An automatic centering adjustment system in an electrode rod welding process

Abstract

The application relates to the technical field of intelligent sensors, in particular to an automatic centering adjustment system in an electrode rod welding process. The system comprises: an interference identification module which constructs time domain characteristic values representing dust shielding degrees and frequency domain characteristic values representing electromagnetic interference degrees by performing time-frequency domain joint analysis on a diameter measurement signal; an adaptive filtering parameter decision module which generates a comprehensive index by fusing the characteristic values, dynamically adjusts a filtering window size to realize adaptive adaptation of filtering strength; and a centering execution module which calculates an eccentricity parameter based on the filtered signal and drives an execution mechanism to complete automatic centering adjustment. The application solves the problem that dust and electromagnetic interference in a welding site cause measurement signal deviation to affect centering precision, dynamically and adaptively filters noise while protecting signal details, and improves the precision of automatic centering in the electrode rod welding process.

Description

Technical Field

[0001] This application relates to the field of intelligent sensor technology, specifically to an automatic alignment adjustment system in the electrode rod welding process. Background Technology

[0002] Electrode rod welding is crucial in special metallurgy (such as electroslag remelting), new energy vehicles, and high-end materials processing. In electroslag remelting, the welding quality of the electrode rod and dummy electrode directly affects the quality of the steel ingot. Insufficient alignment precision can easily lead to uneven gaps, causing serious accidents such as breakdown or even melting and leakage of the crystallizer wall. Traditional alignment methods rely on manual visual inspection and experience, which suffers from low precision, poor consistency, and the inability to quantify micron-level deviations. In contrast, automatic alignment technology, through real-time sensor detection and closed-loop control, achieves micron-level high-precision alignment and non-contact standardized operation, effectively reducing welding defect rates and safety risks. It is a core technology for ensuring the quality of special metallurgy and new energy manufacturing.

[0003] In the automatic alignment process of electrode rod welding, a laser diameter gauge is usually used to measure the diameter, which is then pre-processed and transmitted to the PLC control system. However, pre-welding grinding generates a large amount of suspended dust, and the operation of high-power motors, frequency converters, and other equipment on site is accompanied by high-frequency electromagnetic interference. These suspended particles and electromagnetic noise can cause serious deviations in the measurement signal, resulting in abnormal automatic alignment adjustment and reducing the accuracy of automatic alignment during electrode rod welding. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide an automatic alignment and adjustment system for the electrode rod welding process. The specific technical solution adopted is as follows: This application proposes an automatic alignment adjustment system for electrode rod welding process, the system comprising an interference identification module, an adaptive filtering parameter decision module, and an alignment execution module: The interference identification module is used to segment the diameter measurement signal of the electrode rod and perform joint time-frequency domain analysis to construct time-domain and frequency-domain feature values. The time-domain feature values ​​are constructed based on the randomness and asymmetry of the signal peak shape and are used to characterize the degree of physical obstruction of the optical path by dust particles. The frequency-domain feature values ​​are constructed based on the abnormal distribution characteristics of the signal spectrum energy and are used to characterize the degree of coupling interference of high-frequency electromagnetic equipment on the measurement signal. The adaptive filtering parameter decision module is used to fuse the time-domain feature values ​​and frequency-domain feature values ​​to generate a comprehensive feature index, and to establish a nonlinear mapping relationship between the comprehensive feature index and the filtering window size to achieve dynamic adaptation of the filtering intensity. The centering execution module is used to calculate the eccentricity parameters based on the filtered diameter measurement signal and drive the actuator to complete the automatic centering adjustment.

[0005] Preferably, the construction logic of time-domain feature values ​​is as follows: The local morphological feature sequence of each peak in each diameter measurement signal is extracted. By analyzing the degree of disorder between the local morphological feature sequences of all peaks and the peak asymmetry, the time domain feature value of each diameter measurement signal is constructed.

[0006] Preferably, the local morphological feature sequence is composed of the peak in the diameter measurement signal and all amplitudes in the neighborhood on both sides in chronological order.

[0007] Preferably, the quantification process of peak asymmetry is as follows: In the local morphological feature sequence of each peak, the ratio of the sum of the amplitudes on the left side to the sum of the amplitudes on the right side of each peak is calculated and recorded as the amplitude ratio. The standard deviation of the amplitude ratios of all peaks in each diameter measurement signal is taken as the peak asymmetry of each diameter measurement signal.

[0008] Preferably, the process of constructing the frequency domain feature values ​​is as follows: The energy intensity of each diameter measurement signal at all frequencies in the frequency domain is fitted. By analyzing the dispersion of the fitting error corresponding to the energy intensity at all frequencies, and combining the ratio of the total energy of the first half of the frequency to the total energy of the second half of the frequency in each diameter measurement signal, the frequency domain characteristic value of each diameter measurement signal is constructed.

[0009] Preferably, the comprehensive feature index is the sum of time-domain feature values ​​and frequency-domain feature values.

[0010] Preferably, the process of establishing the nonlinear mapping relationship between the comprehensive feature index and the filter window size is as follows: Optimize filter window length The expression is: In the formula, Indicates the pre-defined multiplicative factor; represents the normalized value of the comprehensive characteristic index; round[] represents the rounding function.

[0011] Preferably, the process of filtering the diameter measurement signal is as follows: Each diameter measurement signal is used as input to the filtering algorithm, and the corresponding optimized filtering window length is used as the filtering window parameter in the filtering algorithm. The filtered diameter measurement signal is then output.

[0012] Preferably, the calculation process for the eccentricity parameter is as follows: The center position coordinates (x, y) of the electrode rod when it is stationary are obtained based on the filtered diameter measurement signal. Calculate the deviation between the x-value of the electrode rod center position and the corresponding X-axis coordinate of the X-axis diameter measuring instrument, and the deviation between the y-value and the corresponding Y-axis coordinate of the Y-axis diameter measuring instrument. Take the root mean square of the two deviations and record it as the eccentricity.

[0013] Preferably, the drive actuator performs automatic centering adjustment, including: The obtained eccentricity, the two deviations corresponding to the center position coordinates of the electrode rod, and the preset allowable centering error range are used as inputs, and automatic centering control is performed through a PLC control terminal. This application has the following beneficial effects: (1) This application collects and segments the diameter measurement signal and performs joint analysis in the time and frequency domains to construct time-domain feature values ​​based on peak shape differences and asymmetry, as well as frequency-domain feature values ​​based on discrete distribution of spectral energy. This method can accurately quantify the degree of physical obstruction caused by dust and coupling interference of high-frequency electromagnetic equipment, effectively distinguish between irregular burrs and specific frequency noise, and provide accurate and reliable interference intensity assessment basis for subsequent adaptive filtering, thereby ensuring the acquisition of high-precision measurement data in complex industrial environments. (2) This application generates a comprehensive feature index by fusing time-domain and frequency-domain feature values ​​and establishes a nonlinear mapping relationship between the index and the filter window size, thereby realizing dynamic adaptation of the filter intensity. The median filter window size is automatically adjusted according to the degree of signal interference. When the interference is severe, a large window is used to completely filter out noise, and a small window is used to finely protect the details when the signal is clean. Thus, noise suppression and signal fidelity are effectively balanced under complex working conditions, providing a high-quality data foundation for subsequent high-precision alignment.

[0014] (3) This application calculates the eccentricity parameter accurately based on the filtered diameter measurement signal and drives the actuator to achieve automatic centering adjustment; by analyzing the center position deviation of the electrode rod during static and dynamic scanning, the deviation and eccentricity are obtained, and the PLC is used for closed-loop control in combination with the preset allowable error range. The movement of the dummy electrode is precisely adjusted according to the direction and magnitude of the offset until the eccentricity falls back to the allowable range, thereby effectively overcoming manufacturing and installation errors and improving the accuracy of automatic centering during electrode rod welding. Attached Figure Description

[0015] Figure 1 A block diagram of an automatic centering adjustment system for electrode rod welding process provided in one embodiment of this application; Figure 2 This is a schematic diagram of the optimized filter window length extraction process provided in one embodiment of this application. Detailed Implementation

[0016] The following description, in conjunction with the accompanying drawings, details the specific scheme of an automatic alignment and adjustment system for electrode rod welding provided in this application.

[0017] Please see Figure 1 The diagram illustrates an automatic alignment adjustment system for electrode rod welding process according to an embodiment of this application. The system includes: an interference identification module 101, an adaptive filtering parameter decision module 102, and an alignment execution module 103.

[0018] Interference identification module 101 is used to segment the diameter measurement signal of the electrode rod and perform joint time-frequency domain analysis to construct time-domain feature values ​​and frequency-domain feature values. The time-domain feature values ​​are constructed based on the randomness and asymmetry of the signal peak shape and are used to characterize the degree of physical obstruction of the optical path by dust particles. The frequency-domain feature values ​​are constructed based on the abnormal distribution characteristics of the signal spectrum energy and are used to characterize the degree of coupling interference of high-frequency electromagnetic equipment to the measurement signal.

[0019] a: Measurement signal acquisition and transmission.

[0020] During the automatic alignment and adjustment process of electrode rod welding, a laser diameter gauge is typically used to measure the diameter of the electrode rod. In practice, the electrode rod is assembled onto the electrode rod rotating roller of the robotic welding device. XY bidirectional laser automatic diameter gauges, installed on the right side (aligned with the chuck center) and the lower side (aligned with the chuck center) of the device, acquire the horizontal and vertical diameter measurement signals (i.e., the distance between the horizontal and vertical boundary points). The sampling rate of these diameter gauges is set to 2kHz. Subsequently, the real-time acquired signal is transmitted to the back-end terminal through the RS485 serial communication interface standard. This transmission method has strong anti-interference capabilities and can effectively ensure the reliability of data transmission.

[0021] b: Signal preprocessing.

[0022] First, the real-time acquired diameter measurement signal is segmented, with the length of each segment set to 0.512 seconds. This ensures sufficient data to meet analysis requirements while also ensuring that the number of sampling points is an integer power of 2, facilitating rapid transformation processing for subsequent frequency domain analysis.

[0023] Subsequently, each diameter measurement signal was analyzed separately. Since smoke and particulate matter interference mainly manifests as random isolated spikes in the time domain, while electromagnetic interference and mechanical vibration manifest as continuous noise at specific frequencies and their harmonics in the frequency domain, it is necessary to perform characteristic differentiation analysis of the signals in the time and frequency domains. Each diameter measurement signal was used as the input of the Fast Fourier Transform (FFT) algorithm, where the number of FFT points was set to 1024, and the corresponding frequency domain signal was output based on the Nyquist theorem.

[0024] The process of converting a time-domain signal to a frequency-domain signal using the Fast Fourier Transform algorithm is a well-known technique and will not be elaborated further.

[0025] c: Time-domain feature analysis and construction of time-domain feature values.

[0026] In terms of temporal characteristic analysis, the grinding operation before the electrode rod is aligned results in a large amount of dust and particulate matter suspended in the air. This type of interference is a physical blockage in the optical path, exhibiting extremely high randomness and transient characteristics. When dust or particulate matter passes through the laser beam, it causes an instantaneous attenuation of the receiver's light intensity, which in turn causes an instantaneous jump in the measured diameter value, presenting extremely narrow high-amplitude spikes. Due to the randomness of the concentration of interfering particles and the duration of blockage, the timing, width, and slope of these spikes are all asymmetrical. In contrast, the normal electrode rod signal waveform exhibits a slow, regular sine wave change with rotation, with a wider overall waveform and no abnormal jumps.

[0027] Based on the above analysis, the time-domain eigenvalues ​​are constructed as follows: (1) Local morphological feature sequence: Peaks in the diameter measurement signal are obtained using peak-finding algorithms (such as Gaussian fitting, orthogonal polynomial least squares fitting, zero-crossing detection, or automatic multi-scale peak finding algorithm). The k amplitude values ​​in the left and right neighborhoods of each peak in the diameter measurement signal are obtained respectively (k is 5 in this embodiment). The sequence composed of each peak and the 2k amplitude values ​​in its left and right neighborhoods is recorded as the local morphological feature sequence; if there are less than k amplitude values ​​in the neighborhood, the mean interpolation method is used to supplement them.

[0028] In this embodiment, the peak finding algorithm uses an automatic multi-scale peak finding algorithm to extract the peak value in the signal. The process of extracting the peak value using the automatic multi-scale peak finding algorithm and the process of filling the data using the mean interpolation method are well-known techniques and will not be described in detail here.

[0029] (2) Analyze the degree of disorder between the local morphological feature sequences of all peaks: For each diameter measurement signal, calculate the difference between the local morphological feature sequences of any two peaks, and then map all the differences to a preset number of equally spaced intervals for statistical analysis to obtain the probability distribution of each interval. Finally, use the Shannon entropy formula to calculate the entropy value based on the probability distribution of all intervals as the quantitative result of the degree of disorder between the local morphological feature sequences of all peaks. The larger the entropy value, the more dispersed the peak morphological differences are, thus characterizing the more serious the irregular spikes caused by dust particle interference in the signal.

[0030] It should be noted that the preset quantity in this embodiment is 10. In actual application, as other implementation methods, implementers can also set it according to specific circumstances. This embodiment does not impose any special restrictions.

[0031] It should be noted that in this embodiment, the Euclidean distance between the local morphological feature sequences of any two peaks is taken as the difference between the local morphological feature sequences of any two peaks. The calculation process of the Euclidean distance is a well-known technique and will not be described in detail here. In practical applications, as other implementation methods, implementers may also choose other methods to calculate the difference between sequences according to the specific circumstances. This embodiment does not impose any special restrictions.

[0032] (3) Calculation of energy asymmetry on both sides of each peak: In the local morphological feature sequence of each peak, calculate the ratio of the sum of amplitudes on the left side of each peak to the sum of amplitudes on the right side, and record it as the amplitude ratio. Take the standard deviation of the amplitude ratio of all peaks in each diameter measurement signal as the peak asymmetry of each diameter measurement signal.

[0033] (4) Construction of time-domain feature values: In this embodiment, the product of the normalized value of the disorder quantification result of the difference between the local morphological feature sequences of all peaks in each diameter measurement signal and the normalized value of the peak asymmetry is used as the time-domain feature value of each diameter measurement signal. To further clarify, in this embodiment, the maximum and minimum values ​​of the disorder quantization results of all segment diameter measurement signals are used as the maximum and minimum values ​​of the minimum-maximum normalization method to normalize the disorder quantization results, mapping them to the interval [0,1]. Similarly, the maximum and minimum values ​​of the peak asymmetry of all segment diameter measurement signals are used as the maximum and minimum values ​​of the minimum-maximum normalization method to normalize the peak asymmetry, mapping it to the interval [0,1], thereby eliminating the influence of dimensions and orders of magnitude. The process of normalizing the data using the minimum-maximum normalization method is a well-known technique and will not be described in detail here.

[0034] The time-domain eigenvalue is used to characterize the degree of interference from physical obstruction by dust particles in the electrode rod diameter measurement signal. Its calculation is mainly affected by two factors: the degree of disorder (entropy) between the differences in the local morphological feature sequences of the peaks and the degree of energy asymmetry (standard deviation) on both sides of the peaks. The larger these factors are, the larger the time-domain eigenvalue is, indicating that there are more irregular high-amplitude spikes in the diameter measurement signal caused by different random dust concentrations and obstruction areas, and the peak shape differences are more dispersed and the asymmetry is more severe. This reflects that the physical obstruction interference on the optical path is stronger, which will cause drastic random jumps in the diameter measurement signal. Without strong filtering, it will seriously affect the accuracy of the eccentricity calculation. Conversely, the smaller these factors are, the smaller the time-domain eigenvalue is, indicating that the peak shape of the diameter measurement signal is more regular, consistent and symmetrical. This reflects that the signal is mainly composed of a stable sine wave generated by the rotation of the electrode rod, and the optical path is minimally affected by dust. At this time, the diameter measurement signal quality is high and the influence is weak, which can ensure high-precision measurement.

[0035] d: Frequency domain feature analysis and construction of frequency domain eigenvalues: In terms of frequency domain characteristic analysis, the automatic alignment operation of the electrode rod involves equipment such as lifting roller table, lifting cylinder, guide column, driven roller assembly, and driving roller assembly. Its operation relies on instruments such as high-power motors and frequency converters. These devices operate at different frequencies and powers, and the high-frequency electromagnetic interference generated will interfere with different high-frequency components of the diameter measurement signal when it is acquired. The interference and superposition of multiple high-frequency devices will cause a surge in the high-frequency energy of the diameter measurement signal. Moreover, due to the different frequencies and powers of different devices and the interference superposition, the high-frequency energy distribution presents an irregular discrete spike shape.

[0036] Based on the above analysis, frequency domain eigenvalues ​​are constructed as follows: (1) Signal spectrum characteristics: The normal electrode rod diameter measurement signal is a slowly changing signal with energy mainly concentrated in the low frequency band. The spectrum is a continuous smooth curve that gradually decays from low frequency to high frequency. However, under high frequency electromagnetic interference, the signal energy will be biased to the high frequency part and discrete spikes will appear at specific frequencies.

[0037] (2) Analyze the degree of dispersion of the fitting error of the energy intensity at all frequencies in the frequency domain for each diameter measurement signal: Use the least squares method to fit the energy intensity at all frequencies in the frequency domain signal of each diameter measurement signal, calculate the standard deviation of the absolute difference between the energy intensity at all frequencies and its fitting value, and use it as the quantitative result of the degree of dispersion of the fitting error of the energy intensity at all frequencies in the frequency domain for each diameter measurement signal. For ease of expression, it will be referred to as the energy dispersion of each diameter measurement signal.

[0038] Energy dispersion is used to characterize the degree of fluctuation in the frequency domain energy distribution of the diameter measurement signal from the ideal smooth curve. It reflects whether the signal is mixed with specific frequency interference caused by the operation of equipment such as high-power motors and frequency converters. The larger the energy dispersion, the more significant the discrete energy peaks in the signal spectrum at specific frequencies, and the greater the deviation between the actual energy distribution and the fitted smooth curve. This reflects the more serious the high-frequency electromagnetic interference and the more complex the signal spectrum structure, which will lead to the mixing of high-frequency noise other than the workpiece itself into the measurement signal. Conversely, the smaller the energy dispersion, the more concentrated the signal spectrum energy distribution and the closer it is to the fitted curve. This reflects that there is almost no external electromagnetic interference mixing in the signal, and it mainly presents the low-frequency smooth attenuation characteristics caused by mechanical motion, with high signal purity.

[0039] The process of fitting data using the least squares method is a well-known technique and will not be elaborated further.

[0040] (3) Band energy ratio calculation: Calculate the ratio of the total energy of the first half of the frequency to the total energy of the second half of the frequency in each diameter measurement signal. For ease of expression, it is called the band energy ratio. The larger the band energy ratio, the more concentrated the energy of the diameter measurement signal is in the low frequency part, while the energy in the high frequency part is relatively less. This reflects that the diameter measurement signal is mainly dominated by low frequency mechanical motion changes, the high frequency noise interference is weak, and the signal is true and reliable. Conversely, the smaller the band energy ratio, the more the energy of the diameter measurement signal leaks to the high frequency band, and the low frequency ratio decreases. This reflects that a large amount of high frequency electromagnetic interference or mechanical vibration noise may be superimposed in the diameter measurement signal, causing the high frequency energy to surge abnormally and the signal effectiveness to decrease.

[0041] (4) Construction of frequency domain eigenvalues: frequency domain eigenvalues ​​of the diameter measurement signal of segment h. The expression is: In the formula, This represents the normalized value of the energy dispersion of the diameter measurement signal in segment h; This represents the normalized value of the bandwidth energy ratio of the diameter measurement signal in segment h; This represents a preset constant greater than 0, used to prevent the denominator from being 0. In this embodiment... The value of is 0.01. Provided that the denominator is not zero and does not excessively affect the calculation result, the implementer may also set it according to the specific situation. This embodiment does not impose any special restrictions.

[0042] In this embodiment, the maximum and minimum values ​​of the energy dispersion of all segment diameter measurement signals are used as the maximum and minimum value parameters in the minimum-maximum normalization method to normalize the energy dispersion measurement results and map them to the interval [0,1], thereby eliminating the influence of dimensions and orders of magnitude. Similarly, the frequency band energy ratio and frequency domain characteristic values ​​are normalized according to the same normalization process as the energy dispersion. The specific process will not be described in detail.

[0043] Frequency domain eigenvalues ​​are used to characterize the overall severity of interference from high-frequency electromagnetic equipment coupling in the electrode rod diameter measurement signal. Their calculation is mainly influenced by two factors: the dispersion of the frequency domain energy distribution (energy dispersion) and the proportion of energy distribution in the frequency band (frequency band energy ratio). Among these factors, a larger energy dispersion and a smaller frequency band energy ratio result in a larger frequency domain eigenvalue, indicating a large number of discrete abnormal peaks in the signal spectrum and a shift of energy towards higher frequencies. This reflects extremely strong high-frequency electromagnetic interference generated by high-power motors, frequency converters, etc., causing serious pollution to the measurement signal. Conversely, a smaller energy dispersion and a larger frequency band energy ratio result in a smaller frequency domain eigenvalue, indicating a smooth spectrum of the diameter measurement signal and energy concentrated in the low frequencies. This reflects a clean measurement environment, free from significant electromagnetic interference, and the signal accurately reflects the diameter change of the electrode rod.

[0044] The adaptive filtering parameter decision module 102 is used to fuse the time-domain feature values ​​and frequency-domain feature values ​​to generate a comprehensive feature index, and to establish a nonlinear mapping relationship between the comprehensive feature index and the filter window size to achieve dynamic adaptation of the filtering intensity.

[0045] The core strategy is to use a larger filter window to completely filter out noise for diameter measurement signals that are severely affected by dust, particulate matter and high-frequency electromagnetic interference; while for diameter measurement signals with less interference and mainly effective signals, a smaller filter window is used to finely protect signal details.

[0046] The specific implementation process is as follows: First, the sum of the time-domain and frequency-domain eigenvalues ​​of each diameter measurement signal is used as a comprehensive characteristic index. Next, set the optimized filter window length. Its expression is: In the formula, Indicates the pre-defined multiplicative factor; The normalized value of the comprehensive feature index is represented by ; round[ ] represents the rounding function; where the preset multiplicative factor is used to convert the dimensionless index to the time scale domain corresponding to the sampling rate, and its value range is [10,30]; at a sampling rate of 2kHz, in order to effectively eliminate the influence of long dust blockage lasting for tens of milliseconds, this embodiment sets .

[0047] Preferably, the schematic diagram of the optimized filter window length extraction process provided in this embodiment is as follows: Figure 2 As shown.

[0048] Furthermore, each diameter measurement signal is used as the input to the filtering algorithm, and the corresponding optimized filtering window length is used as the filtering window parameter in the filtering algorithm to output the filtered diameter measurement signal. In this embodiment, the filtering algorithm adopts the median filtering algorithm. The process of using the median filtering algorithm to filter the signal is a well-known technique and will not be described in detail.

[0049] The centering execution module 103 is used to calculate the eccentricity parameters based on the filtered diameter measurement signal and drive the actuator to complete the automatic centering adjustment.

[0050] After preprocessing the diameter measurement signal, given the difficulty in achieving perfect cylindricity during electrode rod manufacturing and the potential initial eccentricity during chuck installation, it is necessary to accurately calculate its spatial position through analysis of the measurement data. Specifically, the diameter measurement signal acquired by the diameter gauge reflects the distance changes between the two boundary points along the X and Y axes. When the sensor measures the maximum distance to the nearest boundary point, it indicates that the electrode rod surface is far from the sensor, and the electrode rod center is biased in the opposite direction to the sensor; conversely, when the distance is at its minimum, it indicates that the electrode rod surface is close to the sensor, and the center is biased in the direction directly opposite to the sensor.

[0051] Based on the above principle, the X-axis coordinates of the left and right boundary points of the electrode rod when it is stationary are first obtained, and the average of the two values ​​is calculated as the X-axis coordinate of the initial center position of the electrode rod. Similarly, obtain the Y-axis coordinates of the upper and lower boundary points on the Y-axis when the electrode rod is stationary, and calculate the average of the two values ​​as the Y-axis coordinate of the initial center position of the electrode rod. .

[0052] During the dynamic scanning process, the X-axis midpoint and the X-axis coordinate value of the X-axis diameter measuring instrument are calculated for each scan. The deviation between them is used to obtain the center offset component in the X-axis direction. Among them, if This indicates that the center has shifted to the left relative to the rest point; if This indicates a rightward offset; the offset component in the Y-axis direction is calculated using the same method. Finally, the eccentricity is calculated by combining the offset components in both directions. .

[0053] Furthermore, the calculated eccentricity ,deviation The preset allowable alignment error range is input into the background PLC control terminal, and the PLC performs automatic alignment control logic judgment based on the above parameters: If the eccentricity is within the preset centering error range, the centering is considered complete and the adjustment is stopped. If the eccentricity exceeds the preset allowable centering error range, corresponding adjustments will be made according to the direction and magnitude of the deviation: For X-axis offset, when When the value is positive, the dummy electrode moves to the left; when the value is negative, it moves to the right. Regarding the Y-axis offset, when... When the value is positive, the dummy electrode moves upward; when the value is negative, it moves downward. Corresponding direction and distance control signals are output to continuously drive the electrode rod to adjust until the eccentricity falls back within the allowable error range, ultimately completing the automatic centering operation.

[0054] It should be noted that the preset allowable centering error range in this embodiment is usually set to ±0.05 mm to ±0.1 mm. In actual application, as other implementation methods, implementers can also set it according to specific circumstances. This embodiment does not impose any special restrictions.

Claims

1. An automatic alignment and adjustment system for electrode rod welding process, characterized in that, The system includes an interference identification module, an adaptive filter parameter decision module, and a centering execution module. The interference identification module is used to segment the diameter measurement signal of the electrode rod and perform joint time-frequency domain analysis to construct time-domain and frequency-domain feature values. The time-domain feature values ​​are constructed based on the randomness and asymmetry of the signal peak shape and are used to characterize the degree of physical obstruction of the optical path by dust particles. The frequency-domain feature values ​​are constructed based on the abnormal distribution characteristics of the signal spectrum energy and are used to characterize the degree of coupling interference of high-frequency electromagnetic equipment on the measurement signal. The adaptive filtering parameter decision module is used to fuse the time-domain feature values ​​and frequency-domain feature values ​​to generate a comprehensive feature index, and to establish a nonlinear mapping relationship between the comprehensive feature index and the filtering window size to achieve dynamic adaptation of the filtering intensity. The centering execution module is used to calculate the eccentricity parameters based on the filtered diameter measurement signal and drive the actuator to complete the automatic centering adjustment.

2. The automatic alignment and adjustment system for electrode rod welding process according to claim 1, characterized in that, The logic for constructing time-domain feature values ​​is as follows: The local morphological feature sequence of each peak in each diameter measurement signal is extracted. By analyzing the degree of disorder between the local morphological feature sequences of all peaks and the peak asymmetry, the time domain feature value of each diameter measurement signal is constructed.

3. The automatic alignment and adjustment system for electrode rod welding process according to claim 2, characterized in that, The local morphological feature sequence is composed of the peak in the diameter measurement signal and all amplitude values ​​in the neighborhood on both sides in chronological order.

4. The automatic alignment and adjustment system for electrode rod welding process according to claim 3, characterized in that, The quantification process of peak asymmetry is as follows: In the local morphological feature sequence of each peak, the ratio of the sum of the amplitudes on the left side to the sum of the amplitudes on the right side of each peak is calculated and recorded as the amplitude ratio. The standard deviation of the amplitude ratios of all peaks in each diameter measurement signal is taken as the peak asymmetry of each diameter measurement signal.

5. The automatic alignment and adjustment system for electrode rod welding process according to claim 1, characterized in that, The process of constructing the frequency domain feature values ​​is as follows: The energy intensity of each diameter measurement signal at all frequencies in the frequency domain is fitted. By analyzing the dispersion of the fitting error corresponding to the energy intensity at all frequencies, and combining the ratio of the total energy of the first half of the frequency to the total energy of the second half of the frequency in each diameter measurement signal, the frequency domain characteristic value of each diameter measurement signal is constructed.

6. The automatic alignment and adjustment system for electrode rod welding process according to claim 1, characterized in that, The comprehensive feature index is the sum of the time-domain feature value and the frequency-domain feature value.

7. The automatic alignment and adjustment system for electrode rod welding process according to claim 1, characterized in that, The process of establishing the nonlinear mapping relationship between the comprehensive feature index and the filter window size is as follows: Optimize filter window length The expression is: In the formula, Indicates the pre-defined multiplicative factor; represents the normalized value of the comprehensive characteristic index; round[] represents the rounding function.

8. An automatic alignment and adjustment system for electrode rod welding process according to claim 7, characterized in that, The process of filtering the diameter measurement signal is as follows: Each diameter measurement signal is used as input to the filtering algorithm, and the corresponding optimized filtering window length is used as the filtering window parameter in the filtering algorithm. The filtered diameter measurement signal is then output.

9. An automatic centering adjustment system for electrode rod welding process according to claim 1, characterized in that, The calculation process for the eccentricity parameter is as follows: The center position coordinates (x, y) of the electrode rod when it is stationary are obtained based on the filtered diameter measurement signal. Calculate the deviation between the x-value of the electrode rod center position and the corresponding X-axis coordinate of the X-axis diameter measuring instrument, and the deviation between the y-value and the corresponding Y-axis coordinate of the Y-axis diameter measuring instrument. Take the root mean square of the two deviations and record it as the eccentricity.

10. An automatic centering adjustment system for electrode rod welding process according to claim 9, characterized in that, The drive actuator completes automatic centering adjustment, including: The obtained eccentricity, the two deviations corresponding to the center position coordinates of the electrode rod, and the preset allowable centering error range are used as inputs to perform automatic centering control through the PLC control terminal.

Images