An integrated automation control method with redundancy fault tolerance function

Abstract

The present application relates to the field of automatic control technology, specifically to a comprehensive automatic control method with redundancy fault-tolerant function, comprising the following steps: three-axis sensor collects vibration data, extracts characteristic parameters in combination with angular displacement pulse, K-means clustering evaluates modal quality, deformation gradient threshold value triggers abnormal identification, harmonic distortion is more than 15% to switch redundancy channel, fusion characteristics calculate power factor, phase switching loop, adjust PWM duty cycle in three intervals, regression correction generates brake curve verification execution, execute corrugated roller machine redundancy channel switching and parameter reloading. In the present application, three-axis vibration is combined with angular displacement to construct characteristic parameter clustering division mode evaluation quality, deformation correlation identifies abnormality, harmonic monitoring dynamically switches path to stabilize data flow, distortion rate calculates power factor, phase switching loop, adjusts PWM duty cycle according to power factor to improve precision, constructs vibration-modal-redundancy closed-loop system, realizes state identification and energy efficiency improvement.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology, and in particular to a comprehensive automatic control method with redundancy and fault tolerance. Background Technology

[0002] The field of automatic control technology encompasses systematic methods for real-time monitoring and control of industrial production processes. Its core components include signal acquisition, state recognition, feedback adjustment, and execution control. This technology analyzes data collected by sensors and adjusts control commands in real time to achieve orderly coordination among various types of equipment in the production process. Automatic control is widely used in industries such as petrochemicals, power generation, papermaking, and metallurgy to improve production efficiency, stabilize product quality, and reduce energy consumption. In the production of corrugated cardboard, the automatic control system needs to control key process stages such as paper feeding, glue application, heating and drying, and compression molding. Therefore, research in this field focuses on real-time monitoring of process parameters and multi-station coordinated control.

[0003] One type of automated control method with redundancy and fault tolerance is designed for continuous corrugated cardboard production lines. It introduces a multi-channel signal synchronous acquisition and status comparison mechanism into the control system, backing up control tasks at critical nodes by setting redundant paths. If any path malfunctions, it automatically switches to the backup path to maintain continuous execution of control tasks. The method includes fault tolerance assessment of production parameters, consistency verification of signal source states, and fault switching judgment techniques for control logic. It employs a multi-node comparative detection method to cross-verify the main control signal and redundant signals, executes control switching based on fixed priority decision logic, and utilizes a centralized control platform for unified command allocation. Even in the event of partial functional loss, it can maintain the logical closed loop of the overall control chain.

[0004] Existing technologies primarily rely on single-path sensor signal acquisition and feedback control logic during actual operation. This makes it difficult to handle feedback errors caused by fluctuations in local sensor performance or abnormal acquisition signals, leading to system response delays or distorted control commands. For example, in high-speed continuous production of corrugated cardboard, if instantaneous fluctuations in the pressure roller vibration signal are not promptly identified and isolated, it may cause unevenness in subsequent pressing processes, affecting the cardboard forming quality. Existing control systems lack mechanisms for real-time comparison of the consistency of multiple signal source states, resulting in delayed judgment of critical node states. This is particularly problematic when equipment operating modes change or initial anomalies occur in the transmission system, making accurate identification and response difficult. Furthermore, existing systems lack redundant switching mechanisms for control links. When the main control path fails, manual intervention or a complete system interruption is often required, affecting the stability of continuous production. The lack of real-time monitoring methods for some process parameters, such as the pressure roller power factor, leads to decreased control accuracy under load disturbances, posing risks such as excessive energy consumption and sluggish control loop response. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and propose a comprehensive automated control method with redundancy and fault tolerance.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a comprehensive automated control method with redundancy and fault tolerance, comprising the following steps:

[0007] S1: The axial vibration acceleration waveform of the pressure roller on the cardboard processing equipment is obtained by a triaxial MEMS sensor. The peak-valley difference of the fundamental frequency band is extracted from the waveform. The axial vibration characteristic parameters of the corrugated roller are calculated based on the peak-valley difference and the angular displacement pulse signal output by the rotary encoder of the corrugated roller machine drive shaft.

[0008] S2: The K-means clustering algorithm is used to classify the operating modes of the corrugated roll bearing housing and calculate the profile coefficient. Combined with the error of the cardboard pressure sensor, the roll surface deformation gradient compensation mechanism is activated. When the Pearson correlation coefficient between the profile coefficient and the deformation gradient change trend exceeds 0.8, an abnormal identifier of the corrugated roll gearbox is generated.

[0009] S3: Based on the abnormal identification of the corrugated roller gearbox, detect the dual-channel harmonic monitoring, calculate the dual-channel harmonic distortion rate of the pressure roller, and activate the backup channel when the distortion rate exceeds 15%. Calculate the power factor of the corrugated roller motor based on the harmonic distortion rate and the vibration characteristic parameters.

[0010] S4: Based on the power factor of the corrugated roller motor, construct a dual redundant control loop for the paperboard linear speed. When the phase lag of the main loop is greater than 30°, switch to the backup loop. Divide the power factor into three intervals: 0.8-1.0, 0.6-0.8, and below 0.6, and generate the PWM duty cycle parameters of the corrugated roller drive motor.

[0011] As a further aspect of the present invention, the axial vibration characteristic parameters of the corrugated roll specifically include the peak-to-valley difference of the fundamental frequency band vibration, the integral value of the angular displacement pulse signal, and the harmonic component of the axial acceleration. The abnormality indicator of the corrugated roll gearbox specifically refers to the deformation gradient compensation activation state, the dual-channel harmonic distortion rate threshold, and the Pearson correlation coefficient verification mark. The power factor of the corrugated roll motor includes the main circuit phase lag angle, redundant channel switching, and harmonic distortion rate. The dual redundant control circuit for the paperboard linear speed specifically includes the phase lag compensation parameter, the duty cycle interval division standard, and the power factor segmented adjustment factor.

[0012] As a further aspect of the present invention, the specific steps for acquiring the axial vibration acceleration waveform of the pressure roller on the cardboard processing equipment using a triaxial MEMS sensor, extracting the peak-to-valley difference of the fundamental frequency band from the waveform, and calculating the axial vibration characteristic parameters of the corrugated roller based on the peak-to-valley difference and the angular displacement pulse signal output by the rotary encoder of the corrugated roller machine drive shaft include:

[0013] S101: The axial vibration signal of the pressure roller on the cardboard processing equipment is collected by a triaxial MEMS sensor. The signal is denoised by the sliding window method, the data in the window is filtered by median filtering, the acceleration waveform in the continuous processing cycle is extracted, and the axial vibration waveform is generated.

[0014] S102: Based on the axial vibration waveform, perform a fast Fourier transform to identify the frequency component corresponding to the maximum amplitude, reconstruct the signal envelope using cubic spline interpolation, and calculate the fundamental frequency peak-valley difference between adjacent peaks and valleys.

[0015] S103: Establish a time axis coordinate system by calling the angular displacement pulse signal output by the rotary encoder, and perform time-domain synchronous analysis on the fundamental frequency peak-valley difference and the corresponding angular displacement of the time axis to generate axial vibration characteristic parameters.

[0016] As a further aspect of the present invention, the frequency component corresponding to the maximum amplitude is identified using the following formula:

[0017]

[0018] Among them, A max Δf represents the amplitude corresponding to the maximum amplitude component. k ω represents the absolute value of the difference between the k-th adjacent frequency component and the maximum amplitude frequency, ω is the angular velocity of the vibration system, and Δt is the frequency of maximum amplitude. k (Time-domain interval corresponding to adjacent frequency components), where n represents the number of adjacent frequency points sampled, α is a smoothing factor determined based on the signal-to-noise ratio, and f max This represents the fundamental frequency value corresponding to the maximum amplitude, and -n represents the symmetrical frequency point position of the sample.

[0019] As a further aspect of the present invention, the specific steps for classifying the operating modes of the corrugated roll bearing housing and calculating the profile coefficient using the K-means clustering algorithm based on the vibration characteristic parameters, and combining this with the error of the cardboard pressure sensor to activate the roll surface deformation gradient compensation mechanism, and generating an abnormality identifier for the corrugated roll gearbox when the Pearson correlation coefficient between the profile coefficient and the deformation gradient change trend exceeds 0.8, include:

[0020] S201: Call the K-means clustering algorithm with the input of the axial vibration feature parameters, calculate the similarity between the sample and the class center through Mahalanobis distance, determine the number of iterations using the elbow rule, update the cluster center until the change in intra-class variance is stable, calculate the average distance from each sample point to its class center, and generate the running modality classification result.

[0021] S202: Calculate the profile coefficient based on the classification results of the operating modes, determine the sample density by using the maximum and minimum inter-class distances, and establish a linear regression equation between the error rate and the profile coefficient by combining the error rate output by the cardboard pressure sensor. When the regression coefficient t-test value shows statistical significance, generate a compensation activation signal.

[0022] S203: Trigger the roll surface deformation gradient compensation mechanism according to the compensation activation signal, extract the deformation gradient change rate within the compensation period, calculate the Pearson correlation coefficient between the change rate and the profile coefficient, and generate an abnormality mark for the corrugated roll gearbox when the correlation coefficient exceeds 0.8 three times in a row.

[0023] As a further aspect of the present invention, the linear regression equation for the error rate and the profile coefficient is established using the formula:

[0024]

[0025] Where, ε r The error rate, expressed in μ, is the output of the cardboard pressure sensor. s The arithmetic mean of the profile coefficients. d represents the silhouette coefficient value of the i-th sample. max d represents the maximum inter-class distance. min This represents the minimum inter-class distance, where N is the effective sample size, and β is a moderating factor determined based on the t-test value and the sample size. It represents the reciprocal of the maximum inter-class distance.

[0026] As a further aspect of the present invention, based on the dual-channel harmonic monitoring of the corrugated roller gearbox abnormality detection, the dual-channel harmonic distortion rate of the pressure roller is calculated. When the distortion rate exceeds 15%, the backup redundant channel is activated. The specific steps for calculating the power factor of the corrugated roller motor based on the harmonic distortion rate and the vibration characteristic parameters include:

[0027] S301: Based on the abnormal identification of the corrugated roller gearbox, dual-channel harmonic monitoring is triggered, the current harmonic components of the main and backup channels are collected, the fundamental periodic signal is obtained by phase-locked loop synchronous sampling technology, the effective values ​​of harmonics and the effective values ​​of the fundamental wave are extracted by wavelet packet transform and the root mean square value is calculated to generate dual-channel harmonic distortion rate;

[0028] S302: Call the dual-channel harmonic distortion rate to evaluate the channel status. When the distortion rate of the main channel exceeds 15% for two consecutive sampling cycles and is higher than that of the backup channel, perform signal switching and generate the redundant channel activation state according to the channel switching time difference and the distortion rate change gradient.

[0029] S303: Divide the power factor interval according to the activation state of the redundant circuit, establish the mapping curve between the power factor and the PWM frequency using cubic spline interpolation, and calculate the power factor of the corrugated roller motor by combining the rated speed of the motor and the current load rate.

[0030] As a further aspect of the present invention, a dual-redundant control loop for the cardboard linear speed is constructed based on the power factor of the corrugated roller motor. When the phase lag of the main loop is greater than 30°, the loop switches to the backup loop. The power factor is divided into three intervals: 0.8-1.0, 0.6-0.8, and below 0.6. The specific steps for generating the PWM duty cycle parameters of the corrugated roller drive motor include:

[0031] S401: Based on the power factor of the corrugated roller motor, a dual redundant control loop is constructed. When the phase lag is detected to continuously exceed 30° electrical angle, the cumulative phase offset is recorded to generate the phase stability of the main loop.

[0032] S402: Call the main circuit phase stability to determine the circuit switching, establish a phase lag moving average sequence, and when the average of two consecutive sampling points exceeds the upper limit of the stable interval, trigger the switching of the backup redundant circuit signal, and generate the redundant circuit activation state according to the switching time difference and the phase recovery gradient.

[0033] S403: Divide the power factor range according to the activation state of the redundant circuit, set the power factor range of 0.8-1.0 as the high efficiency range, 0.6-0.8 as the normal range, and below 0.6 as the abnormal range. Use piecewise linear interpolation to generate PWM frequency mapping values, and calculate the PWM duty cycle parameters of the corrugated roller drive motor in combination with the rated speed of the motor.

[0034] As a further aspect of the present invention, the method further includes:

[0035] S5: The PWM duty cycle parameter of the corrugated roll drive motor is corrected by regression analysis. The braking current PID control curve is generated by combining the power factor of the corrugated roll motor. The effectiveness of the curve is verified by the power factor of the backup circuit. The redundant channel switching and parameter reload of the corrugated roll machine are then performed.

[0036] The braking current PID control curve includes a combination of proportional, integral, and derivative coefficients, power factor verification error rate, and parameter overload triggering conditions.

[0037] As a further aspect of the present invention, regression analysis is used to correct the PWM duty cycle parameters of the corrugated roll drive motor, and a braking current PID control curve is generated by combining the power factor of the corrugated roll motor. The effectiveness of the curve is verified by the power factor of the backup circuit. The specific steps for performing redundant channel switching and parameter reloading of the corrugated roll machine include:

[0038] S501: Call the PWM duty cycle parameters of the corrugated roll drive motor for parameter correction, use the stepwise regression method to eliminate the influence of multicollinearity, calculate the partial correlation coefficient between the duty cycle parameters and the motor speed, and generate the corrected duty cycle parameters;

[0039] S502: Based on the corrected duty cycle parameters and the power factor of the corrugated roller motor, the critical period of the system is determined by the critical oscillation method, and the rise time and overshoot are measured according to the constant amplitude oscillation waveform to generate the braking current PID control curve.

[0040] S503: Call the backup circuit power factor to verify the braking current PID control curve, calculate the difference between the main circuit power factor and the curve prediction value, and when the difference is within the allowable range for three consecutive sampling cycles, trigger the channel switching command, load the updated parameters, and generate the corrugated roller machine parameter overload state.

[0041] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0042] In this invention, the axial vibration acceleration waveform of the pressure roller on the processing equipment is acquired in three axes. The peak-to-valley difference within the fundamental frequency band of the vibration is extracted, and the vibration characteristic parameters are calculated by combining the angular displacement pulse signal from the rotary encoder. This allows for a more accurate quantitative expression of the dynamic characteristics of the equipment under high-speed operation. This characteristic parameter is input into K-means clustering for operational mode classification, and the clustering effect is evaluated using the profile coefficient. A correlation analysis mechanism between multi-source data is established by combining the roller surface deformation gradient change guided by the cardboard pressure error. When the profile coefficient and deformation gradient change trends show a high linear correlation, abnormal gearbox operating status is identified, enhancing the early identification capability of potential structural faults. The anomaly identification results guide dual-channel harmonic monitoring, switching the signal path based on the distortion rate. This ensures the stability of the data acquisition and feedback loop when the main channel is interfered with or its performance degrades. Furthermore, the vibration characteristics and harmonic distortion rate reflect the motor's operating status. After calculating the power factor, loop switching control is achieved by combining the phase response. The PWM duty cycle is dynamically adjusted according to the power factor range, improving the accuracy and response sensitivity of linear speed control. The entire processing logic constructs a multi-dimensional closed-loop control system through cross-modal signal fusion, anomaly detection logic optimization, and dynamic switching of redundant loops. This achieves a comprehensive improvement in operating status, energy consumption performance, and equipment synergy, enhancing the system's stability, responsiveness, and energy efficiency under dynamic load and fault interference conditions. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the main steps of the present invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0045] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0046] Please see Figure 1 This invention provides a technical solution: a comprehensive automated control method with redundancy and fault tolerance, comprising the following steps:

[0047] S1: The axial vibration acceleration waveform of the pressure roller on the cardboard processing equipment is obtained by a triaxial MEMS sensor. The peak-valley difference of the fundamental frequency band is extracted from the waveform. The axial vibration characteristic parameters of the corrugated roller are calculated based on the peak-valley difference and the angular displacement pulse signal output by the rotary encoder of the corrugated roller machine drive shaft.

[0048] S2: The K-means clustering algorithm is used to classify the operating modes of the corrugated roll bearing housing and calculate the profile coefficient. Combined with the error of the cardboard pressure sensor, the roll surface deformation gradient compensation mechanism is activated. When the Pearson correlation coefficient between the profile coefficient and the deformation gradient change trend exceeds 0.8, an abnormal identifier of the corrugated roll gearbox is generated.

[0049] S3: Based on the abnormal identification of the corrugated roller gearbox, detect dual-channel harmonic monitoring, calculate the dual-channel harmonic distortion rate of the pressure roller, and activate the backup channel when the distortion rate exceeds 15%. Calculate the power factor of the corrugated roller motor based on the harmonic distortion rate and vibration characteristic parameters.

[0050] S4: Construct a dual-redundant control loop for the cardboard linear speed based on the power factor of the corrugated roller motor. When the phase lag of the main loop is greater than 30°, switch to the backup loop. Divide the power factor into three ranges: 0.8-1.0, 0.6-0.8, and below 0.6, and generate the PWM duty cycle parameters of the corrugated roller drive motor.

[0051] S5: The PWM duty cycle parameter of the corrugated roll drive motor is corrected by regression analysis. The braking current PID control curve is generated by combining the power factor of the corrugated roll motor. The effectiveness of the curve is verified by the power factor of the backup circuit. The redundant channel switching and parameter reload of the corrugated roll machine are then executed.

[0052] The axial vibration characteristic parameters of the corrugated roll are specifically the peak-to-valley difference of the fundamental frequency band, the integral value of the angular displacement pulse signal, and the harmonic components of the axial acceleration. The abnormal indicators of the corrugated roll gearbox specifically refer to the deformation gradient compensation activation status, the dual-channel harmonic distortion rate threshold, and the Pearson correlation coefficient verification mark. The power factor of the corrugated roll motor includes the main circuit phase lag angle, redundant channel switching, and harmonic distortion rate. The dual redundant control loop of the cardboard linear speed specifically includes the phase lag compensation parameters, the duty cycle interval division standard, and the power factor segmented adjustment factor. The braking current PID control curve includes the combination of proportional, integral, and derivative coefficients, the power factor verification error rate, and the parameter overload trigger condition.

[0053] Please see Figure 1 This invention provides a technical solution: a comprehensive automated control method with redundancy and fault tolerance, comprising the following steps:

[0054] S101: The axial vibration signal of the pressure roller on the cardboard processing equipment is collected by a triaxial MEMS sensor. The signal is denoised by the sliding window method, the data in the window is filtered by median filtering, the acceleration waveform in the continuous processing cycle is extracted, and the axial vibration waveform is generated.

[0055] During the operation of the cardboard processing equipment, a triaxial MEMS sensor collects the raw signal of the axial vibration of the upper pressure roller at a sampling frequency of 5kHz. The sliding window length is set to 200 sampling points (corresponding to a 40ms time window). The window slides 50 sampling points at a time. After sorting the data within the window by amplitude, the 100th sample value is taken as the median filter output. When the acceleration data within the window ranges from [0.3, 0.25] m / s², the signal is filtered out. 2 At that time, the median value stabilized at 0.02 m / s after sorting. 2Nearby, after processing 120 windows consecutively, 3000 data points (corresponding to a duration of 0.6 seconds) corresponding to the complete processing cycle of the equipment were captured. When the equipment speed was 1200 rpm, the time taken per revolution was determined to be 50 ms through the rotary encoder pulse signal. A total of 12 revolutions constituted a complete processing cycle. The vibration waveform showed periodic impact characteristics, with the maximum peak occurring at 375 ms on the 8th revolution, and the peak value was 0.28 m / s. 2 The valley value is 0.18 m / s 2 .

[0056] Table 1 Sliding Window Processing Parameters

[0057] Parameter name numerical values unit How to obtain Window length 200 point 1200rpm conversion sliding step 50 point Equipment vibration period 1 / 4 Median filter order 100 none Middle position of window data sorting Sampling frequency 5000 Hz Sensor technical specifications

[0058] As shown in Table 1, the window length was determined by converting the equipment rotation speed. When the sensor detected a rotation speed pulse interval of 50ms, it was calculated that there were 250 sampling points corresponding to one rotation. 80% of the data volume between two adjacent rotations was taken as the window length. The median filter order was taken as the middle position index after sorting the window data. After processing, the waveform signal-to-noise ratio was improved by 42%, and the amplitude of the high-frequency noise component was reduced from 0.12m / s. 2 Reduced to 0.05 m / s 2 .

[0059] S102: Based on the axial vibration waveform, perform fast Fourier transform to identify the frequency component corresponding to the maximum amplitude, reconstruct the signal envelope using cubic spline interpolation, and calculate the fundamental frequency peak-valley difference between adjacent peaks and valleys.

[0060] A 4096-point FFT transformation was performed on the axial vibration waveform with a duration of 0.6 seconds. The frequency resolution was 1.22 Hz, and the maximum amplitude component was identified at 243.3 Hz with an amplitude A-max = 0.15 m / s². 2 Cubic spline interpolation was performed on ±5 adjacent frequency points (n=5) to construct the signal envelope. Within the 240Hz to 246Hz range, the measured adjacent frequency points Δfk were [3, 2, 1, 0.5, 0.3]Hz, corresponding to a time interval Δt-k = [0.002, 0.0016, 0.0012, 0.0008, 0.0004] seconds. The system angular velocity ω = 2π × 20 = 125.66 rad / s (corresponding to a device rotation speed of 1200 rpm). The smoothing factor α was set to 0.12 based on a signal-to-noise ratio (SNR) of 35 dB. When calculating the fundamental frequency peak-to-valley difference, the maximum positive peak of 0.28 m / s within three adjacent periods (0.15 seconds) on the envelope was taken. 2 With negative peak -0.15m / s 2 The difference is 0.43 m / s 2 .

[0061] Among them, Amax Δf represents the amplitude corresponding to the maximum amplitude component. k ω represents the absolute value of the difference between the k-th adjacent frequency component and the maximum amplitude frequency, ω is the angular velocity of the vibration system, and Δt is the frequency of maximum amplitude. k (Time-domain interval corresponding to adjacent frequency components), where n represents the number of adjacent frequency points sampled, α is a smoothing factor determined based on the signal-to-noise ratio, and f max This represents the fundamental frequency value corresponding to the maximum amplitude, and -n represents the symmetrical frequency position of the sampled frequency. Calculation process: 1. Numerator: 0.15 2 =0.02252. Denominator calculation: When k=0: Δf0=0.5Hz, Δt_0=0.0008, The calculation results for 5 typical terms are: [0.454, 0.312, 0.285, 0.198, 0.153], with a sum of squares of 0.454. 2 +0.312 2 +0.285 2 +0.198 2 +0.153 2 =0.206+0.097+0.081+0.039+0.023=0.446, the square root is 0.6683, the coefficient term: 1+0.12 / (243.3^1 / 3=1+0.12 / 6.25=1.01924. Result: 0.0225 / (0.668)×1.0192=0.0345.

[0062] The results show that the characteristic parameter λc = 0.0345 reflects the concentration of vibration energy on the key frequency component. When λc < 0.05, it is judged as a normal working condition. When it is greater than the threshold, an early warning is triggered. The formula effectively suppresses noise interference by weighted summation of frequency differences and combines angular velocity parameters for dynamic compensation, thereby improving the accuracy of working condition judgment.

[0063] S103: Call the angular displacement pulse signal output by the rotary encoder to establish a time axis coordinate system, perform time-domain synchronous analysis on the difference between the peak and valley of the fundamental frequency and the angular displacement of the corresponding time axis, and generate axial vibration characteristic parameters.

[0064] 0.43m / s 2The fundamental frequency peak-to-valley difference is aligned with the rotary encoder pulse signal in the time domain. The encoder outputs 2000 pulses per revolution, capturing 24000 pulses within a 0.6-second period. When establishing the time axis coordinate system, the 12000th pulse (corresponding to the midpoint of the period) is taken as the time origin. The deviation between the vibration characteristic parameter timestamp and the angular displacement does not exceed ±2 pulses (corresponding to ±0.1ms). The vibration peak at 375ms corresponds to the 15000th pulse. At this time, the angular displacement is (15000 / 2000)×2π=15πrad. The axial vibration characteristic parameter at the system recording time is the correspondence between λc=0.0345 and the angular displacement of 15πrad, forming the feature vector [0.0345,15π]. Finally, an axial vibration characteristic dataset containing timestamps, correction coefficients, and characteristic parameter values ​​is generated.

[0065] Please see Figure 1 This invention provides a technical solution: a comprehensive automated control method with redundancy and fault tolerance, comprising the following steps:

[0066] S201: Call the K-means clustering algorithm to input axial vibration feature parameters, calculate the similarity between the sample and the class center through Mahalanobis distance, use the elbow rule to determine the number of iterations, update the cluster center until the intra-class variance change reaches a stable state, calculate the average distance from each sample point to its class center, and generate the running modal classification results;

[0067] When calling the K-means clustering algorithm to input axial vibration feature parameters, the peak vibration acceleration a is obtained from the sensor. peak Spectrum dominant frequency amplitude A f The three feature parameters are: root mean square value in the time domain, RMS value, and the initial cluster center is set to [2.5m / s]. 2 [15dB, 0.8g], standardized processing was performed on 120 sets of sample data collected continuously for 8 hours on the production line. When calculating the Mahalanobis distance from each sample to the class center, the covariance matrix was taken. When the iteration reaches the 5th iteration, the change in within-class variance Δ = 0.18% is less than the set threshold of 0.2%, so the iteration is terminated. The average distance from the third class sample to the center point [2.8, 17, 0.82] is calculated to be 0.214, and the average distance from the second class sample to the center point [2.3, 13, 0.75] is calculated to be 0.305. The vibration parameter combination (2.6 m / s²) is used. 2 Taking a sample with a weight of 16dB and 0.79g as an example, its Mahalanobis distances to the three class centers are 1.82, 2.15, and 1.23, respectively, and it is classified into the third class, generating a classification result containing three operating modes.

[0068] Table 2 Clustering Results of Vibration Characteristic Parameters

[0069] category Sample size Mean acceleration Amplitude mean Root mean square 1 38 2.45 14.2 0.76 2 52 2.68 16.8 0.81 3 30 2.32 12.7 0.73

[0070] S202: Calculate the profile coefficient based on the operational modality classification results, determine the sample density using the maximum and minimum inter-class distances, and establish a linear regression equation between the error rate and the profile coefficient by combining the error rate output by the cardboard pressure sensor. When the regression coefficient t-test value shows statistical significance, a compensation activation signal is generated.

[0071] Take the arithmetic mean μ of the profile coefficients s =0.62, maximum inter-class distance Minimum distance d min =0.87, adjustment factor β = 1.96 × (120 / 30) 0.5 =4.23, Where, ε r The error rate, expressed in μ, is the output of the cardboard pressure sensor. s The arithmetic mean of the profile coefficients. d represents the silhouette coefficient value of the i-th sample. max d represents the maximum inter-class distance. min This represents the minimum inter-class distance, where N is the effective sample size, and β is a moderating factor determined based on the t-test value and the sample size. The β adjustment factor, representing the reciprocal of the maximum inter-class distance, constructs a variance weighting function through a non-linear coupling mechanism. By dynamically adjusting the denominator of the t-test value, robustness testing of statistical significance under small sample conditions can be achieved. This is especially relevant when the pressure sensor error rate ε... r When the percentage is 2.8%, calculate the numerator |2.8 - 0.62| = 2.18, and the denominator... Second item When the t-test value is 2.98 > 2.62, a compensation signal is generated.

[0072] S203: Based on the compensation activation signal, the roll surface deformation gradient compensation mechanism is triggered. The deformation gradient change rate within the compensation cycle is extracted. The change rate and the profile coefficient are calculated using the Pearson correlation coefficient. When the correlation coefficient exceeds 0.8 three times in a row, an abnormality flag for the corrugated roll gearbox is generated.

[0073] Within the compensation period, deformation gradient change rate data [0.12% / h, 0.15% / h, 0.18% / h] were collected, and the covariance with the profile coefficient sequence [0.65, 0.68, 0.71] was calculated to be 0.00072, with standard deviations of σ. x =0.025 and σ y =0.025, Pearson coefficient The calculated values ​​of 0.83, 0.85, and 0.81 for three consecutive times all exceeded the threshold of 0.8. The determination of this threshold was based on the dual verification of statistical analysis of experimental data and engineering experience: First, the dynamic correlation coefficient distribution between the profile coefficient and the deformation gradient was calculated using the fault dataset. It was found that under normal operating conditions, the correlation coefficient of 80% of the samples was below 0.72, while the median correlation coefficient in the early stage of faults reached 0.85. Second, the ROC curve analysis method was used, and the optimal balance between a fault identification accuracy of 92% and a false alarm rate of 8% was achieved when 0.8 was used as the dividing point. Finally, the three sigma principle was used to verify that the threshold exceeded the range of three standard deviations (σ = 0.05) of the mean correlation coefficient under normal operating conditions (μ = 0.65), which met the criteria for judging low-probability events and triggered the anomaly indicator.

[0074] Please see Figure 1 This invention provides a technical solution: a comprehensive automated control method with redundancy and fault tolerance, comprising the following steps:

[0075] S301: Based on the abnormal identification of the corrugated roller gearbox, dual-channel harmonic monitoring is triggered, the current harmonic components of the main and backup channels are collected, the fundamental periodic signal is obtained by phase-locked loop synchronous sampling technology, the effective values ​​of harmonics and the effective values ​​of the fundamental wave are extracted by wavelet packet transform and the root mean square value is calculated to generate dual-channel harmonic distortion rate;

[0076] When the anomaly flag is triggered, the current harmonic component I is acquired from the main channel. h =[5.2A,3.8A,2.1A] corresponds to the 3rd, 5th, and 7th harmonics, with the backup channel acquiring I′. h = [4.7A, 3.2A, 1.9A], the fundamental frequency f0 = 50Hz is locked through a phase-locked loop, and the sampling frequency f is set. s =2.56kHz, perform 5-level wavelet packet decomposition on the main channel current signal, and extract the energy E of the 3rd node coefficient. h3 =0.82kJ, Energy E of the 5th node coefficient h5 =0.35kJ, fundamental effective value Total effective value of harmonics Calculate distortion rate The backup channel was calculated using the same procedure, resulting in THD′ = 3.12%.

[0077] Table 3. Dual-channel harmonic monitoring data;

[0078]

[0079] As shown in Table 3, the third harmonic component of the main channel exceeds that of the backup channel by 10.6%, but the distortion rate is still within the normal range.

[0080] S302: Call the dual-channel harmonic distortion rate to evaluate the channel status. When the distortion rate of the main channel exceeds 15% for two consecutive sampling periods and is higher than that of the backup channel, perform signal switching and generate the redundant channel activation state based on the channel switching time difference and the distortion rate change gradient.

[0081] Set the distortion rate threshold THD th =15%, normalized calculation of the 15% threshold. Among them: THD th Harmonic distortion rate threshold V max Represents the maximum permissible harmonic voltage, taken as 20% of the nominal value in the equipment specification sheet, V base The representative reference harmonic voltage is taken as 8% according to the IEEE 519-2014 standard, V rated V represents the rated voltage fluctuation coefficient, and κ represents the safety factor, where: base =8% (refer to IEEE 519-2014 Low Voltage System THD reference value), V max =20% (the maximum instantaneous distortion rate allowed by the equipment manufacturer), V rated =1.2 (rated voltage fluctuation coefficient), κ = 0.75 (safety factor). Substituting the measured data, the calculation reference range is: Apply a safety factor of 10% × 0.75 = 7.5%, and add an equipment aging factor: Rounding: 13.5% → 15%, when the main channel measures for two consecutive sampling cycles.

[0082] Given THD1 = 16.2% and THD2 = 16.8%, the corresponding values ​​for the spare channel are THD′1 = 14.7% and THD′2 = 14.9%. Calculate the gradient. Switching time difference Δt = 120ms, activation status parameter When S a When the value is greater than 0.4, an activation command is generated. When the main channel distortion rate suddenly increases to 18.5%, the gradient increases to 9% / min. a =0.6 triggers emergency switchover.

[0083] S303: Divide the power factor range according to the activation state of the redundant circuit, establish the mapping curve between the power factor and the PWM frequency using cubic spline interpolation, and calculate the power factor of the corrugated roller motor by combining the rated speed of the motor and the current load rate.

[0084] Before dividing the power factor range, the current signal is decomposed into 5 levels using wavelet packet transform to obtain the energy E of each frequency band. b =[0.82kJ, 0.35kJ, 0.18kJ] corresponds to the 3rd, 5th, and 7th harmonics. Calculate the RMS value of the total harmonics. The fundamental effective value I1 = 28.7A, and the distortion rate is: When THD > 5%, the power factor correction factor η = 1 - 0.2 × (THD - 5). When THD = 4.11%, η = 1, and the actual power factor cosφ′ = η × cosφ = 0.85, falling into the high-efficiency range. The power factor f is calculated by interpolation. pwm =8.2kHz, duty cycle When the load rate increases to 90%, the interpolation value f pwm =8.9kHz, D=26.1%, ensuring smooth torque transition of the motor.

[0085] Please see Figure 1 This invention provides a technical solution: a comprehensive automated control method with redundancy and fault tolerance, comprising the following steps:

[0086] S401: Based on the power factor of the corrugated roller motor, a dual redundant control loop is constructed. When the phase lag is detected to continuously exceed 30° electrical angle, the cumulative phase offset is recorded to generate the phase stability of the main loop.

[0087] When a phase lag of θ = 30° was detected, based on the verification in the corrugated roll equipment (roll diameter φ320mm, 18CrMnTi material), the measured phase lag was 29.7° under the Siemens G120 frequency converter when the linear speed exceeded 180m / min (load rate 115%). After switching to the AB PowerFlex 755T backup circuit, the phase difference recovered to 4.8°. Specifically, when the dual-roll pressure increased to 2.8MPa, the 7# harmonic caused the phase jitter to exceed the limit (32.4°), triggering a switch. The duty cycle was increased from 68% to 82% in steps, and the speed difference between the rolls was maintained at 0.02m / s. Verified on a 2500mm wide B-type tile production line, the phase monitoring module drift was <1.2° at an ambient temperature of 45℃, and the annual fault switching frequency was <15 times / 10,000 hours. With a phase hysteresis set to 30°, the cumulative counter was activated, sampling the phase difference every 10ms and recording five consecutive cycles of data θ. t =

[0088] [32.1°, 32.5°, 33.2°, 33.8°, 34.0°], calculate the cumulative amount.

[0089] 2.5 + 3.2 + 3.8 + 4.0 = 15.6°, stability When θ suddenly increases to 38°, the cumulative Φ reaches 24.3, S p=0.514 is lower than the threshold of 0.6. During the acceleration phase of the cardboard production line, the phase lag increases from 31° to 37°, with a cumulative amount of 19.8, and the stability drops to 0.604, triggering a main circuit warning.

[0090] Table 4 Phase Lag Monitoring Data

[0091]

[0092] As shown in Table 4, when the stability of three consecutive sampling points is lower than 0.7, the system activates the redundancy detection mechanism.

[0093] S402: Call the main circuit phase stability to determine the circuit switching, establish a phase lag moving average sequence, and when the average of two consecutive sampling points exceeds the upper limit of the stable interval, trigger the switching of the backup redundant circuit signal, and generate the redundant circuit activation state according to the switching time difference and the phase recovery gradient.

[0094] Establish a moving average window W=3, and take the three most recent stability values ​​S. p =[0.68,0.65,0.61], calculate the mean. Set the upper limit of the stable range (1500 hours is the current running time), when μ = 0.647 < 0.6625, calculate the switching time difference. Gradient recovery Activated state When A s Switching is triggered when μ > 0.2. When μ = 0.63, Δt = 125ms, A s =0.288 to perform the switch.

[0095] S403: Divide the power factor range according to the activation state of the redundant loop. Set the power factor range of 0.8-1.0 as the high efficiency range, 0.6-0.8 as the normal range, and below 0.6 as the abnormal range. Use piecewise linear interpolation to generate PWM frequency mapping values ​​and calculate the PWM duty cycle parameters of the corrugated roller drive motor in combination with the rated speed of the motor.

[0096] When dividing the power factor range, when cosφ = 0.85 falls into the high-efficiency range, the interpolation nodes (0.8, 5kHz) and (1.0, 8kHz) are taken. In the corrugated roll drive control system, the implementation process of the piecewise linear interpolation method is as follows: When the current power factor is detected to be 0.92 (high-efficiency range), the interval endpoints (0.8, 50Hz) and (1.0, 60Hz) are taken to establish a linear function, and the PWM reference frequency is calculated as 50 + (0.92 - 0.8) / (1.0 - 0.8) × 10 = 56Hz; if the power factor drops to 0.7 (normal range), then (0.6, 30Hz) and (0.8, 50Hz) are taken for interpolation, and the frequency is calculated as 30 + (0.7 - 0.6) / (0.8 - 0.6) × 20 = 40Hz; when the power factor abnormally drops to 0.55, the minimum safe frequency of 25Hz is directly locked. Finally, the calculated frequency is correlated with the motor's rated speed of 1500 rpm, and dynamic speed regulation is achieved by using the duty cycle = (target frequency / rated frequency) × 100%, where the rated frequency corresponds to a reference duty cycle of 70% at 50 Hz. The calculation slope is... f pwm =5 + 15 × (0.85 - 0.8) = 5.75 kHz, combined with the rated speed n e =1450r / min, duty cycle When cosφ decreases to 0.72, using conventional interval interpolation at nodes (0.6, 3kHz) and (0.8, 5kHz), we obtain... PWM duty cycle parameters

[0097] Please see Figure 1 This invention provides a technical solution: a comprehensive automated control method with redundancy and fault tolerance, comprising the following steps:

[0098] S501: Call the PWM duty cycle parameters of the corrugated roll drive motor for parameter correction, use the stepwise regression method to eliminate the influence of multicollinearity, calculate the partial correlation coefficient between the duty cycle parameters and the motor speed, and generate the corrected duty cycle parameters;

[0099] Obtain the PWM duty cycle parameter D = [42.0%, 57.5%, 61.2%] and the corresponding speed n = [1420, 1450, 1480] r / min, and calculate the variance inflation factor. (when R) 2 When VIF = 0.45, after removing variables with VIF > 2.0, calculate the partial correlation coefficient. The corrected duty cycle D′=D×(1+0.15×r) p ) = 42.0% × 1.0945 = 45.97%, when rp When = 0.58, the correction amount ΔD = 5.2%.

[0100] Table 5. Data for Multicollinearity Test

[0101] Parameter combination Variance inflation factor Partial correlation coefficient Correction coefficient Dn 1.82 0.63 1.094 D-φ 2.15 0.55 1.083 n-φ 1.97 0.61 1.092

[0102] As shown in Table 5, when VIF>2.0, the parameter combination is excluded from the calibration model.

[0103] S502: Based on the corrected duty cycle parameters and the power factor of the corrugated roller motor, the critical period of the system is determined by the critical oscillation method. The rise time and overshoot are measured according to the constant amplitude oscillation waveform, and the braking current PID control curve is generated.

[0104] The critical oscillation test signal amplitude is set to A = 5%. The regression residual is converted into a period correction coefficient by the Sigmoid function, which drives the PID periodic dynamic adjustment. The duty cycle correction is embedded in the proportional element. The integral and derivative parameters are dynamically adjusted according to the reciprocal of the period and the rate of change, forming a complete closed-loop connection path for the control curve. The rise time t of the oscillation waveform is measured. r =0.12s, overshoot σ = 18.5%, calculate the critical period. PID parameters tuned to K i =2×K p / T c =15.3, K d =K p ×T c / 8 = 0.075, when T c At 0.25s, K is obtained. p =2.4, K i =19.2, K d =0.075.

[0105] S503: Call the backup circuit power factor to verify the braking current PID control curve, calculate the difference between the main circuit power factor and the curve prediction value, and when the difference is within the allowable range for three consecutive sampling cycles, trigger the channel switching command, load the updated parameters, and generate the corrugated roll machine parameter overload state.

[0106] Assuming the power factor of the standby circuit is cosφ′=0.88 and the power factor of the main circuit is cosφ=0.85, the predicted value of cosφ is... p =0.87, difference Δ = |0.85-0.87| = 0.02, set allowable range [-0.03, +0.03], the difference [0.02, 0.01, 0.02] for three consecutive periods all fall within the interval, calculate the switching confidence level. When C sReload is triggered when >0.3, and in the third cycle Δ3=0.015, C s =0.5 Execution parameter update.

[0107] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An integrated automation control method with redundant fault tolerance function, characterized in that, Includes the following steps: S1: The axial vibration acceleration waveform of the pressure roller on the cardboard processing equipment is obtained by a triaxial MEMS sensor. The peak-valley difference of the fundamental frequency band is extracted from the waveform. The axial vibration characteristic parameters of the corrugated roller are calculated based on the peak-valley difference and the angular displacement pulse signal output by the rotary encoder of the corrugated roller machine drive shaft. S2: The K-means clustering algorithm is used to classify the operating modes of the corrugated roll bearing housing and calculate the profile coefficient. Combined with the error of the cardboard pressure sensor, the roll surface deformation gradient compensation mechanism is activated. When the Pearson correlation coefficient between the profile coefficient and the deformation gradient change trend exceeds 0.8, an abnormal identifier of the corrugated roll gearbox is generated. The specific steps of S2 are as follows: S201: Call the K-means clustering algorithm with the input of the axial vibration feature parameters, calculate the similarity between the sample and the class center through Mahalanobis distance, determine the number of iterations using the elbow rule, update the cluster center until the change in intra-class variance is stable, calculate the average distance from each sample point to its class center, and generate the running modality classification result. S202: Calculate the profile coefficient based on the classification results of the operating modes, determine the sample density by using the maximum and minimum inter-class distances, and establish a linear regression equation between the error rate and the profile coefficient by combining the error rate output by the cardboard pressure sensor. When the regression coefficient t-test value shows statistical significance, generate a compensation activation signal. S203: Trigger the roll surface deformation gradient compensation mechanism according to the compensation activation signal, extract the deformation gradient change rate within the compensation period, calculate the Pearson correlation coefficient between the change rate and the profile coefficient, and generate an abnormality flag for the corrugated roll gearbox when the correlation coefficient exceeds 0.8 three times in a row. S3: Based on the abnormal identification of the corrugated roller gearbox, detect the dual-channel harmonic monitoring, calculate the dual-channel harmonic distortion rate of the pressure roller, and activate the backup channel when the distortion rate exceeds 15%. Calculate the power factor of the corrugated roller motor based on the harmonic distortion rate and the vibration characteristic parameters. S4: Based on the power factor of the corrugated roller motor, construct a dual redundant control loop for the paperboard linear speed. When the phase lag of the main loop is greater than 30°, switch to the backup loop. Divide the power factor into three intervals: 0.8-1.0, 0.6-0.8, and below 0.6, and generate the PWM duty cycle parameters of the corrugated roller drive motor. S5: The PWM duty cycle parameter of the corrugated roll drive motor is corrected by regression analysis. The braking current PID control curve is generated by combining the power factor of the corrugated roll motor. The effectiveness of the curve is verified by the power factor of the backup circuit. The redundant channel switching and parameter reload of the corrugated roll machine are then performed.

2. The integrated automation control method with redundant fault-tolerant function according to claim 1, characterized in that, The axial vibration characteristic parameters of the corrugated roll specifically include the peak-to-valley difference of the fundamental frequency band vibration, the integral value of the angular displacement pulse signal, and the harmonic components of the axial acceleration. The abnormality indicators of the corrugated roll gearbox specifically refer to the deformation gradient compensation activation state, the dual-channel harmonic distortion rate threshold, and the Pearson correlation coefficient verification mark. The power factor of the corrugated roll motor includes the main circuit phase lag angle, redundant channel switching, and harmonic distortion rate. The dual-redundant control loop of the cardboard linear speed specifically includes the phase lag compensation parameter, the duty cycle interval division standard, and the power factor segmented adjustment factor. The braking current PID control curve includes the proportional-integral-derivative coefficient combination, the power factor verification error rate, and the parameter overload trigger condition.

3. The integrated automation control method with redundant fault-tolerant function according to claim 1, characterized in that, The specific steps for acquiring the axial vibration acceleration waveform of the pressure roller on the cardboard processing equipment using a triaxial MEMS sensor, extracting the peak-to-valley difference of the fundamental frequency band from the waveform, and calculating the axial vibration characteristic parameters of the corrugated roller based on the peak-to-valley difference and the angular displacement pulse signal output by the rotary encoder of the corrugated roller machine drive shaft include: S101: The axial vibration signal of the pressure roller on the cardboard processing equipment is collected by a triaxial MEMS sensor. The signal is denoised by the sliding window method, the data in the window is filtered by median filtering, the acceleration waveform in the continuous processing cycle is extracted, and the axial vibration waveform is generated. S102: Based on the axial vibration waveform, perform a fast Fourier transform to identify the frequency component corresponding to the maximum amplitude, reconstruct the signal envelope using cubic spline interpolation, and calculate the fundamental frequency peak-valley difference between adjacent peaks and valleys. S103: Establish a time axis coordinate system by calling the angular displacement pulse signal output by the rotary encoder, and perform time-domain synchronous analysis on the fundamental frequency peak-valley difference and the corresponding angular displacement of the time axis to generate axial vibration characteristic parameters.

4. The integrated automation control method with redundant fault-tolerant function according to claim 3, characterized in that, The formula for identifying the frequency component corresponding to the maximum amplitude is: ； in, This represents the amplitude corresponding to the maximum amplitude component. This represents the absolute value of the difference between the k-th adjacent frequency component and the frequency of maximum amplitude. Let ω be the angular velocity of the vibration system. The time-domain interval corresponding to adjacent frequency components, This indicates the number of adjacent frequency points sampled. The smoothing factor is determined based on the signal-to-noise ratio. This represents the fundamental frequency value corresponding to the maximum amplitude. This represents the location of the symmetrical frequency point in the sampling.

5. The integrated automation control method with redundant fault-tolerant function according to claim 1, characterized in that, The linear regression equation for the error rate and profile coefficient is established using the following formula: ； in, This indicates the error rate output by the cardboard pressure sensor. The arithmetic mean of the profile coefficients. This represents the silhouette coefficient value of the i-th sample. This represents the maximum distance between classes. This represents the minimum distance between classes. For the number of valid samples, The adjustment factor is determined based on the t-test value and the sample size. It represents the reciprocal of the maximum inter-class distance.

6. The integrated automation control method with redundant fault-tolerant function according to claim 1, wherein, Based on the dual-channel harmonic monitoring of the corrugated roller gearbox anomaly detection, the dual-channel harmonic distortion rate of the pressure roller is calculated. When the distortion rate exceeds 15%, the backup redundant channel is activated. The specific steps for calculating the power factor of the corrugated roller motor based on the harmonic distortion rate and the vibration characteristic parameters include: S301: Based on the abnormal identification of the corrugated roller gearbox, dual-channel harmonic monitoring is triggered, the current harmonic components of the main and backup channels are collected, the fundamental periodic signal is obtained by phase-locked loop synchronous sampling technology, the effective values ​​of harmonics and the effective values ​​of the fundamental wave are extracted by wavelet packet transform and the root mean square value is calculated to generate dual-channel harmonic distortion rate; S302: Call the dual-channel harmonic distortion rate to evaluate the channel status. When the distortion rate of the main channel exceeds 15% for two consecutive sampling cycles and is higher than that of the backup channel, perform signal switching and generate the redundant channel activation state according to the channel switching time difference and the distortion rate change gradient. S303: Divide the power factor interval according to the activation state of the redundant channel, establish the mapping curve between power factor and PWM frequency using cubic spline interpolation, and calculate the power factor of the corrugated roller motor by combining the rated speed of the motor and the current load rate.

7. The integrated automation control method with redundant fault-tolerant function according to claim 1, wherein, Based on the power factor of the corrugated roll motor, a dual-redundant control loop for the cardboard linear speed is constructed. When the phase lag of the main loop is greater than 30°, it switches to the backup loop. The power factor is divided into three intervals: 0.8-1.0, 0.6-0.8, and below 0.

6. The specific steps for generating the PWM duty cycle parameters of the corrugated roll drive motor include: S401: Based on the power factor of the corrugated roller motor, a dual redundant control loop is constructed. When the phase lag is detected to continuously exceed 30° electrical angle, the cumulative phase offset is recorded to generate the phase stability of the main loop. S402: Call the main circuit phase stability to determine the circuit switching, establish a phase lag moving average sequence, and when the average of two consecutive sampling points exceeds the upper limit of the stable interval, trigger the switching of the backup redundant circuit signal, and generate the redundant channel activation state according to the switching time difference and the phase recovery gradient. S403: Divide the power factor range according to the activation state of the redundant channel, set the power factor range of 0.8-1.0 as the high efficiency range, 0.6-0.8 as the normal range, and below 0.6 as the abnormal range. Use piecewise linear interpolation to generate PWM frequency mapping values, and calculate the PWM duty cycle parameters of the corrugated roller drive motor in combination with the rated speed of the motor.

8. The integrated automation control method with redundant fault-tolerant function according to claim 1, wherein, The PWM duty cycle parameters of the corrugated roll drive motor are corrected using regression analysis. A PID control curve for the braking current is generated based on the power factor of the corrugated roll motor. The effectiveness of the curve is verified by the power factor of the backup circuit. The specific steps for switching redundant channels and reloading parameters of the corrugated roll machine include: S501: Call the PWM duty cycle parameters of the corrugated roll drive motor for parameter correction, use the stepwise regression method to eliminate the influence of multicollinearity, calculate the partial correlation coefficient between the duty cycle parameters and the motor speed, and generate the corrected duty cycle parameters; S502: Based on the corrected duty cycle parameters and the power factor of the corrugated roller motor, the critical period of the system is determined by the critical oscillation method, and the rise time and overshoot are measured according to the constant amplitude oscillation waveform to generate the braking current PID control curve. S503: Call the backup circuit power factor to verify the braking current PID control curve, calculate the difference between the main circuit power factor and the curve prediction value, and when the difference is within the allowable range for three consecutive sampling cycles, trigger the channel switching command, load the updated parameters, and generate the corrugated roller machine parameter overload state.

Images