A steel-based ceramic driving roller self-adaptive speed control method

Abstract

The application relates to the technical field of speed regulation control, and discloses a steel-based ceramic driving roller self-adaptive speed regulation control method. The method comprises the following steps: based on a zero-position pulse trigger signal, a driving motor current time domain signal is resampled according to a roller rotation angle to obtain a current circumferential distribution sequence, a phase alignment superposition average and a mean value removal process are performed to obtain a circumferential current fluctuation sequence, a circumferential discrete Fourier transform is performed on the circumferential current fluctuation sequence to extract each order harmonic amplitude and initial phase and reconstruct a friction degradation spatial atlas, a feedforward compensation amount is superimposed at a leading angle before a low friction zone reaches a belt surface contact arc to obtain an angle synchronous speed given value driving frequency converter output. The application solves the problem that the existing slip detection method cannot implement active compensation before a low friction zone reaches a belt surface contact arc due to the dependence of a time domain threshold trigger, and the speed regulation response is lagged.

Description

Technical Field

[0001] This application relates to the field of speed control technology, and in particular to an adaptive speed control method for a steel-based ceramic drive drum. Background Technology

[0002] Belt conveyors are essential equipment in industrial bulk material transportation. The drive drum, as a key component for power transmission, directly impacts transmission reliability due to the choice of its surface coating material. Steel-based ceramic drive drums use a steel cylinder as the base, with engineering ceramic patches such as alumina embedded in the outer surface. Compared to traditional rubber-coated drums, the ceramic coating offers higher surface hardness and a higher coefficient of friction, resulting in superior transmission stability under humid, dusty, and heavy-load conditions. Therefore, it is widely used in large belt conveyor systems in coal mines, ports, and metallurgy. Existing belt conveyor drive speed control methods typically employ variable frequency speed control strategies based on material flow rate, adjusting the inverter's output frequency by detecting the material flow rate, or protective speed control methods based on slip rate threshold triggering, reducing the drive speed to prevent slippage when the speed difference between the belt and the drum exceeds a preset threshold.

[0003] However, the aforementioned existing speed control methods all treat the friction coefficient as a fixed design parameter, using the static friction coefficient as the basis for preset speed control benchmarks. This ignores the localized detachment or wear of ceramic patches on steel-based ceramic drive rollers due to impact, fatigue, and thermal stress during long-term operation. This localized degradation of the ceramic surface causes the actual friction coefficient to continuously drift relative to the design value. Existing methods, based on the static friction coefficient, create a systematic deviation between the preset speed control benchmark and the actual friction transmission state, resulting in the speed control system operating under an incorrect control benchmark for extended periods. Furthermore, existing slippage detection methods rely on slip rate thresholds for triggering, representing a reactive, after-the-fact response mechanism. Adjustments are only made after slippage has occurred and the slip rate exceeds the threshold, failing to predict and proactively suppress slippage before it occurs, thus exhibiting an inherent lag in speed control response.

[0004] The localized detachment of ceramic patches physically occurs at specific angular positions along the circumference of the roller. This means that the degradation of the friction coefficient is not uniformly distributed across the entire circumference, but rather exhibits a spatially non-uniform distribution characteristic that strictly corresponds to the circumferential angle of the roller. This spatial non-uniformity along the circumference causes the low-friction area to periodically pass through the surface contact arc during each rotation, generating periodic micro-slip pulses with the same frequency characteristics as the rotation angle. However, existing technologies, whether based on time-domain slip rate detection methods or current amplitude monitoring methods, limit signal processing to the time axis, failing to separate and extract the circumferential spatial distribution structure of friction degradation from the current signal. Consequently, it is impossible to establish the spatial location of the degradation position or to implement advanced angular synchronous compensation for specific degradation positions. Existing speed control methods have a fundamental technical blind spot when facing this specific physical scenario of non-uniform circumferential degradation of the ceramic surface. Summary of the Invention

[0005] This application provides an adaptive speed control method for a steel-based ceramic driven roller, which solves the problems in the prior art of being unable to sense the spatial distribution of friction degradation in the circumferential direction on the surface of the steel-based ceramic driven roller and the long-term deviation of the speed control reference from the actual friction state. It also solves the problems in the existing slip detection methods, which rely on time-domain threshold triggering, resulting in a lag in speed response and the inability to implement active compensation before reaching the contact arc of the belt surface in the low friction zone.

[0006] This application provides an adaptive speed control method for a steel-based ceramic driven drum, the method comprising: Step S1: Based on the zero-position pulse trigger signal, resample the time-domain signal of the drive motor current according to the drum rotation angle to obtain the current circumferential distribution sequence; Step S2: Perform phase alignment, superposition, averaging, and mean removal processing on the multiple current circumferential distribution sequences to obtain a circumferential current fluctuation sequence; Step S3: Perform a circumferential discrete Fourier transform on the circumferential current fluctuation sequence to extract the amplitude and initial phase of each harmonic. The amplitude of each harmonic corresponds to the motor current deviation caused by insufficient friction due to local detachment or wear at different angular positions in the circumferential direction of the ceramic surface. The amplitude of each harmonic is converted into the friction coefficient deviation at the corresponding circumferential angular position based on the normal pressure, wrap angle and rated friction coefficient of the drive roller under rated working conditions. The friction coefficient deviation and the initial phase are reconstructed into a friction degradation space map according to the circumferential angle. Step S4: Combine the friction degradation spatial map with the real-time rotation angle, and superimpose the feedforward compensation at the leading angle before reaching the contact arc of the belt surface in the low friction zone to obtain the angle synchronization speed setpoint, and drive the frequency converter output with the angle synchronization speed setpoint.

[0007] The technical solution provided in this application introduces a zero-position pulse trigger signal at the shaft end of the steel-based ceramic drive drum as a physical reference for the starting angle of each revolution. The time-domain signal of the drive motor current is resampled into a current circumferential distribution sequence according to the drum rotation angle. This fundamentally establishes an explicit correspondence between the current signal and the drum circumferential angle, overcoming the inherent defect of existing technologies that limit current signal processing to the time axis and cannot perceive spatial information in the circumferential direction. Based on this, the multi-revolution current circumferential distribution sequence is phase-aligned, superimposed, averaged, and mean-removed to obtain a circumferential current fluctuation sequence with polarity markings. Utilizing the suppression effect of multi-revolution superposition on asynchronous noise caused by random load fluctuations, the friction degradation synchronous component corresponding to the fixed drum circumferential angle is separated from the noise background. This allows the weak circumferential features left by localized ceramic surface detachment or wear in the current signal to be clearly presented, solving the problem in existing technologies where the current signal at a single moment cannot effectively reflect the circumferential distribution of friction degradation due to insufficient signal-to-noise ratio.

[0008] Furthermore, this application inputs the circumferential current fluctuation sequence into a circumferential discrete Fourier transform to extract the amplitude and initial phase of each harmonic. Based on the Euler friction transmission equation, the harmonic amplitude is converted into the friction coefficient deviation, which, together with the initial phase, is reconstructed into a friction degradation spatial spectrum. This elevates the degradation state of the ceramic surface from a scalar time function to a multi-harmonic spectrum expression with a circumferential spatial distribution structure. This allows for the simultaneous realization of angular location of degradation and amplitude quantification of degradation degree, an information dimension that existing scalar monitoring methods based on slip rate or current amplitude cannot achieve at all. Based on the combination of the friction degradation spatial map and the real-time rotation angle, a feedforward compensation is superimposed at the leading angle before the low friction zone reaches the contact arc of the surface, to obtain the angle synchronous speed setpoint and drive the inverter output. This transforms the traditional passive speed regulation mechanism based on post-feedback into an active feedforward speed regulation mechanism based on the prediction of the circumferential position. This ensures that the drive torque reserve is established before the low friction zone enters the contact arc, fundamentally eliminating the periodic micro-slip pulses caused by the threshold trigger lag in the existing technology. This achieves adaptive speed regulation control of the steel-based ceramic drive roller under the condition of uneven circumferential degradation of the ceramic surface. Attached Figure Description

[0009] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a schematic diagram of an embodiment of the adaptive speed control method for steel-based ceramic driven drum in this application. Figure 2This is a schematic diagram of the phase alignment superposition and averaging of the multi-turn current circular distribution sequence and the polarity distribution of the circular current fluctuation sequence in the embodiments of this application; Figure 3 This is a schematic diagram comparing the given value of the angular synchronization velocity with the circumferential distribution of the rated angular velocity in an embodiment of this application. Detailed Implementation

[0011] This application provides an adaptive speed control method for a steel-based ceramic driven roller. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0012] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the adaptive speed control method for steel-based ceramic driven drum in this application includes: Step S1: Based on the zero-position pulse trigger signal, resample the time-domain signal of the drive motor current according to the drum rotation angle to obtain the current circumferential distribution sequence; Specifically, the zero-position pulse trigger signal refers to the single electrical pulse signal output by the single-turn zero-position pulse encoder installed at the end of the steel-based ceramic drive roller shaft when it completes one full rotation. This signal marks the start time of each rotation and is independent of the drive motor speed; it is only used to determine the time boundary of each rotation. The drive motor current time-domain signal refers to the effective value of the three-phase composite current calculated after the three-phase current of the three-phase AC motor is collected by the Hall sensor. The sampling frequency is set to 10,000 Hz. This value is based on the fact that at the rated speed of 120 rpm, a single rotation lasts 0.5 seconds, and a sampling frequency of 10,000 Hz can obtain 5,000 original sampling points within a single rotation. After resampling to 360 angle sampling points, each sampling point corresponds to approximately 14 original points. The interpolation accuracy meets the detection requirement that the width of the ceramic detachment zone is not less than 5 degrees. Resampling by drum rotation angle refers to mapping 5000 time-domain sampling points to 360 angular positions at equal angular intervals. The current value at each angular position is obtained by linear interpolation of adjacent time-domain sampling points. Finally, after each rotation, a current circumferential distribution sequence containing 360 data points is generated, and each data point in the sequence is bound to the corresponding circumferential angle.

[0013] Step S2: Perform phase alignment, superposition, averaging, and mean removal processing on the multiple current circumferential distribution sequences to obtain a circumferential current fluctuation sequence; Specifically, the premise of phase-aligned superposition averaging is that the circumferential distribution sequence of current in each revolution starts from the zero-position pulse trigger signal. Therefore, the angular reference frames of each revolution sequence are naturally aligned, and no additional phase correction operation is required. The superposition averaging is completed by directly taking the arithmetic mean of the current values ​​of multiple revolutions at the same angular position. The cumulative number of revolutions is set to 16 revolutions, based on the fact that 16 revolutions take about 8 seconds at rated speed. Ceramic surface degradation is a slow time-varying process, and the degradation distribution does not change substantially within 8 seconds. The non-synchronous component of random load fluctuations is attenuated to one-quarter of the root mean square value of a single revolution after averaging over 16 revolutions, which meets the signal-to-noise separation requirements. The mean-reduction process refers to calculating the average current value at 360 angular positions and then subtracting it point by point to obtain the circumferential current fluctuation sequence. Positive values ​​in the fluctuation sequence correspond to the intervals where friction is insufficient and the motor needs to increase the current, while negative values ​​correspond to the intervals where friction is sufficient. The polarity mark directly reflects the circumferential distribution direction of ceramic surface degradation.

[0014] Step S3: Perform a circumferential discrete Fourier transform on the circumferential current fluctuation sequence to extract the amplitude and initial phase of each harmonic. The amplitude of each harmonic corresponds to the motor current deviation caused by insufficient friction due to local detachment or wear at different angular positions in the circumferential direction of the ceramic surface. The amplitude of each harmonic is converted into the friction coefficient deviation at the corresponding circumferential angular position based on the normal pressure, wrap angle and rated friction coefficient of the drive roller under rated working conditions. The friction coefficient deviation and the initial phase are reconstructed into a friction degradation space map according to the circumferential angle. Specifically, the circumferential discrete Fourier transform transforms the 360-point circumferential current fluctuation sequence from the angular domain to the circumferential frequency domain, extracting the amplitudes and initial phases of the first to tenth harmonics. The tenth order is chosen because the number of simultaneous local detachment points on the ceramic roller surface under actual working conditions does not exceed ten, and the amplitudes of components above the tenth order can be ignored. The amplitudes of each harmonic are linearized using the Euler friction drive equation and converted into friction coefficient deviations. The surface normal pressure, wrap angle, and rated friction coefficient required for linearization are all design parameters of the drive roller and are known quantities. The friction degradation spatial map refers to the point-by-point superposition of the friction coefficient deviation amplitudes and corresponding initial phases into a cosine function, reconstructing a circumferential distribution curve of the friction coefficient deviation at 360 angular positions. The positive value intervals in the curve are marked as low-friction areas, representing the locations of local detachment or wear of the ceramic.

[0015] Step S4: Combine the friction degradation spatial map with the real-time rotation angle, and superimpose the feedforward compensation at the leading angle before reaching the contact arc of the belt surface in the low friction zone to obtain the angle synchronization speed setpoint, and drive the frequency converter output with the angle synchronization speed setpoint.

[0016] Specifically, the angle synchronization speed setpoint refers to the inverter speed setpoint signal obtained by superimposing the friction coefficient deviation at the corresponding position of the friction degradation space map at each sampling moment, based on the real-time rotation angle plus the lead angle, the calculation of the feedforward compensation amount, and the rated angular velocity. This signal is updated every 0.1 milliseconds and synchronized with the drum rotation angle in real time, driving the inverter to complete the torque boost before reaching the contact arc of the belt surface in the low friction zone.

[0017] In one specific embodiment, step S1 includes: Based on the zero-position pulse trigger signal output by the zero-position pulse encoder when the steel-based ceramic drive drum shaft completes one full rotation, the time interval between two adjacent zero-position pulse trigger signals is recorded to obtain the single rotation cycle. Based on the product of the single-cycle rotation period and the preset sampling frequency, the number of time-domain sampling points for this cycle is determined. The effective value of the three-phase composite current calculated after the three-phase current of the drive motor is collected by the Hall current sensor is truncated by the window of this cycle to obtain the time-domain current sequence of this cycle. The current sequence of this circle is resampled by linear interpolation at equal angles according to a preset number of sampling points at the circumferential angle. The current value of each sampling point at the angle is obtained by weighted summation of the current values ​​at the adjacent integer indices of its corresponding time-domain sampling index, thus obtaining the current circumferential distribution sequence of this circle. The current circumferential distribution sequence of multiple consecutive turns is accumulated and stored according to the turn number to obtain a multi-turn current circumferential distribution sequence.

[0018] Specifically, a zero-position pulse encoder is installed at the end of the steel-based ceramic drive roller shaft. It outputs one electrical pulse for each complete rotation. The controller records the time interval between two adjacent pulses as the single-rotation cycle. The preset sampling frequency is set to 10000 Hz. This value is based on the following: at a rated speed of 120 rpm, the single-rotation cycle is 0.5 seconds. At a sampling frequency of 10000 Hz, 5000 raw time-domain sampling points can be obtained within a single rotation. The number of time-domain sampling points per rotation is obtained by multiplying the single-rotation cycle by 10000 Hz and rounding to the nearest integer. This number of points adaptively adjusts with changes in rotational speed. The three-phase current of the drive motor is continuously acquired by a Hall current sensor at 10000 Hz. The composite effective value of the three-phase current is calculated at each sampling moment. The composite effective value sequence within the range of the current rotation is extracted using two adjacent zero-position pulse trigger signals as boundaries, resulting in the time-domain current sequence for that rotation. The number of points in this sequence is the number of time-domain sampling points for that rotation.

[0019] The preset number of circumferential angle sampling points is set to 360, corresponding to a circumferential resolution of 1 degree per point. When the current sequence of this circle is resampled to 360 equal angular positions, the time-domain sampling index corresponding to the m-th angle sampling point (m is an integer from 0 to 359) is m multiplied by the number of time-domain sampling points of this circle and then divided by 360. When the index is not an integer, the current values ​​at the lower and upper rounding indices are taken, and linear interpolation is performed with the decimal part as the weight. Specifically, the current value at the lower rounding index is multiplied by one and the decimal part is subtracted, and the current value at the upper rounding index is multiplied by the decimal part. The sum of the two is the current value of the m-th angle sampling point. Thus, the current circumferential distribution sequence of this circle is obtained. This sequence has a total of 360 data points, and each point is bound to a corresponding circumferential angle. The current circumferential distribution sequence obtained after each complete cycle is appended and stored in sequence according to the cycle number. After accumulating 16 cycles, a multi-cycle current circumferential distribution sequence is obtained. The 16 cycles are set based on the following: at the rated speed, 16 cycles take about 8 seconds. The degradation of the ceramic surface is a slow time-varying process. Within 8 seconds, the degradation distribution does not change substantially, which meets the signal-noise separation requirements of subsequent superposition and averaging.

[0020] In one specific embodiment, step S2 includes: Based on the starting angle corresponding to the zero-position pulse trigger signal of each loop in the multi-loop current circumferential distribution sequence, the multi-loop current circumferential distribution sequence is aligned by circumferential angle to obtain a phase-aligned current sequence group. The average current circumferential sequence is obtained by arithmetically averaging the current values ​​of each circle at the same circumferential angle position in the phase-aligned current sequence group. The mean value of the entire circle is calculated based on the current values ​​of all angle sampling points in the average current circumferential sequence. The difference between the current value of each angle sampling point in the average current circumferential sequence and the mean value of the entire circle is obtained to obtain the circumferential current fluctuation sequence. The deviation direction of each angle sampling point in the circumferential current fluctuation sequence is recorded as positive and negative marks. The angle interval with the current value higher than the average value of the whole circle is marked as a positive value interval, and the friction coefficient of the circumferential position in the corresponding angle interval of the ceramic surface is lower than the average level. The angle interval with the current value lower than the average value of the whole circle is marked as a negative value interval, and the friction coefficient of the circumferential position in the corresponding angle interval of the ceramic surface is higher than the average level, thus obtaining a circumferential current fluctuation sequence with polarity markings.

[0021] Specifically, since each round of the multi-turn current circumferential distribution sequence is triggered by a zero-position pulse trigger signal as the starting time for resampling, the 0th angle sampling point of each round of current circumferential distribution sequence naturally corresponds to the same physical position on the circumference of the roller. Therefore, the circumferential angle alignment operation does not require additional phase offset calculation. The 16 rounds of current circumferential distribution sequence are directly arranged according to the round number to form a phase-aligned current sequence group. This sequence group is a two-dimensional data array of 16 rows and 360 columns, with the row direction corresponding to the round number and the column direction corresponding to the circumferential angle position. The arithmetic mean of the phase-aligned current sequence group is taken column by column along the row direction. That is, the sum of 16 current values ​​in the same column (the same circumferential angle position) is divided by 16 to obtain the average current value at that angle position. After calculating the average current circumferential sequence for 360 angle positions in sequence, a total of 360 data points are obtained. Each point is the arithmetic mean of the current values ​​of 16 revolutions at the corresponding angle position. The root mean square value of the asynchronous current disturbance caused by random load fluctuations is reduced to one-quarter of that of a single revolution after averaging over 16 revolutions, while the synchronous current deviation caused by frictional degradation corresponding to the fixed circumferential angle of the roller is completely preserved.

[0022] The circumferential mean is obtained by summing the current values ​​of all 360 angular sampling points in the circumferential current average sequence and dividing by 360. This mean represents the reference level of the drive current under the average friction transmission state of the roller throughout the entire circumference. Subtracting the circumferential mean from the current value of each angular sampling point in the circumferential current average sequence yields the current deviation value at that angular position. Calculations are performed sequentially at all 360 angular positions to obtain the circumferential current fluctuation sequence. A positive current deviation value indicates that the drive motor needs to output a current higher than the average level at that angular position. The physical mechanism is as follows: when the ceramic surface experiences localized peeling or wear at that circumferential position, the friction coefficient decreases, and the effective tension of the belt surface decreases. To maintain stable speed, the motor automatically increases the output current to compensate for insufficient transmission. Therefore, the positive value range corresponds one-to-one with the low-friction degradation area of ​​the ceramic surface in the circumferential direction. A negative current deviation value indicates that the friction coefficient at that position is higher than the average level, corresponding to the intact area of ​​the ceramic surface. Record the positive or negative sign of each data point in the circumferential current fluctuation sequence as a polarity marker to obtain a circumferential current fluctuation sequence with polarity markers. This sequence also carries information on the current deviation amplitude and friction state direction at each angular position.

[0023] Figure 2 This is a schematic diagram of the phase alignment superposition and averaging of the multi-turn current circumferential distribution sequence and the polarity distribution of the circumferential current fluctuation sequence in the embodiments of this application. Figure 2 The upper part shows a comparison between the original single-turn current circumferential sequence and the 16-turn superimposed average sequence. The original single-turn sequence exhibits a large amount of random fluctuation due to random load fluctuations. In the 16-turn superimposed average sequence, the asynchronous noise component is effectively suppressed, and the friction degradation synchronous component corresponding to the fixed roller circumferential angle is clearly presented. Figure 2The lower half shows the circumferential current fluctuation sequence obtained after removing the mean. The dark-filled area represents the positive value range, corresponding to the low friction area where the friction coefficient within the angle range of the ceramic surface is lower than the average value of the whole circle. The light-filled area represents the negative value range, corresponding to the high friction area where the friction coefficient within the angle range of the ceramic surface is higher than the average value of the whole circle. The angle boundary and amplitude information of the two types of ranges together constitute the circumferential current fluctuation sequence with polarity marking.

[0024] In one specific embodiment, step S3 performs a circular discrete Fourier transform on the circular current fluctuation sequence to extract the amplitude and initial phase of each harmonic, including: The polarity-marked circular current fluctuation sequence is input into the circular discrete Fourier transform, and the current values ​​of all angle sampling points in the circular current fluctuation sequence are expanded into complex spectra in the order of circular angles to obtain the circular spectrum sequence. Based on the circumferential spectrum sequence, the magnitude of the complex spectral component corresponding to each harmonic order is extracted sequentially from the first to the tenth order. The magnitude is then multiplied by twice the reciprocal of the number of circumferential angle sampling points to obtain the harmonic amplitude of each order. Based on the circumferential spectrum sequence, the argument of each order complex spectral component is extracted to obtain the initial phase of each order. The initial phase of each order represents the peak position angle of the corresponding order friction degradation component in the circumferential direction. The harmonic amplitudes and initial phases of each order are stored in a one-to-one correspondence according to the harmonic order, resulting in a harmonic amplitude and initial phase lookup table.

[0025] Specifically, the circumferential current fluctuation sequence with polarity markings contains 360 real data points. Arranged in order of circumferential angle from 0 degrees to 359 degrees, these points are input into a circumferential discrete Fourier transform. The transform performs a complete discrete Fourier transform calculation on the 360 ​​data points, outputting 360 complex spectral components, forming a circumferential spectrum sequence. The nth complex component in the circumferential spectrum sequence corresponds to the nth harmonic in the circumferential direction. The physical meaning of the nth harmonic is the periodic change in current caused by n equally spaced degradation distributions along the circumferential direction of the ceramic surface. The first harmonic corresponds to a single local detachment, the second harmonic corresponds to two symmetrical detachments around the circumference, and so on. The harmonic order is chosen from the first to the tenth order because the number of simultaneous local detachment points on the surface of the steel-based ceramic roller under actual working conditions does not exceed ten, and the amplitudes of components above the tenth order are negligible relative to the noise background; therefore, the tenth order is sufficient to cover all actual degradation modes. The magnitude of the nth-order complex spectral component is taken and multiplied by twice the reciprocal of 360, which is equivalent to dividing the magnitude by 180, to obtain the amplitude of the nth-order harmonic. This normalization operation restores the accumulated result of the discrete Fourier transform to the physical quantity of the actual current deviation amplitude in the corresponding circular current fluctuation sequence, in amperes, consistent with the current unit of the input sequence.

[0026] The argument of each complex spectral component from the first to the tenth order in the circumferential spectrum sequence is calculated. The argument ranges from negative π to positive π, and the unit is radians. After converting the argument value to angle units, the peak position angle of the corresponding order friction degradation component in the circumferential direction is obtained, which is the initial phase of each order. The physical meaning of the nth order initial phase is: the circumferential angle position corresponding to the maximum amplitude of the nth order friction degradation component in a 360-degree circle. This angle position directly corresponds to the angular coordinate of the most severely degraded part of the ceramic surface in the circumferential direction. The harmonic amplitudes of each order from the first to the tenth order and the corresponding initial phases are arranged in ascending order. The harmonic order is used as the index key, and the harmonic amplitude and initial phase are stored as two columns in a reference structure to obtain a harmonic amplitude and initial phase reference table. This reference table has ten rows, and each row contains the harmonic amplitude (in amperes) and initial phase (in degrees) corresponding to a harmonic order, which completely records the frequency domain structure information of the circumferential degradation distribution of the ceramic surface.

[0027] In one specific embodiment, step S3 converts the amplitude values ​​of each harmonic order into friction coefficient deviations at corresponding circumferential angle positions based on the normal pressure, wrap angle, and rated friction coefficient of the drive roller under rated operating conditions. Based on the normal pressure, wrap angle and rated friction coefficient of the belt surface under the rated working condition of the drive roller, the change in effective tension of the belt surface corresponding to the change in unit friction coefficient under the rated working condition is calculated according to the Euler friction transmission equation, and the friction-tension linearization coefficient is obtained. Based on the friction-tension linearization coefficient, the rated torque coefficient of the drive motor, and the radius of the drive drum, the change in friction coefficient corresponding to the unit change in current is calculated to obtain the current-friction conversion coefficient. Multiply the harmonic amplitude of each order in the harmonic amplitude and initial phase comparison table with the current-friction conversion coefficient to obtain the deviation amplitude of each order friction coefficient; The friction coefficient deviation amplitude and the corresponding initial phase in the harmonic amplitude and initial phase comparison table are stored one-to-one according to the harmonic order to obtain the friction coefficient deviation amplitude and initial phase comparison table.

[0028] Specifically, the Euler friction drive equation describes the exponential relationship between the ratio of tight-side tension to slack-side tension on the drive roller's belt surface and the coefficient of friction and the wrap angle. The effective tension on the belt surface is the difference between the tight-side tension and the slack-side tension. Taking the first partial derivative of the effective tension with respect to the coefficient of friction at the rated coefficient of friction yields the change in effective tension corresponding to a unit change in the coefficient of friction under rated operating conditions, i.e., the friction-tension linearization coefficient. The unit of this coefficient is Newtons per dimensionless unit. The required input parameters are: the unit of the belt surface normal force is Newtons; the unit of the wrap angle is radians; and the rated coefficient of friction is dimensionless. All three are drive roller design parameters and are known quantities, requiring no online measurement. The physical premise for linearization is that the deviation in the coefficient of friction caused by local degradation of the ceramic surface is small relative to the rated coefficient of friction, and the first-order linearization error at the rated coefficient of friction is within the engineering allowable range. This premise holds true when the ceramic detachment area does not exceed 20% of the roller's contact area.

[0029] Based on the friction-tension linearization coefficient, combined with the rated torque coefficient of the drive motor (in Newton-meters per ampere) and the radius of the drive roller (in meters), the process of calculating the current-friction conversion coefficient is as follows: Multiply the effective tension change of the belt surface by the radius of the drive roller to obtain the output torque change of the drive motor. Divide the output torque change of the drive motor by the rated torque coefficient to obtain the corresponding current change. Divide the friction-tension linearization coefficient by the ratio of the current change to the unit friction coefficient change to obtain the friction coefficient change corresponding to the unit current change, i.e., the current-friction conversion coefficient, in dimensionless units per ampere. Multiply the harmonic amplitudes (in amperes) of the first to tenth orders in the harmonic amplitude and initial phase comparison table by the current-friction conversion coefficient (in dimensionless units per ampere) to obtain the friction coefficient deviation amplitude (in dimensionless units) for each order. The physical meaning of this product is: the magnitude of the friction coefficient deviation caused by the circumferential direction of the corresponding order ceramic surface degradation. The initial phases of each order corresponding to the friction coefficient deviation amplitude and harmonic amplitude and initial phase comparison table are arranged and stored in order from the first to the tenth harmonic order to form the friction coefficient deviation amplitude and initial phase comparison table. The comparison table has ten rows, and each row contains the friction coefficient deviation amplitude and initial phase corresponding to a harmonic order. The initial phase uses the original value in the harmonic amplitude and initial phase comparison table, and the unit is degrees.

[0030] In one specific embodiment, step S3 reconstructs the friction coefficient deviation and the initial phase into a friction degradation space map by circumferential angle, including: Based on the aforementioned table of friction coefficient deviation amplitude and initial phase, the friction coefficient deviation amplitude and corresponding initial phase of each order are substituted into the cosine function of each order to obtain the friction degradation circumferential components of each order. The amplitudes of the friction degradation circumferential components of each order at each circumferential angle sampling point are superimposed point by point to obtain the full circumferential friction coefficient deviation distribution sequence. Based on the angle range with positive amplitude in the full circumferential friction coefficient deviation distribution sequence, it is marked as a low friction zone. The low friction zone corresponds to the location of local detachment or wear on the ceramic surface. Based on the angle range with negative amplitude in the full circumferential friction coefficient deviation distribution sequence, it is marked as a high friction zone. The angle range and corresponding amplitude of the low friction zone and the high friction zone together constitute a friction degradation spatial map.

[0031] Specifically, for the data of the first to tenth orders in the table of friction coefficient deviation amplitude and initial phase comparison, the calculation method of the nth order friction degradation circumferential component is as follows: taking the circumferential angle position (range from 0 degrees to 359 degrees, step size of 1 degree, a total of 360 angle points) as the independent variable, the nth order friction coefficient deviation amplitude is used as the amplitude coefficient of the cosine function, n is multiplied by the current circumferential angle position and converted to radians, and then the value of the nth order initial phase converted to radians is subtracted as the phase independent variable of the cosine function. After calculating the cosine function value, it is multiplied by the amplitude coefficient to obtain the amplitude of the nth order friction degradation circumferential component at that angle position. After calculating for the 360 ​​angle positions in sequence, the nth order friction degradation circumferential component sequence is obtained. This sequence has a total of 360 data points, the unit is dimensionless, and the physical meaning is the contribution of the nth order degradation mode to the friction coefficient deviation at each circumferential angle position. For the ten circumferential component sequences of friction degradation from the first to the tenth order, the amplitudes of the tenth order components are added at each circumferential angle sampling point to obtain the total deviation value of the friction coefficient at that angle position. After calculating the deviation values ​​at 360 angle positions in sequence, a full circumferential friction coefficient deviation distribution sequence is obtained. This sequence has a total of 360 data points, and the value of each point is the deviation of the ceramic surface friction coefficient at the corresponding circumferential angle position from the rated friction coefficient, with the unit being dimensionless.

[0032] The symbols of the full-circumferential friction coefficient deviation distribution sequence are checked point by point. For continuous angular intervals with amplitudes greater than zero, the starting angle, ending angle, and amplitude values ​​at each point within the interval are recorded and marked as low-friction zones. The physical meaning of a low-friction zone is that there is localized detachment or wear on the ceramic surface within this circumferential angle range, resulting in a friction coefficient lower than the rated value. This corresponds to the physical relationship of the positive value intervals in the circumferential current fluctuation sequence with polarity markings. The two maintain logical consistency through the transmission link of harmonic decomposition and reconstruction. For continuous angular intervals with amplitudes less than zero, the starting angle, ending angle, and amplitude values ​​at each point within the interval are recorded and marked as high-friction zones. High-friction zones correspond to circumferential positions where the ceramic surface is intact and the friction coefficient is higher than the rated value. The friction degradation spatial map is composed of the angular ranges and corresponding amplitudes of all low-friction and high-friction zones. Its data structure is a 360-point sequence indexed by circumferential angles and numerically represented by friction coefficient deviations, with interval boundary markings for low-friction and high-friction zones. This map fully describes the spatial distribution of friction degradation on the surface of the steel-based ceramic drive roller in the circumferential direction.

[0033] In one specific embodiment, step S4 includes: Based on the ratio of the cumulative pulse count output by the zero-position pulse encoder in the current circle to the number of time-domain sampling points in the previous complete circle, the ratio is multiplied by 360° to obtain the real-time rotation angle. Based on the angle range of the low friction zone in the friction degradation space map, the real-time rotation angle is added to a leading angle determined by half of the contact arc angle to obtain the feedforward query angle. At the feedforward query angle, the corresponding friction coefficient deviation is extracted from the friction degradation space map. Based on the ratio of the product of the friction coefficient deviation and the rated angular velocity and the inverter speed-torque droop coefficient to the rated friction coefficient, the feedforward compensation amount is obtained. The angle interval in the feedforward compensation amount where the circumferential gradient exceeds the preset gradient threshold is identified as the gradient abrupt interval. The width of the smoothing window is determined by weighted summation of the harmonic amplitudes of each order in the friction degradation spatial spectrum. The feedforward compensation amount in the gradient abrupt interval is smoothed by Hanning window weighted moving average to obtain the smoothed feedforward compensation amount. The smooth feedforward compensation is superimposed on the rated angular velocity to obtain the angular synchronization speed setpoint, and the inverter output is driven by the angular synchronization speed setpoint.

[0034] Specifically, the cumulative pulse count refers to the accumulated sampling count value recorded by the zero-position pulse encoder from the moment the zero-position pulse trigger signal is output to the current sampling moment. This count value is cleared and restarted from zero each time a zero-position pulse is triggered. Its value range is from zero to the number of time-domain sampling points of the previous complete revolution minus one. Dividing the cumulative pulse count by the number of time-domain sampling points of the previous complete revolution gives the proportion of rotation completed in the current revolution. Multiplying this by 360 degrees gives the real-time rotation angle. This calculation is performed once every sampling cycle, i.e., every 0.1 milliseconds. The angular resolution of the real-time rotation angle is 360 degrees divided by the number of time-domain sampling points of the previous complete revolution, which is 0.072 degrees at the rated speed of 120 rpm. The lead angle is taken as half of the wrap angle of the contact arc of the belt surface. The physical basis is that the circumferential rotation angle required for the low-friction zone to travel from the beginning of the contact arc of the belt surface to the middle of the contact arc is about half of the wrap angle. Within the time corresponding to this rotation angle, the frequency converter completes the torque boost response, so that the driving torque reserve is fully established when the low-friction zone reaches the middle of the contact arc. The real-time rotation angle is added to the lead angle, and the modulus is taken by 360 degrees to obtain the feedforward query angle. Using the feedforward query angle as an index, the friction coefficient deviation at the corresponding position is extracted from the 360-point sequence of the friction degradation spatial map. The feedforward compensation is calculated as follows: the friction coefficient deviation is multiplied by the rated angular velocity, then multiplied by the inverter speed-torque droop coefficient, and the result is divided by the rated friction coefficient to obtain the feedforward compensation, in radians per second. The inverter speed-torque droop coefficient is a dimensionless parameter, ranging from 0.03 to 0.08, reflecting the torque adjustment ratio corresponding to a unit speed deviation. The rated friction coefficient is a known quantity and a design parameter of the drive roller.

[0035] The feedforward compensation at each of the 360 ​​angular positions along the entire circumference is calculated point-by-point, with the difference between adjacent angle points to obtain the circumferential gradient sequence. The preset gradient threshold is set to 0.003 radians per second per degree. This value is based on the following: the longitudinal propagation speed of the belt is approximately 80 meters per second, the diameter of the drive roller is 0.8 meters, and under the constraint that the belt tension fluctuation does not exceed 2% of the rated tension, the upper limit of the allowable rate of change of the speed setpoint within a 1-degree angle variation range is calculated to be 0.003 radians per second per degree. The angle points in the circumferential gradient sequence whose absolute values ​​exceed the preset gradient threshold, and the half-angle unit of the width of the extended smoothing window on both sides, constitute the gradient abrupt change interval. The smoothing window width is determined by the ratio of 360 degrees to the weighted sum of the harmonic amplitudes of each order in the friction degradation space spectrum. Specifically, the smoothing window width is obtained by multiplying each harmonic amplitude by its corresponding order, summing the results, and then dividing the sum by 360. The minimum window width is limited to 5 degrees, and the maximum to 30 degrees. This setting is based on the principle that a narrow degradation region and a small required smoothing range are needed when higher-order components dominate, while a wide degradation region and a large required smoothing range are needed when lower-order components dominate. These two boundary values ​​cover the width range of the degradation region in the actual working conditions of the steel-based ceramic roller. The Hanning window weighted moving average is only performed within the gradient abrupt change interval. For each angle point within the interval, a sequence of feedforward compensation values ​​within the smoothing window width is taken, centered on that point. This sequence is then weighted by the Hanning window weight coefficients, summed, and normalized to obtain the smoothed feedforward compensation value for that point. Angle points outside the gradient abrupt change interval retain their original feedforward compensation values. The smooth feedforward compensation at each of the 360 ​​angular positions of the full circle is added point by point to the rated angular velocity to obtain a sequence of angular synchronous speed setpoints. The angular synchronous speed setpoint corresponding to the current moment is extracted from this sequence using the real-time rotation angle as an index. After digital-to-analog conversion, it is continuously output to the external speed setpoint port of the frequency converter with a period of 0.1 milliseconds to drive the frequency converter output.

[0036] Figure 3 This is a schematic diagram comparing the given value of the angular synchronization velocity with the circumferential distribution of the rated angular velocity in an embodiment of this application. Figure 3 The dashed line represents the rated angular velocity reference value, and the solid line represents the angular synchronous speed setpoint corresponding to each circumferential angle position under the drive of the friction degradation space map. The dark area filled by the diagonal line represents the positive value range of the feedforward compensation, corresponding to the speed setpoint being higher than the rated angular velocity when the low friction zone reaches the lead angle of the contact arc, driving the frequency converter to build up torque reserve in advance. The light area filled by the diagonal line represents the negative value range of the feedforward compensation, corresponding to the speed setpoint being lower than the rated angular velocity when the high friction zone passes through the lead angle, achieving torque balance in the circumferential direction. The angular synchronous speed setpoint is continuously updated with the real-time rotation angle at a period of 0.1 milliseconds and output to the external speed setpoint port of the frequency converter.

[0037] In one specific embodiment, the leading angle is half of the wrap angle of the belt contact arc, which is determined by the actual wrap angle between the drive roller and the belt, and the value ranges from 180° to 240°.

[0038] The width of the smoothing window is determined by taking the reciprocal of the sum of the harmonic amplitudes of each order in the triboelectric space spectrum according to their order, and then multiplying it by 360°. The minimum value of the smoothing window width is 5° and the maximum value is 30°.

[0039] The speed-torque droop coefficient of the frequency converter ranges from 0.03 to 0.08. When the superposition result of the feedforward compensation and the rated angular velocity exceeds ±10% of the rated angular velocity, the superposition result is limited to within ±10°% of the rated angular velocity and then output as the angular synchronization speed setpoint.

[0040] Specifically, the belt contact arc wrap angle refers to the central angle corresponding to the arc of actual contact between the drive drum and the conveyor belt. This angle is determined by the diameter of the drive drum, the belt tension, and the conveyor layout. In a single-drum drive layout, the typical range is 180 to 240 degrees. The lead angle is half of the belt contact arc wrap angle, i.e., 90 to 120 degrees. Within this range, the dynamic response time of the frequency converter corresponds to the time required for the drum to rotate 90 to 120 degrees at the rated speed. Calculated at a rated speed of 120 rpm, it takes approximately 125 milliseconds to rotate 90 degrees and approximately 167 milliseconds to rotate 120 degrees. The above time range covers the torque build-up response time of mainstream frequency converters under load conditions, ensuring that feedforward compensation is completed before reaching the middle of the belt contact arc in the low-friction zone. In calculating the width of the smoothing window, the weighted summation of the harmonic amplitudes by order means multiplying the amplitude of the first harmonic by 1, the amplitude of the second harmonic by 2, and so on up to the amplitude of the tenth harmonic by 10. The ten products are then added together to obtain a weighted sum. The reciprocal of the weighted sum is then multiplied by 360 degrees to obtain the width of the smoothing window. When the amplitude of higher-order harmonic components is larger, the weighted sum is larger and the reciprocal is smaller, resulting in a narrower smoothing window width, which corresponds to a narrower distribution of the ceramic degradation area in the circumferential direction. When lower-order harmonic components dominate, the weighted sum is smaller and the reciprocal is larger, resulting in a wider smoothing window width, which corresponds to a large-area degradation situation. The minimum value of 5 degrees and the maximum value of 30 degrees are determined by the physical range of the circumferential width of the degradation area in the actual working conditions of the steel-based ceramic drum.

[0041] The reason for setting the speed-torque droop coefficient of the frequency converter to a range of 0.03 to 0.08 is as follows: if the droop coefficient is too small, the feedforward compensation is insufficient to drive the frequency converter to generate an effective torque boost; if the droop coefficient is too large, the feedforward compensation exceeds the dynamic adjustment range of the frequency converter, leading to over-adjustment. 0.03 to 0.08 covers the effective adjustment range of commonly used industrial frequency converters under belt conveyor conditions. The reason for setting the amplitude limit threshold to ±10% of the rated angular velocity is as follows: the allowable speed fluctuation range of the conveyor belt is constrained by the longitudinal elasticity of the belt and the stroke of the tensioning device. When the speed deviation exceeds 10% of the rated value, the belt tension fluctuation exceeds the compensation capacity of the tensioning device, posing a risk of belt misalignment. Therefore, the angular synchronization speed setpoint is limited to ±10% of the rated angular velocity. When it exceeds this range, it is truncated to the boundary value before being output to the frequency converter to prevent damage to the conveyor belt caused by excessive speed setpoint due to abnormal amplitude in the friction degradation space spectrum.

[0042] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for adaptive speed control of a steel-based ceramic driven drum, characterized in that, The method includes: Step S1: Based on the zero-position pulse trigger signal, resample the time-domain signal of the drive motor current according to the drum rotation angle to obtain the current circumferential distribution sequence; Step S2: Perform phase alignment, superposition, averaging, and mean removal processing on the multiple current circumferential distribution sequences to obtain a circumferential current fluctuation sequence; Step S3: Perform a circumferential discrete Fourier transform on the circumferential current fluctuation sequence to extract the amplitude and initial phase of each harmonic. The amplitude of each harmonic corresponds to the motor current deviation caused by insufficient friction due to local detachment or wear at different angular positions in the circumferential direction of the ceramic surface. The amplitude of each harmonic is converted into the friction coefficient deviation at the corresponding circumferential angular position based on the normal pressure, wrap angle and rated friction coefficient of the drive roller under rated working conditions. The friction coefficient deviation and the initial phase are reconstructed into a friction degradation space map according to the circumferential angle. Step S4: Combine the friction degradation spatial map with the real-time rotation angle, and superimpose the feedforward compensation at the leading angle before reaching the contact arc of the belt surface in the low friction zone to obtain the angle synchronization speed setpoint, and drive the frequency converter output with the angle synchronization speed setpoint.

2. The adaptive speed control method for steel-based ceramic driven drum according to claim 1, characterized in that, Step S1 includes: Based on the zero-position pulse trigger signal output by the zero-position pulse encoder when the steel-based ceramic drive drum shaft completes one full rotation, the time interval between two adjacent zero-position pulse trigger signals is recorded to obtain the single rotation cycle. Based on the product of the single-cycle rotation period and the preset sampling frequency, the number of time-domain sampling points for this cycle is determined. The effective value of the three-phase composite current calculated after the three-phase current of the drive motor is collected by the Hall current sensor is truncated by the window of this cycle to obtain the time-domain current sequence of this cycle. The current sequence of this circle is resampled by linear interpolation at equal angles according to a preset number of sampling points at the circumferential angle. The current value of each sampling point at the angle is obtained by weighted summation of the current values ​​at the adjacent integer indices of its corresponding time-domain sampling index, thus obtaining the current circumferential distribution sequence of this circle. The current circumferential distribution sequence of multiple consecutive turns is accumulated and stored according to the turn number to obtain a multi-turn current circumferential distribution sequence.

3. The adaptive speed control method for steel-based ceramic driven drum according to claim 2, characterized in that, Step S2 includes: Based on the starting angle corresponding to the zero-position pulse trigger signal of each loop in the multi-loop current circumferential distribution sequence, the multi-loop current circumferential distribution sequence is aligned by circumferential angle to obtain a phase-aligned current sequence group. The average current circumferential sequence is obtained by arithmetically averaging the current values ​​of each circle at the same circumferential angle position in the phase-aligned current sequence group. The mean value of the entire circle is calculated based on the current values ​​of all angle sampling points in the average current circumferential sequence. The difference between the current value of each angle sampling point in the average current circumferential sequence and the mean value of the entire circle is obtained to obtain the circumferential current fluctuation sequence. The deviation direction of each angle sampling point in the circumferential current fluctuation sequence is recorded as positive and negative marks. The angle interval with the current value higher than the average value of the whole circle is marked as a positive value interval, and the friction coefficient of the circumferential position in the corresponding angle interval of the ceramic surface is lower than the average level. The angle interval with the current value lower than the average value of the whole circle is marked as a negative value interval, and the friction coefficient of the circumferential position in the corresponding angle interval of the ceramic surface is higher than the average level, thus obtaining a circumferential current fluctuation sequence with polarity markings.

4. The adaptive speed control method for steel-based ceramic driven drum according to claim 3, characterized in that, Step S3 involves performing a circular discrete Fourier transform on the circular current fluctuation sequence to extract the amplitude and initial phase of each harmonic, including: The polarity-marked circular current fluctuation sequence is input into the circular discrete Fourier transform, and the current values ​​of all angle sampling points in the circular current fluctuation sequence are expanded into complex spectra in the order of circular angles to obtain the circular spectrum sequence. Based on the circumferential spectrum sequence, the magnitude of the complex spectral component corresponding to each harmonic order is extracted sequentially from the first to the tenth order. The magnitude is then multiplied by twice the reciprocal of the number of circumferential angle sampling points to obtain the harmonic amplitude of each order. Based on the circumferential spectrum sequence, the argument of each order complex spectral component is extracted to obtain the initial phase of each order. The initial phase of each order represents the peak position angle of the corresponding order friction degradation component in the circumferential direction. The harmonic amplitudes and initial phases of each order are stored in a one-to-one correspondence according to the harmonic order, resulting in a harmonic amplitude and initial phase lookup table.

5. The adaptive speed control method for steel-based ceramic driven drum according to claim 4, characterized in that, Step S3 converts the amplitude values ​​of each harmonic order into friction coefficient deviations at corresponding circumferential angle positions based on the normal pressure, wrap angle, and rated friction coefficient of the drive roller under rated operating conditions. Based on the normal pressure, wrap angle and rated friction coefficient of the belt surface under the rated working condition of the drive roller, the change in effective tension of the belt surface corresponding to the change in unit friction coefficient under the rated working condition is calculated according to the Euler friction transmission equation, and the friction-tension linearization coefficient is obtained. Based on the friction-tension linearization coefficient, the rated torque coefficient of the drive motor, and the radius of the drive drum, the change in friction coefficient corresponding to the unit change in current is calculated to obtain the current-friction conversion coefficient. Multiply the harmonic amplitude of each order in the harmonic amplitude and initial phase comparison table with the current-friction conversion coefficient to obtain the deviation amplitude of each order friction coefficient; The friction coefficient deviation amplitude and the corresponding initial phase in the harmonic amplitude and initial phase comparison table are stored one-to-one according to the harmonic order to obtain the friction coefficient deviation amplitude and initial phase comparison table.

6. The adaptive speed control method for steel-based ceramic driven drum according to claim 5, characterized in that, Step S3 reconstructs the friction coefficient deviation and the initial phase into a friction degradation space map according to the circumferential angle, including: Based on the aforementioned table of friction coefficient deviation amplitude and initial phase, the friction coefficient deviation amplitude and corresponding initial phase of each order are substituted into the cosine function of each order to obtain the friction degradation circumferential components of each order. The amplitudes of the friction degradation circumferential components of each order at each circumferential angle sampling point are superimposed point by point to obtain the full circumferential friction coefficient deviation distribution sequence. Based on the angle range with positive amplitude in the full circumferential friction coefficient deviation distribution sequence, it is marked as a low friction zone. The low friction zone corresponds to the location of local detachment or wear on the ceramic surface. Based on the angle range with negative amplitude in the full circumferential friction coefficient deviation distribution sequence, it is marked as a high friction zone. The angle range and corresponding amplitude of the low friction zone and the high friction zone together constitute a friction degradation spatial map.

7. The adaptive speed control method for steel-based ceramic driven drum according to claim 1, characterized in that, Step S4 includes: Based on the ratio of the cumulative pulse count output by the zero-position pulse encoder in the current circle to the number of time-domain sampling points in the previous complete circle, the ratio is multiplied by 360° to obtain the real-time rotation angle. Based on the angle range of the low friction zone in the friction degradation space map, the real-time rotation angle is added to a leading angle determined by half of the contact arc angle to obtain the feedforward query angle. At the feedforward query angle, the corresponding friction coefficient deviation is extracted from the friction degradation space map. Based on the ratio of the product of the friction coefficient deviation and the rated angular velocity and the inverter speed-torque droop coefficient to the rated friction coefficient, the feedforward compensation amount is obtained. The angle interval in the feedforward compensation amount where the circumferential gradient exceeds the preset gradient threshold is identified as the gradient abrupt interval. The width of the smoothing window is determined by weighted summation of the harmonic amplitudes of each order in the friction degradation spatial spectrum. The feedforward compensation amount in the gradient abrupt interval is smoothed by Hanning window weighted moving average to obtain the smoothed feedforward compensation amount. The smooth feedforward compensation is superimposed on the rated angular velocity to obtain the angular synchronization speed setpoint, and the inverter output is driven by the angular synchronization speed setpoint.

8. The adaptive speed control method for steel-based ceramic driven drum according to claim 7, characterized in that, The leading angle is half of the surface contact arc wrap angle, which is determined by the actual wrap angle between the drive roller and the belt, and ranges from 180° to 240°.

9. The adaptive speed control method for steel-based ceramic driven drum according to claim 7, characterized in that, The width of the smoothing window is determined by taking the reciprocal of the sum of the harmonic amplitudes of each order in the triboelectric space spectrum according to their order, and then multiplying it by 360°. The minimum value of the smoothing window width is 5° and the maximum value is 30°.

10. The adaptive speed control method for steel-based ceramic driven drum according to claim 7, characterized in that, The speed-torque droop coefficient of the frequency converter ranges from 0.03 to 0.

08. When the superposition result of the smooth feedforward compensation and the rated angular velocity exceeds ±10% of the rated angular velocity, the superposition result is limited to ±10% of the rated angular velocity and then output as the angular synchronization speed setpoint.

Images