Arcing machine control method based on multi-parameter collaborative optimization

Abstract

The invention relates to the technical field of arc discharge control, in particular to an arc discharge machine control method based on multi-parameter collaborative optimization, which comprises the following steps of: generating a first regulation and control instruction by adopting a fuzzy PID (Proportion Integration Differentiation) algorithm based on a deviation value between a dynamic parameter feedback signal and a preset threshold value, the axial feeding speed of the servo motor, the power output curve of the arc generator and the coolant spraying time sequence are synchronously adjusted, so that the arc energy density gradient and the workpiece heat conduction rate form dynamic balance, the mechanical resonance frequency in the machining process is monitored through the vibration spectrum analysis unit, and when it is detected that the resonance intensity exceeds a safety threshold value, the coolant spraying time sequence is started. And a second regulation and control instruction is generated by triggering a frequency compensation algorithm, and the driving frequency of the servo motor is corrected in real time, so that the control precision is improved, the arc discharge process is more stable, the response speed is increased, the arc discharge time is shortened, the adaptability is enhanced, the requirements of different materials and processes can be met, the energy consumption is reduced, and the production efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of arcing control, and particularly relates to an arcing machine control method based on multi-parameter collaborative optimization. Background Art

[0002] Arcing control refers to preventing or reducing the arcing phenomenon generated when a switch disconnects a circuit through specific technical means and management measures in an electrical system. At present, there are problems in arcing control such as low control accuracy, slow response speed, poor adaptability, and high energy consumption.

[0003] In view of this, we propose an arcing machine control method based on multi-parameter collaborative optimization to solve the existing problems. Summary of the Invention

[0004] The purpose of the present invention is to provide an arcing machine control method based on multi-parameter collaborative optimization to solve the problems proposed in the above background art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: An arcing machine control method based on multi-parameter collaborative optimization, the working steps include:

[0006] S1: Real-time collect dynamic arc parameters through an arc voltage sensor, a current sensor and an infrared thermal imager; wherein, the dynamic parameters include voltage value, current intensity, arc morphology characteristics and workpiece surface temperature distribution;

[0007] S2: The central controller matches the target processing mode according to the preset process parameter library, and based on the deviation amount between the dynamic parameter feedback signal and the preset threshold, uses the fuzzy PID algorithm to generate the first regulation instruction;

[0008] S3: The motion control module receives the first regulation instruction, synchronously adjusts the axial feed speed of the servo motor, the power output curve of the arc generator and the coolant injection timing, so that the arc energy density gradient and the workpiece heat conduction rate form a dynamic balance;

[0009] S4: Monitor the mechanical resonance frequency during the processing through the vibration spectrum analysis unit. When the detected resonance intensity exceeds the safety threshold, trigger the frequency compensation algorithm to generate the second regulation instruction and correct the servo motor drive frequency in real time.

[0010] Further, in S2, the fuzzy PID algorithm adopts a double-input and triple-output structure. The input variables are the arc length deviation amount ΔL and the heat affected zone temperature gradient ΔT. The output variables are the servo motor speed correction coefficient Kv, the arc current duty cycle correction coefficient Kc and the coolant flow correction coefficient Kw. The association logic of the fuzzy rule base includes: when ΔL > 0 and ΔT < 0, increase Kv and decrease Kc; when ΔL < 0 and ΔT > the critical molten pool maintenance temperature, start the exponential compensation mode of Kw.

[0011] Further, in S3, the power output curve adjustment adopts a segmented pulse modulation strategy, including: applying a strong current pulse of 200 - 300 A with a pulse width of 50 - 100 ms during the arc starting stage; after entering the stable processing stage, switching to a square wave output with a variable duty cycle, and its fundamental frequency is negatively correlated with the thermal diffusivity of the workpiece material.

[0012] Further, in S4, the specific steps for the trigger frequency compensation algorithm to generate the second regulation instruction include:

[0013] A1: Synchronously collect mechanical vibration signals through a triaxial accelerometer array distributed on the spindle box, electrode clamping arm, and base platform at a sampling rate greater than or equal to 5 kHz;

[0014] A2: After the vibration signal processing module performs band - pass filtering on the original signal, extract the current dominant resonance frequency fr and the corresponding amplitude Ar through fast Fourier transform; among them, the cut - off frequency of the band - pass filtering is 100 Hz - 2 kHz;

[0015] A3: The frequency compensation decision unit compares fr with the equipment natural frequency database, and triggers the compensation mechanism when any of the following conditions is met: the frequency offset Δf = |fr - fn| ≤ 50 Hz, where fn is the nearest natural frequency; the vibration energy integral value Er = ∫Ar 2 dt exceeds the threshold Eth within a 0.1 - second window;

[0016] A4: Generate a regulation instruction including the frequency offset compensation amount Δfc = Kp·Δf + Kd·d(Δf) / dt to drive the servo motor to perform anti - phase vibration suppression; where Kp is the proportional coefficient and Kd is the differential coefficient.

[0017] Further, it also includes an abnormal working condition handling mechanism. When any of the following is detected, start the emergency stop protocol: the arc voltage volatility exceeds ±15% within three consecutive sampling periods; the molten pool area expansion rate monitored by the thermal imager is greater than 2 times the standard deviation of the preset safety value; a resonance peak coinciding with the equipment natural frequency appears in the vibration spectrum and the amplitude continues to increase.

[0018] Further, integrate a digital twin system, including a processing process simulation module based on finite element analysis, a dynamic optimization engine, and a process parameter self - learning database; among them, the processing process simulation module based on finite element analysis predicts the thermal stress distribution within the next 2 seconds in real - time, the dynamic optimization engine feeds back the prediction result to the central controller to adjust the control parameters in advance, and the process parameter self - learning database records historical processing data and generates a material - process mapping relationship model.

[0019] Further, in A3, the compensation mechanism includes the coordinated control of an adaptive notch filter. The specific operations include: dynamically adjusting the notch center frequency fc = fr ± Δf according to the Δf calculated in real time offset , where Δf offset = 0.2·(fn - fr); adopting a variable Q-value control strategy, where the Q-value increases linearly with the increase of the Er / Eth ratio, and the maximum Q-value is limited to 15.

[0020] Further, in A4, the generation of the regulation instruction adopts a phase compensation mechanism, which specifically includes: calculating the phase difference θ between the vibration signal and the motor drive signal through cross-correlation analysis; superimposing a phase correction term Δθ = π·sgn(θ - θ0)·min(|θ - θ0| / 180, 0.25) on the basis of Δfc, where θ0 is the ideal cancellation phase angle; finally outputting the frequency instruction f cmd = f base + Δfc·(1 + a·Δθ), where a is the phase coupling coefficient, and f base is the frequency reference value.

[0021] Further, a harmonic coupling suppression strategy is also included. When the following situations are detected, multi-frequency point joint compensation is started: in the frequency domain analysis, there are ±n×50Hz harmonic components centered on fr, where n = 1, 2, 3; the sum of the energies of each harmonic exceeds 30% of the fundamental wave energy; at this time, a composite regulation instruction f cmd = Σ[Δf cn ·exp(-j2πkΔf cn )], where Δf cn is the offset of the center frequency, and k is the harmonic order weight factor.

[0022] Further, an integrated resonance warning mechanism is included, including establishing a monitoring of the change rate of the vibration energy accumulation index CEI = Σ(Er / Eth) over time; when the slope of CEI is greater than the preset safety curve within a 10-second period, triggering preventive frequency compensation in advance; automatically adjusting the threshold Eth in step A3 to 80% of the original value.

[0023] Compared with the prior art, the beneficial effects of the present invention are:

[0024] The present invention matches the target processing mode according to a preset process parameter library, and based on the deviation between the dynamic parameter feedback signal and the preset threshold, uses a fuzzy PID algorithm to generate a first control command to synchronously adjust the axial feed speed of the servo motor, the power output curve of the arc generator, and the coolant injection timing, so as to form a dynamic balance between the arc energy density gradient and the workpiece heat conduction rate. The mechanical resonance frequency during the processing is monitored by a vibration spectrum analysis unit. When the detected resonance intensity exceeds the safety threshold, a frequency compensation algorithm is triggered to generate a second control command to real-time correct the driving frequency of the servo motor. This not only improves the control accuracy, makes the arc striking process more stable, but also improves the response speed, shortens the arc striking time, enhances the adaptability, can adapt to the requirements of different materials and processes, reduces the energy consumption, and improves the production efficiency. Brief Description of the Drawings

[0025] Figure 1 It is a schematic flow chart of a control method for an arc striking machine based on multi-parameter collaborative optimization of the present invention;

[0026] Figure 2 It is a schematic flow chart of the trigger frequency compensation algorithm of the present invention to generate a second control command. Detailed Embodiments

[0027] The technical solutions of the present invention will be further described below in conjunction with the drawings and specific embodiments.

[0028] Embodiment 1

[0029] As Figure 1 shown, a control method for an arc striking machine based on multi-parameter collaborative optimization, the working steps include:

[0030] S1: The dynamic parameters of the arc are collected in real time through an arc voltage sensor, a current sensor, and an infrared thermal imager; among them, the dynamic parameters include voltage value, current intensity, arc shape characteristics, and workpiece surface temperature distribution;

[0031] S2: The central controller matches the target processing mode according to the preset process parameter library, and based on the deviation between the dynamic parameter feedback signal and the preset threshold, uses a fuzzy PID algorithm to generate a first control command;

[0032] S3: The motion control module receives the first control command and synchronously adjusts the axial feed speed of the servo motor, the power output curve of the arc generator, and the coolant injection timing, so as to form a dynamic balance between the arc energy density gradient and the workpiece heat conduction rate;

[0033] S4: The mechanical resonance frequency during the processing is monitored by a vibration spectrum analysis unit. When the detected resonance intensity exceeds the safety threshold, a frequency compensation algorithm is triggered to generate a second control command to real-time correct the driving frequency of the servo motor.

[0034] As shown Figure 2 in FIG. 4, in S4, the specific steps for triggering the frequency compensation algorithm to generate the second regulation instruction include:

[0035] A1: Synchronously collect mechanical vibration signals through a triaxial accelerometer array distributed on the spindle box, electrode clamping arm, and base platform at a sampling rate greater than or equal to 5 kHz;

[0036] A2: After the vibration signal processing module performs band-pass filtering on the original signal, extract the current dominant resonance frequency fr and the corresponding amplitude Ar through fast Fourier transform; among them, the cut-off frequency of the band-pass filtering is 100 Hz - 2 kHz;

[0037] A3: The frequency compensation decision unit compares fr with the device natural frequency database, and triggers the compensation mechanism when any of the following conditions is met: the frequency offset Δf = |fr - fn| ≤ 50 Hz, where fn is the nearest natural frequency; the vibration energy integral value Er = ∫Ar 2 dt exceeds the threshold Eth within a 0.1-second window;

[0038] A4: Generate a regulation instruction including the frequency offset compensation amount Δfc = Kp·Δf + Kd·d(Δf) / dt, and drive the servo motor to perform anti-phase vibration suppression; where Kp is the proportional coefficient and Kd is the differential coefficient.

[0039] The working principle of a kind of arc welding machine control method based on multi-parameter collaborative optimization according to Embodiment 1 is:

[0040] In S2, the fuzzy PID algorithm adopts a double-input and three-output structure. The input variables are the arc length deviation amount ΔL and the temperature gradient ΔT in the heat affected zone. The output variables are the servo motor speed correction coefficient Kv, the arc current duty cycle correction coefficient Kc, and the coolant flow rate correction coefficient Kw. The associated logic of the fuzzy rule base includes: when ΔL > 0 and ΔT < 0, increase Kv and decrease Kc; when ΔL < 0 and ΔT > the critical molten pool maintenance temperature, start the exponential compensation mode of Kw.

[0041] During the process of starting the exponential compensation mode of Kw, the arc voltage sensor collects the voltage volatility δV, the infrared thermal imager detects the core temperature T of the molten pool core and the temperature gradient ▽T in the heat affected zone, and the spectral analyzer obtains the metal vapor particle concentration C plasma ; when one of the following composite trigger conditions is met, activate the exponential compensation mode:

[0042] The first condition: δV > 15% and ▽T > 200 °C / mm;

[0043] The second condition: C plasma exceeds the material gasification threshold Cvapor 80% of

[0044] The third condition: dT core / dt > 50 °C / ms;

[0045] The variable exponent flow correction algorithm is where K base is the coefficient reference value, T crit is the critical molten pool maintenance temperature specific to the material, α and β are dynamic adjustment coefficients, automatically matched according to the material thermal property database, where λ material is the material thermal conductivity, λ ref is the reference benchmark value of the material thermal conductivity, t dwell is the arc residence time, t base is the time reference value, σy is the material yield strength, σy ref is the reference benchmark value of the material yield strength; The coolant flow valve and the injection angle servo are synchronously adjusted through the double closed-loop control module, so that the coolant coverage area always envelopes the molten pool expansion front.

[0046] In S3, the power output curve adjustment adopts a segmented pulse modulation strategy, including: applying a strong current pulse of 200 - 300 A during the arc starting stage, with a pulse width of 50 - 100 ms; after entering the stable processing stage, it switches to a square wave output with a variable duty cycle, and its fundamental frequency is negatively correlated with the thermal diffusion coefficient of the workpiece material.

[0047] An arc welding machine control method based on multi-parameter collaborative optimization further includes an abnormal condition handling mechanism. When any of the following is detected, an emergency shutdown protocol is initiated: the arc voltage fluctuation rate exceeds ±15% within 3 consecutive sampling periods; the molten pool area expansion rate monitored by the thermal imager is greater than 2 times the standard deviation of the preset safety value; a resonance peak that coincides with the natural frequency of the equipment appears in the vibration spectrum and the amplitude continues to increase.

[0048] An arc welding machine control method based on multi-parameter collaborative optimization integrates a digital twin system, including a machining process simulation module based on finite element analysis, a dynamic optimization engine, and a process parameter self-learning database; among them, the machining process simulation module based on finite element analysis predicts the thermal stress distribution within the next 2 seconds in real time, the dynamic optimization engine feeds the prediction results back to the central controller to adjust the control parameters in advance, and the process parameter self-learning database records historical machining data and generates a material-process mapping relationship model.

[0049] The machining process simulation module based on finite element analysis predicts the thermal stress distribution within the next 2 seconds in real time, and its specific steps include:

[0050] B1. Build a multi-physics coupling finite element model and integrate the thermal-mechanical-electrical coupling equations. Among them, the dynamic Gaussian distribution function of the moving arc heat source is introduced into the heat conduction equation, the Johnson-Cook plasticity model is used in the constitutive equation to characterize the high-temperature rheological properties of materials, and the modified Maxwell equations are used for electromagnetic field calculation to describe the characteristics of high-frequency arcs.

[0051] B2. Receive the current processing parameters collected by the sensor through the real-time data interface, including the arc position (x, y, z), power P(t), and coolant flow rate v(t).

[0052] B3. Perform dynamic model order reduction processing: Use the Krylov subspace projection method to compress the full-order model dimension to 5%-10% of the original model, and extract the first 20 leading modes based on POD (Proper Orthogonal Decomposition) to construct the reduced basis.

[0053] B4. On the basis of the reduced model, integrate the LSTM time series prediction module. Using the thermal history data of the previous 1 second as the input, predict the temperature field T(x, y, z, t + Δt) and thermal stress field σ(x, y, z, t + Δt) within the next 2 seconds, and the update period is less than or equal to 50 ms.

[0054] The dynamic optimization engine feeds back the prediction results to the central controller to adjust the control parameters in advance. The specific steps include:

[0055] C1. Build a multi-level optimization framework, including the process layer, control layer, and decision layer. Among them, the process layer receives the thermal stress distribution σ within the future Δt time window predicted by the finite element simulation module pred and the temperature gradient field ▽T pred , the control layer analyzes the current control parameter set U = {u1, u2,..., un} of the central controller, and the decision layer runs the multi-objective optimization algorithm based on NSGA-II and generates the Pareto front solution set.

[0056] C2. Calculate the contribution matrix of the control parameters to the prediction index through the sensitivity analysis module where σ max is the predicted maximum thermal stress, and T peak is the temperature peak;

[0057] C3. The dynamic optimization engine selects the optimal control strategy according to the real-time working conditions: Use the gradient descent method to optimize along the main direction of the J matrix in the stable processing stage, and switch to the particle swarm optimization algorithm for global search in the transient mutation stage.

[0058] C4. Send the optimized parameter set U* to the central controller through the OPC UA protocol to perform preventive adjustment ahead of Δt time.

[0059] The process parameter self-learning database records historical processing data and generates a material-process mapping relationship model. The specific steps include:

[0060] D1. Construct a multi-source heterogeneous data acquisition system to obtain process parameters, quality parameters, and environmental parameters in real time. Among them, the process parameters include arc voltage V(t), current I(t), and feed rate F(t), the quality parameters include penetration depth H, weld appearance S, and residual stress σ res , and the environmental parameters include environmental humidity Substrate pre-tightening force P pre ;

[0061] D2. The process parameter self-learning database performs data purification, including removing abnormal process data points through the Grubbs criterion, using the SMOTE oversampling technique to balance the data distribution of different material categories, and performing dynamic time warping (DTW) alignment processing on spatio-temporal asynchronous data;

[0062] D3. Construct a material-process mapping model based on the Transformer architecture. Among them, the encoder layer fuses material property features (yield strength σ y , thermal conductivity λ, specific heat capacity c p ), the decoder layer outputs the recommended values and confidence intervals of process parameters, and the attention mechanism focuses on the weights of key process influencing factors;

[0063] D4. Update the model parameters through an online incremental learning mechanism, including triggering model fine-tuning when the deviation of the material properties of the new processing task from the historical data is greater than 15%, and using the elastic weight consolidation (EWC) algorithm to prevent catastrophic forgetting.

[0064] In A3, the compensation mechanism includes the coordinated control of an adaptive notch filter. The specific operations include: dynamically adjusting the notch center frequency fc = fr ± Δf according to the Δf calculated in real time offset , where, Δf offset = 0.2·(fn - fr); adopt a variable Q-value control strategy, and the Q value increases linearly with the increase of the Er / Eth ratio, and the maximum Q value is limited to 15.

[0065] For the calculation of the Q value, Q base = 5 + 10*(1 - |fr - fn| / 100); if Er / Eth > 1, Q adj = Q base *[1 + 0.2*(Er / Eth - 1)]; if Er / Eth ≤ 1, Q adj = Q base *0.8.

[0066] In A4, the generation of the regulation instruction adopts a phase compensation mechanism, which specifically includes: calculating the phase difference θ between the vibration signal and the motor drive signal through cross-correlation analysis; adding a phase correction term Δθ = π·sgn(θ - θ0)·min(|θ - θ0| / 180, 0.25) on the basis of Δfc, where θ0 is the ideal cancellation phase angle; finally outputting the frequency instruction f cmd = f base + Δfc·(1 + a·Δθ), where a is the phase coupling coefficient and f base is the frequency reference value.

[0067] Calculating the phase difference θ between the vibration signal and the motor drive signal through cross-correlation analysis, and its specific operations include:

[0068] E1. Constructing a multi-channel synchronous acquisition system, including: an array of vibration accelerometers installed in a three-dimensional orthogonal layout at the power output end, a motor encoder and a current loop detection module synchronously capturing the drive signal, and a temperature sensor real-time monitoring the thermal deformation of the transmission chain;

[0069] E2. Performing adaptive signal preprocessing: performing wavelet threshold denoising on the vibration signal, and the threshold function is: where σ is the noise standard deviation, ΔT is the temperature rise, and T ref is the temperature reference value; performing dynamic resampling on the drive signal, and the sampling rate is adaptively adjusted according to the rotational speed fluctuation rate: fs = f base *(1 + |w real - w set | / w set ), where w set is the set phase and w real is the actual phase;

[0070] E3. Calculating the composite cross-correlation function: calculating the delay Δt1 according to the time-domain cross-correlation in the main correlation path, and analyzing and obtaining the phase difference according to the frequency-domain coherence function in the auxiliary path

[0071] E4. Fusing Δt1 and through the particle filter algorithm and outputting the optimized phase difference where the weight coefficients w1 and w2 are dynamically adjusted according to the signal-to-noise ratio.

[0072] The method for triggering the frequency compensation algorithm to generate the second regulation instruction further includes a harmonic coupling suppression strategy. When the following situations are detected, multi-frequency point joint compensation is started: in the frequency-domain analysis, there are ±n×50Hz harmonic components centered on fr, where n = 1, 2, 3; the sum of the energies of each harmonic exceeds 30% of the fundamental wave energy; at this time, a composite regulation instruction f cmd = Σ[Δf cn·exp(-j2πkΔf cn )], where Δf cn is the offset of the center frequency, and k is the harmonic order weighting factor.

[0073] The method for triggering the frequency compensation algorithm to generate the second regulation instruction integrates a resonance warning mechanism, including establishing a monitoring of the change rate of the vibration energy accumulation index CEI = Σ(Er / Eth) over time; when the slope of CEI is greater than the preset safety curve within a 10-second period, triggering preventive frequency compensation in advance; automatically adjusting the threshold Eth in step A3 to 80% of the original value to improve the system sensitivity.

[0074] The above specific embodiments are only several preferred embodiments of the present invention. Based on the technical solution of the present invention and the relevant inspirations of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A control method for an arc-pulling machine based on multi-parameter collaborative optimization, characterized in that, The working steps include: S1: Real-time collect the dynamic parameters of the arc through an arc voltage sensor, a current sensor and an infrared thermal imager; among them, the dynamic parameters include voltage value, current intensity, arc shape characteristics and workpiece surface temperature distribution; S2: The central controller matches the target processing mode according to the preset process parameter library, and based on the deviation between the dynamic parameter feedback signal and the preset threshold, uses the fuzzy PID algorithm to generate the first regulation instruction; S3: The motion control module receives the first regulation instruction and synchronously adjusts the axial feed speed of the servo motor, the power output curve of the arc generator and the coolant injection timing, so that the arc energy density gradient and the workpiece heat conduction rate form a dynamic balance; S4: Monitor the mechanical resonance frequency during the processing through the vibration spectrum analysis unit. When the detected resonance intensity exceeds the safety threshold, trigger the frequency compensation algorithm to generate the second regulation instruction and correct the servo motor drive frequency in real time.

2. The arc striking machine control method based on multi-parameter collaborative optimization according to claim 1, wherein: In S2, the fuzzy PID algorithm adopts a two-input three-output structure. The input variables are the arc length deviation ΔL and the temperature gradient of the heat affected zone ΔT. The output variables are the servo motor speed correction coefficient Kv, the arc current duty cycle correction coefficient Kc and the coolant flow correction coefficient Kw. The associated logic of the fuzzy rule base includes: when ΔL>0 and ΔT<0, increase Kv and decrease Kc; when ΔL<0 and ΔT>the critical molten pool maintenance temperature, start the exponential compensation mode of Kw.

3. A control method for an arc-drawing machine based on multi-parameter collaborative optimization according to claim 1, wherein: In S3, the power output curve adjustment adopts a segmented pulse modulation strategy, including: applying a strong current pulse of 200-300A with a pulse width of 50-100ms during the arc starting stage; after entering the stable processing stage, switch to a square wave output with a variable duty cycle, and its fundamental frequency is negatively correlated with the thermal diffusion coefficient of the workpiece material.

4. A method for controlling an arc-drawing machine based on multi-parameter collaborative optimization according to claim 1, characterized in that In S4, the specific steps for triggering the frequency compensation algorithm to generate the second regulation instruction include: A1: Synchronously collect mechanical vibration signals through a three-axis accelerometer array distributed on the spindle box, the electrode clamping arm and the base platform at a sampling rate greater than or equal to 5kHz; A2: After the vibration signal processing module performs band-pass filtering on the original signal, extract the current dominant resonance frequency fr and the corresponding amplitude Ar through fast Fourier transform; among them, the cut-off frequency of the band-pass filtering is 100Hz-2kHz; A3: The frequency compensation decision unit compares fr with the device's inherent frequency database and triggers the compensation mechanism when any of the following conditions are met: the frequency offset Δf = |fr - fn| ≤ 50 Hz, where fn is the nearest neighbor inherent frequency; the vibration energy integral value Er = ∫Ar 2 dt exceeds the threshold Eth within a 0.1-second window; A4: Generate a regulation instruction including the frequency offset compensation amount Δfc = Kp·Δf + Kd·d(Δf) / dt, and drive the servo motor to perform anti-phase vibration suppression; where Kp is the proportional coefficient and Kd is the differential coefficient.

5. A control method for an arc-pulling machine based on multi-parameter collaborative optimization according to claim 1, characterized in that: It also includes an abnormal working condition handling mechanism. When any of the following is detected, start the emergency stop protocol: the arc voltage volatility exceeds ±15% within 3 consecutive sampling periods; the molten pool area expansion rate monitored by the thermal imager is greater than 2 times the standard deviation of the preset safety value; a resonance peak that coincides with the natural frequency of the equipment appears in the vibration spectrum and the amplitude continues to increase.

6. A control method for an arc-pulling machine based on multi-parameter collaborative optimization according to claim 1, characterized in that: An integrated digital twin system, including a machining process simulation module based on finite element analysis, a dynamic optimization engine, and a process parameter self-learning database; among them, the machining process simulation module based on finite element analysis predicts the thermal stress distribution within the next 2 seconds in real time, the dynamic optimization engine feeds the prediction results back to the central controller to adjust the control parameters in advance, and the process parameter self-learning database records historical machining data and generates a material-process mapping relationship model.

7. A method for controlling an arc-drawing machine based on multi-parameter collaborative optimization according to claim 4, characterized in that: In A3, the compensation mechanism includes the coordinated control of an adaptive notch filter, and the specific operations include: dynamically adjusting the notch center frequency fc = fr ± Δf according to the Δf calculated in real time offset , where Δf offset = 0.2·(fn - fr); adopting a variable Q-value control strategy, the Q-value increases linearly with the increase of the Er / Eth ratio, and the maximum Q-value is limited to 15.

8. A control method for an arc-pulling machine based on multi-parameter collaborative optimization according to claim 4, characterized in that: In A4, the control command generation adopts a phase compensation mechanism, which specifically includes: calculating the phase difference θ between the vibration signal and the motor drive signal through cross-correlation analysis; superimposing a phase correction term Δθ = π·sgn(θ - θ0)·min(|θ - θ0| / 180, 0.25) on the basis of Δfc, where θ0 is the ideal cancellation phase angle; and finally outputting the frequency command f cmd = f base + Δfc·(1 + a·Δθ), where a is the phase coupling coefficient and f base is the frequency reference value.

9. A method for controlling an arc-pulling machine based on multi-parameter collaborative optimization according to claim 4, characterized in that: It also includes a harmonic coupling suppression strategy that initiates multi-frequency point joint compensation when the following conditions are detected: in the frequency domain analysis, there are ±n×50Hz sub-harmonic components centered on fr, where n = 1, 2, 3; the sum of the energies of each sub-harmonic exceeds 30% of the fundamental wave energy; at this time, a composite control instruction f cmd = Σ[Δf cn ·exp(-j2πkΔf cn )], where Δf cn is the offset of the center frequency, and k is the harmonic order weight factor.

10. A method for controlling an arc-drawing machine based on multi-parameter collaborative optimization according to claim 4, characterized in that: An integrated resonance warning mechanism, including establishing a monitoring of the change rate of the vibration energy accumulation index CEI = Σ(Er / Eth) over time; when the slope of CEI is greater than the preset safety curve within a 10-second cycle, triggering preventive frequency compensation in advance; automatically adjusting the threshold Eth in step A3 to 80% of the original value.

Images