulated by the high-frequency square wave modulation signal to obtain a baseband signal; the baseband signal is filtered to obtain the Hall effect sensor after eliminating the offset voltage noise.

[0088] In this embodiment, the chopper stabilization technology process is as shown in Figure 8 , the measured Hall voltage is taken as the input signal V in , a high-frequency square wave signal with a frequency of f chop is added as a modulation signal m1(t), V in is modulated to a high-frequency band through a multiplier to generate a modulation signal V2. The modulated signal V2 is superimposed with the circuit inherent offset voltage V os and noise V noise to form a mixed signal, the signal is amplified through an amplifier to obtain V amp , and then a square wave signal m2(t) with the same frequency and phase as m1(t) is used as a demodulation signal to demodulate the high-frequency signal back to the baseband (the baseband signal is Vdemod ), the demodulated signal passes through a low-pass filter to filter out high-frequency noise and offset voltage, retaining the baseband signal, and outputs a pure useful signal V out , which contains the original low-frequency information.

[0089] Further expand the modulation process, the input voltage is processed by chopper stabilization technology to eliminate offset voltage noise. This technology is based on the principle of modulation and demodulation, by separating the useful signal and noise in the frequency domain, to realize the stable amplification of the signal. Chopper stabilization technology is often used with rotating current technology, widely used in Hall sensors, effectively eliminating offset voltage and 1 / f noise. Its principle is shown in Figure 9 .

[0090] V os and V noise represent the offset voltage and noise of the circuit, respectively, and the chopper clock signals m1(t) and m2(t) are both square waves with a frequency of f chop , and in order to ensure that the input signal and offset voltage, noise and other signals do not overlap, the frequency of the input signal should be limited to within half of f chop .

[0091] Assuming the period of the chopper clock signal is T, the input signal is an analog quantity that varies with time, and the Fourier series expansion of the square wave modulation signal m(t) is:

[0092]

[0093] where C k represents the Fourier coefficient of the kth term and the even term is 0, which is:

[0094]

[0095] Therefore, the input signal V in is converted to odd harmonics by the modulation signal m1(t), amplified by the operational amplifier, and then demodulated by the modulation signal m2(t), whose expression is:

[0096]

[0097] When k, l = 2n (n = 1, 2, 3,...), V2 = 0.

[0098] After two square wave modulation processes, the signal spectrum will finally be located at 2nf chop . After the first modulation, the offset voltage V os and the noise V noise are introduced, and only one modulation process is performed. Let the power spectral density of the offset and noise be S N(f), after modulating by m2(t), we get:

[0099]

[0100] It can be seen that the modulated noise and the misalignment only remain components at odd harmonic frequencies. The input signal after two modulation processes is demodulated back to the baseband. The odd harmonics of the chopping frequency become the frequency band of the misalignment voltage and noise modulation shift. Finally, a low-pass filter is used to filter out the modulated noise and misalignment voltage.

[0101] S130, combine the Hall effect sensor after eliminating the misalignment voltage and the Hall effect sensor after eliminating the misalignment noise to obtain a double magnetic pole type Hall effect sensor.

[0102] S200, based on the double magnetic pole type Hall effect sensor, perform Hall voltage acquisition processing on the electrode to obtain the contact area change amount of the electrode tip;

[0103] Specifically, the electrode is subjected to energization processing to obtain the current density of the electrode; according to the Ampere loop law, the relationship between the electrode magnetic field strength and the current density of the electrode is determined; the Hall voltage of the electrode is obtained by performing Hall voltage acquisition processing on the electrode by the double magnetic pole type Hall effect sensor; the contact area of the electrode is calculated in combination with the Hall voltage of the electrode and the relationship between the electrode magnetic field strength and the current density of the electrode, to obtain the contact area change amount of the electrode tip.

[0104] First of all, it needs to be pointed out that, as shown in Figure 3 Due to mechanical fatigue, the electrode loses its working axis and is displaced by δ relative to its reference axis, as shown in Figure 3 (b), misalignment of the electrode can cause adverse characteristics in the welding process and the quality of the welded joint. Misalignment or misalignment, whether axial or angular, can lead to irregular weld shape and reduced weld size and projection due to asymmetric force distribution and changes in contact surface.

[0105] When the electrode is misaligned, the contact area C a According to the change of the contact surface equation formula, its expression is:

[0106]

[0107] In the above formula, r represents the radius of the electrode tip, δ represents the misalignment between the electrodes, and C a represents the contact area of the electrode tip.

[0108] Therefore, the alignment of the welding electrode has a significant impact on the quality of the spot welded joint. The deformation of the contact surface is caused by misalignment, as shown in Figure 4For the cross-section of the two electrode tips when the electrodes are misaligned, from the figure and the formula, it can be concluded that the reduction of the contact area is closely related to the electrode offset. When the electrodes are not offset, that is, δ = 0, the contact area is maximum (C a0 = πr 2 ), and as δ increases, C a nonlinearly decreases.

[0109] Further, the embodiment of the present application proposes an initial method for detecting electrode offset by measuring the change of the magnetic field. First, the theoretical influence between the electrode arrangement state and the generated magnetic field is studied. By simplifying the electrode as a wire with constant current density, the magnetic field behavior generated by the wire follows Ampere's law, and the relationship between the magnetic field (B) and the current density can be derived from the Ampere integral law, which links the magnetic field strength to its source (current density), where the expression of the Ampere integral law is:

[0110]

[0111] In the above formula, B represents the magnetic field, J represents the current density, and E represents the electric field strength.

[0112] According to the expression of the Ampere integral law, if current flows through the electrode, there is a current density, and a rotating magnetic field will appear around the electrode, and the rotor of the magnetic field points in the same direction as the current density, which is shown in Figure 5 .

[0113] As can be known from the above introduction, under the condition that the welding current (I) is constant, when the electrode misalignment leads to the reduction of C a , the current density J increases significantly, so the current density (J) is inversely proportional to the contact area (C a ), and the relationship can be expressed as:

[0114]

[0115] According to the Ampere loop law, the relationship between the magnetic field strength (B) and the current density (J) in the simplified model can be expressed as:

[0116]

[0117] In the above formula, k is a proportional coefficient.

[0118] It is determined by the electrode geometry, material characteristics and magnetic field distribution experiment calibration. The formula shows that the magnetic field strength B is linearly and positively related to the current density J (i.e., the reciprocal of the contact area C a ). During the detection process, k needs to be repeatedly corrected through data fitting to improve the accuracy of the model.

[0119] The output voltage (VH ) and the magnetic field strength (B) is:

[0120] V H = S·B·cosθ

[0121] In the above formula, S is the Hall coefficient, and θ is the angle between the magnetic field direction and the sensor plane.

[0122] After eliminating the angle influence through calibration experiments, it is simplified to V H = K·B, where K is the comprehensive sensitivity coefficient. Combined with the magnetic field strength formula, the final result is:

[0123]

[0124] It can be concluded that the Hall voltage V H is inversely proportional to the contact area C a , and real-time measurement of V H can indirectly reflect the change of the contact area, so as to calculate the electrode misalignment amount δ.

[0125] S300, the contact area change of the electrode tip is processed by the neural network model, the misalignment amount of the electrode is determined and compared with the standard threshold misalignment amount, and the electrode misalignment detection result is obtained.

[0126] Specifically, the contact area change of the electrode tip is processed by the neural network model, the average and variance of the magnetic field strength are obtained, the misalignment amount of the electrode is determined according to the average and variance of the magnetic field strength, the standard electrode is collected and processed by the Hall voltage based on the double-magnetic-pole Hall effect sensor, the misalignment amount of the standard electrode is obtained as the standard threshold misalignment amount, if the misalignment amount of the electrode is greater than the standard threshold misalignment amount, the electrode currently has the axial misalignment problem, and the audible and light alarm is triggered, if the misalignment amount of the electrode meets the range of the standard threshold misalignment amount, the electrode currently does not have the axial misalignment problem, and the electrode misalignment detection result is output.

[0127] First, as shown in Figure 10 , the flow of electrode offset detection is: after the system starts, the double-magnetic-pole Hall sensor is first initialized, the rotating current offset elimination technology and the chopping stabilization technology are activated to eliminate noise, and the four-axis anti-parallel layout is calibrated; then the magnetic field is generated by electrode energization, the Hall voltage V H is collected in real time, the contact area change is calculated combined with the formula , after the data is received by the ESP32 microcontroller, it is transmitted to the PLC gateway through the LoRaWAN protocol and stored in the MySQL database; the decision tree model extracts the average and variance of the magnetic field strength, judges whether the misalignment amount δ exceeds the threshold δ max , if it exceeds, the audible and light alarm is triggered and the data is recorded, if the misalignment amount is within the threshold, it enters the low-power sleep mode and waits for the next round of detection.

[0128] Further, as shown in Figure 13 , the Hall sensor is integrated with the electrode cap grinder and can be fixed on the side of the electrode cap grinder. After a certain number of welding points are welded, the electrode cap is ground on the electrode cap grinder. After the grinding of the electrode cap is completed, the upper and lower electrodes are moved to the center of the PCB circumference, as shown in Figure 14 , the electrode misalignment detection of the welding robot. The upper and lower electrodes are controlled to contact, and a preset current is designed to trigger the generation of a magnetic field. Four Hall sensors are used to collect magnetic field data and convert the value to voltage, which is compared with the electrode centering group to set the threshold value. The data acquisition and signal processing module is used to calculate the average magnetic field of each sensor for each inspection, and the difference between the axes is sent to the welding database for subsequent data processing and judgment.

[0129] As shown in Figure 15 , the communication protocol between the device and the database is based on the ESP32 microcontroller receiving information to obtain the magnetic field value, calculate the magnetic field difference of each axis, and then send the value to the PLC. The device collects data and sends it to the gateway through the LoRaWAN communication protocol, and the gateway sends the data to the PLC. Finally, the PLC serves as a gateway between the LoRa gateway and the database that stores data for analysis.

[0130] In summary, the specific implementation of the embodiment of the present application is described:

[0131] 1) Selection and installation of Hall effect sensor.

[0132] A dual-pole Hall effect sensor based on CMOS technology is selected, which integrates a rotating current imbalance elimination technology and a chopping stabilization technology, which can effectively eliminate imbalance voltage noise and improve the accuracy of magnetic field measurement. At the same time, considering the temperature stability of the sensor, a Hall effect sensor with low temperature drift is selected to reduce the influence of temperature changes on the measurement results of the magnetic field. The Hall effect sensor is installed near the electrode of the resistance spot welding equipment, ensuring that the measurement plane of the sensor is parallel to the contact plane of the electrode, and the center of the sensor is located on the electrode axis, so as to accurately measure the magnetic field generated by the electrode. The sensor is fixed on the bracket of the equipment by screws to ensure its stability during welding. The installation position should be avoided to be affected by welding current, electromagnetic interference and other factors. In order to realize the collection of multi-directional magnetic field data, the sensor is configured in an anti-parallel layout, which can simultaneously collect magnetic field data on four Cartesian coordinate axes.

[0133] 2) Microcontroller and communication module.

[0134] An ESP32 microcontroller is used, which has strong data processing capabilities and multiple communication interfaces. The ESP32 is connected with the Hall effect sensor through the SPI interface to achieve fast collection of sensor data. A LoRaWAN communication module is installed on the ESP32 microcontroller to send the collected magnetic field data to the gateway. An industrial gateway supporting the LoRaWAN protocol is selected, which can receive the LoRaWAN data packets sent by the ESP32 microcontroller and convert them into a signal format suitable for PLC processing. The gateway is connected with the PLC through Ethernet, and the PLC acts as a data transfer station to store the magnetic field data into the database and make alarm judgments according to the set threshold. A MySQL database is built to store the magnetic field data and alarm records received from the PLC. Corresponding data tables are created in the database, including the magnetic field data table and the alarm record table, to effectively manage and query the data.

[0135] 3) Decision tree algorithm integration.

[0136] Key features are extracted from the magnetic field data collected by the Hall effect sensor, including magnetic field strength, change rate, and other related statistical parameters (such as mean, variance, peak value, etc.). These features are mapped to the geometric parameters of the radar chart, and each feature is defined with a clear dimension in the radar chart, so that these geometric parameters can intuitively represent the alignment state and misalignment degree of the electrodes.

[0137] Using the labeled training dataset containing magnetic field features under different electrode alignment states and corresponding alignment states, a decision tree algorithm is applied to build the model. The decision tree recursively partitions the feature space to form a series of decision rules for classifying the alignment state of the electrodes. Cross-validation and other methods are used to strictly train and evaluate the constructed decision tree model to ensure its accuracy and robustness. By adjusting the parameters of the decision tree (such as tree depth, splitting criteria, etc.), the model performance is optimized to improve its ability to distinguish between normal alignment and misalignment states.

[0138] 4) Real-time monitoring and alarm.

[0139] In practical applications, the real-time collected magnetic field features are input into the trained decision tree model. The model quickly outputs the judgment of the electrode alignment state based on the input feature data. Combined with the output of the decision tree model, an alarm threshold is set. When the system detects electrode misalignment, it automatically sends an alarm and records relevant data, providing a basis for subsequent analysis and optimization.

[0140] 5) Database management and query interface development.

[0141] A web-based database management and query interface is developed, using PHP and HTML technology to build the front-end page, and interacting with the MySQL database through the MySQLi extension. The front-end page uses responsive design and can adapt to different terminal devices, providing a good user experience. The interface is beautified using CSS style sheets, making it simple, intuitive and easy to use.

[0142] In the management and query interface, functions such as magnetic field data query, alarm record query, data statistical analysis, etc. are provided. Users can quickly retrieve the required data by inputting query conditions such as time range, electrode number, magnetic field strength range, etc. The data statistical analysis function can statistically analyze historical data and display them in intuitive charts such as bar charts, line charts, pie charts, etc., providing decision support for users. Set user permission management, different users have different operation permissions, to ensure the security and confidentiality of data.

[0143] Therefore, the embodiments of the present application have the following distinguishing points from the prior art:

[0144] 1) The double-magnetic-pole type Hall effect sensor manufactured based on CMOS technology can replace the traditional continuous working form with a sampling-sleeping working form. In the design of the sensor, rotating current unbalance elimination technology and chopping stabilization technology are used to eliminate unbalanced voltage and noise. The sensor has a wide temperature range of -40℃ to 125℃, and supports double-magnetic-pole switching type detection of magnetic field.

[0145] 2) It can monitor the lateral displacement of the electrode pressure in the resistance spot welding process in real time, and judge the electrode centering and misalignment state in time, so as to timely repair and adjust the electrode, and ensure the stability of the welding process and the welding quality.

[0146] 3) Four Hall effect sensors are arranged in an anti-parallel layout to realize synchronous acquisition of magnetic field data on four Cartesian coordinate axes, so as to fully capture the change of magnetic field.

[0147] 4) By measuring the magnetic field and associating it with the electrode misalignment state, a new monitoring and correction method is provided, which does not rely on traditional destructive testing.

[0148] 5) A data acquisition and communication protocol is designed to transmit data from the sensor to the PLC through the LoRaWAN communication protocol, and then store it in the database, realizing efficient processing and analysis of data.

[0149] 6) The Hall effect sensor, microcontroller, communication module, PLC, database and decision tree algorithm components are seamlessly integrated to form a complete resistance spot welding electrode centering and misalignment monitoring system, realizing real-time monitoring and accurate judgment of the electrode alignment state.

[0150] 7) Extract key features from the magnetic field data collected by the Hall effect sensors, including magnetic field strength, rate of change, and other relevant statistical parameters, and map these features onto the geometric parameters of a radar chart to visually represent the alignment state and misalignment degree of the electrode.

[0151] 8) Use a labeled training dataset containing magnetic field features under different electrode alignment states and corresponding alignment states to apply a decision tree algorithm to build a model that forms a series of decision rules by recursively partitioning the feature space for classifying the alignment state of the electrode.

[0152] 9) Real-time collected magnetic field features are input into the trained decision tree model, which quickly outputs a judgment about the electrode alignment state based on the input feature data, and sets an alarm threshold in combination with the output of the decision tree model to achieve automatic alarm and data recording.

[0153] Therefore, the embodiments of the present application have the following advantages over the prior art:

[0154] 1) The dual-pole Hall effect sensor manufactured using CMOS technology, combined with rotating current offset elimination technology and chopping stabilization technology, effectively eliminates offset voltage and noise, ensuring more accurate output signals when the magnetic field changes. At the same time, the anti-parallel layout of the four Hall effect sensors can simultaneously collect magnetic field data on four Cartesian coordinate axes, fully capturing magnetic field changes and further improving detection accuracy.

[0155] 2) The designed data acquisition and communication protocol uses LoRaWAN communication protocol to transmit sensor data to PLC and then store it to the database, achieving efficient processing and analysis of data. This allows real-time monitoring of electrode misalignment during welding, providing timely and accurate data support for welding quality evaluation.

[0156] 3) The introduction of the decision tree model enables real-time and non-destructive detection of electrode misalignment. By setting a threshold, the system can promptly alert the operator or the automated system to adjust the welding parameters when the detection result exceeds the threshold, avoiding the occurrence of welding defects and improving welding efficiency and quality.

[0157] 4) By monitoring and adjusting welding parameters in real time, welding defects and rework caused by electrode misalignment are reduced, and material waste and repair costs during production are lowered, effectively reducing overall production costs.

[0158] 5) This embodiment combines Hall effect sensor technology, data communication technology, and intelligent monitoring technology to achieve intelligent monitoring and control of the welding process. This helps to promote the development of welding technology towards higher efficiency and intelligence, improving the technical level and production efficiency of the entire welding industry.

[0159] 6) Through real-time monitoring and timely adjustment of electrode misalignment, the accurate alignment of the electrode during welding is ensured, the mechanical properties of the welded joint are improved, the stability and reliability of the welding process are enhanced, and the occurrence of welding defects and failures is reduced.

[0160] In summary, the embodiment of the application first uses the rotating current imbalance elimination technology and the chopper stabilization technology to effectively reduce the imbalance and noise of the Hall element and the amplifier circuit, thereby improving the accuracy of magnetic field measurement. At the same time, in view of the error of the traditional Hall element model, an optimized structure model is proposed, and the related parameters of the Hall element are accurately set through the hardware description language, to ensure the accuracy of the simulation design and further improve the detection accuracy. Further, by measuring the magnetic field of the welding gun and obtaining the magnetic field values on the Cartesian axes, the magnetic field distribution under the electrode alignment state can be more comprehensively and accurately reflected, providing rich data support for accurate detection of electrode centering and misalignment. Combined with a reliable alarm sending protocol and an alarm threshold determined through experiments, whether the electrode has misaligned can be determined in a timely and accurate manner, avoiding the occurrence of welding defects, and finally through the decision tree model, complex pattern recognition and classification can be performed according to the data collected by the sensor, improving the detection accuracy of electrode misalignment and reducing the situation of misjudgment and omission.

[0161] Further, the sleep and sampling switching mode is adopted, which greatly reduces the power consumption on the premise of ensuring the performance of the sensor, so that the sensor can work stably for a long time, providing a basis for real-time monitoring. The system can collect the magnetic field data of the welding gun in real time, and send the data to the data analysis software through the LoRa communication protocol, realizing real-time monitoring and data analysis of the welding process, timely discovering problems in the welding process and adjusting. The decision tree model can quickly process the real-time collected data, realize real-time evaluation of electrode misalignment, and provide strong support for real-time monitoring. And by establishing a reliable alarm sending protocol and determining a reasonable alarm threshold, when the electrode misalignment exceeds the set threshold, the system can timely send an alarm to remind the operator or automatically adjust the welding parameters, avoid the occurrence of welding defects, and improve the welding quality and production efficiency. Based on the evaluation results of the decision tree model, whether the electrode is misaligned can be quickly judged, and the welding parameters can be adjusted in a timely manner according to the output of the model, to realize the automatic control of the welding process.

[0162] Reference Figure 2 , the electrode misalignment detection system based on the double-magnetic-pole type Hall effect sensor comprises:

[0163] The first module 201 is used for eliminating the imbalance voltage and noise of the Hall effect sensor by the rotating current imbalance elimination method and the chopper stabilization technology elimination method, to obtain the double-magnetic-pole type Hall effect sensor.

[0164] The second module 202 is configured to perform Hall voltage acquisition and processing on the electrode based on the double-magnetic-pole Hall effect sensor, and obtain a contact area change amount of the electrode tip;

[0165] The third module 203 is configured to perform feature extraction processing on the contact area change amount of the electrode tip by using a neural network model, determine a misplacement amount of the electrode, and compare the misplacement amount with a standard threshold misplacement amount, to obtain an electrode misplacement detection result.

[0166] The content in the method embodiments is applicable to the system embodiments, the system embodiments specifically implement the same functions as the method embodiments, and achieve the same beneficial effects as the method embodiments.

[0167] The above is a specific description of the preferred implementation of the present application, but the present application is not limited to the above-mentioned embodiments. Those skilled in the art can make various equivalent modifications or replacements without departing from the spirit of the present application. These equivalent modifications or replacements are all included in the scope defined by the claims of the present application.

Claims

1. A method for detecting electrode misalignment based on a dual-pole type Hall effect sensor, characterized by, The method comprises the following steps: The Hall effect sensor is subjected to a rotating current offset elimination method and a chopping stabilization technology elimination method to eliminate offset voltage and noise, and a double-magnetic-pole Hall effect sensor is obtained; The Hall voltage of the electrode is collected and processed based on the double-magnetic-pole Hall effect sensor to obtain the contact area change of the electrode tip; The contact area change of the electrode tip is subjected to feature extraction processing by a neural network model, the dislocation amount of the electrode is determined, and the dislocation detection result of the electrode is obtained by comparing the dislocation amount with a standard threshold dislocation amount.

2. The electrode misalignment detection method based on a dual-pole type Hall effect sensor according to claim 1, characterized by, The double-magnetic-pole Hall effect sensor comprises a rotating current source, a Hall element, a modulation circuit, a chopping amplification circuit, a demodulation circuit, a comparison circuit and an output control circuit, and the rotating current source, the Hall element, the modulation circuit, the chopping amplification circuit, the demodulation circuit, the comparison circuit and the output control circuit are connected in sequence.

3. The electrode misalignment detection method based on the dual-pole type Hall effect sensor according to claim 2, characterized by, The step of eliminating offset voltage and noise from the Hall effect sensor by the rotating current offset elimination method and the chopping stabilization technology elimination method to obtain the double-magnetic-pole Hall effect sensor specifically comprises: Based on the rotating current source, the Hall effect sensor is subjected to offset voltage elimination processing by the rotating current offset elimination method to obtain a Hall effect sensor with eliminated offset voltage; Based on the chopping amplification circuit, the Hall effect sensor is subjected to offset voltage noise elimination processing by the chopping stabilization technology elimination method to obtain a Hall effect sensor with eliminated offset voltage noise; The Hall effect sensor with eliminated offset voltage and the Hall effect sensor with eliminated offset voltage noise are combined to obtain the double-magnetic-pole Hall effect sensor.

4. The electrode misalignment detection method based on the dual-pole type Hall effect sensor according to claim 3, characterized by, The step of eliminating offset voltage from the Hall effect sensor by the rotating current offset elimination method based on the rotating current source to obtain a Hall effect sensor with eliminated offset voltage specifically comprises: The rotating current source generates a Hall element bias voltage of a vertical phase and a Hall element bias voltage of a horizontal phase; Based on the Hall element bias voltage of the vertical phase, the power supply current of the Hall element flows from top to bottom to obtain a vertical phase output voltage of the Hall element; Based on the Hall element bias voltage of the horizontal phase, the power supply current of the Hall element flows from left to right to obtain a horizontal phase output voltage of the Hall element; The vertical phase output voltage of the Hall element and the horizontal phase output voltage of the Hall element are superimposed and averaged to obtain the Hall effect sensor with eliminated offset voltage.

5. The electrode misalignment detection method based on the dual-pole type Hall effect sensor according to claim 4, characterized by, The step of eliminating offset voltage noise from the Hall effect sensor by the chopping stabilization technology elimination method based on the chopping amplification circuit to obtain a Hall effect sensor with eliminated offset voltage noise specifically comprises: The Hall effect sensor is powered to obtain a Hall voltage to be processed; Based on the chopping amplification circuit, a high-frequency square wave modulation signal is generated to modulate the Hall voltage to be processed to a high frequency band to obtain a high-frequency modulation signal; The high-frequency modulation signal is superimposed with the offset voltage and noise of the Hall effect sensor to obtain a mixed signal; After amplification processing of the mixed signal, the mixed signal is demodulated by the high-frequency square wave modulation signal to obtain a baseband signal; The baseband signal is filtered to remove high-frequency noise and offset voltage, and a Hall effect sensor without offset voltage noise is obtained.

6. The electrode misalignment detection method based on the dual-pole type Hall effect sensor according to claim 5, characterized by, The Hall voltage of the electrode is collected by the double-magnetic-pole Hall effect sensor, and the contact area change of the electrode tip is obtained. The electrode is energized to obtain the current density of the electrode. According to the Ampere loop law, the relationship between the electrode magnetic field strength and the current density of the electrode is determined. The Hall voltage of the electrode is collected by the double-magnetic-pole Hall effect sensor. The contact area of the electrode is calculated based on the Hall voltage of the electrode and the relationship between the electrode magnetic field strength and the current density of the electrode, and the contact area change of the electrode tip is obtained.

7. The electrode misalignment detection method based on the dual-pole type Hall effect sensor according to claim 6, characterized by, The expression of the electrode contact area calculation is as follows: In the above formula, V H represents the Hall voltage of the electrode, K represents the sensitivity coefficient, k represents the proportional coefficient, C a represents the contact area of the electrode tip, and I represents the current size through the electrode.

8. The electrode misalignment detection method based on the dual-pole type Hall effect sensor according to claim 7, characterized by, The contact area change of the electrode tip is extracted by the neural network model, and the misalignment amount of the electrode is determined and compared with the standard threshold misalignment amount to obtain the electrode misalignment detection result. The contact area change of the electrode tip is extracted by the neural network model, and the misalignment amount of the electrode is determined and compared with the standard threshold misalignment amount to obtain the electrode misalignment detection result. The contact area change of the electrode tip is extracted by the neural network model, and the misalignment amount of the electrode is determined and compared with the standard threshold misalignment amount to obtain the electrode misalignment detection result. If the misalignment amount of the electrode is greater than the standard threshold misalignment amount, the electrode currently has an axial misalignment problem, and the audible and light alarms are triggered. If the misalignment amount of the electrode meets the range of the standard threshold misalignment amount, the electrode currently does not have an axial misalignment problem, and the electrode misalignment detection result is output. The following modules are included:

9. An electrode misalignment detection system based on a dual-pole type Hall effect sensor, characterized by, The first module is used to eliminate the offset voltage and noise of the Hall effect sensor by the rotating current offset elimination method and the chopper stabilization technique elimination method, and a double-magnetic-pole Hall effect sensor is obtained. The second module is used to collect the Hall voltage of the electrode based on the double-magnetic-pole Hall effect sensor, and the contact area change of the electrode tip is obtained. The third module is used to extract the contact area change of the electrode tip by the neural network model, determine the misalignment amount of the electrode, and compare it with the standard threshold misalignment amount to obtain the electrode misalignment detection result. ​

Images