A servo synchronous tracking control method and system for cold-cutting flying saws

Abstract

This invention relates to the field of flying saw servo control, specifically to a servo synchronous tracking control method and system for cold-cutting flying saws. First, a synchronous acquisition link is used to perform time-series alignment and detrending processing on current and speed signals. Second, based on an inverse dynamics model, the original torque containing full-frequency information is calculated. Then, fast and slow dual-channel filters are constructed to extract high-frequency impact and steady-state load characteristics, respectively. Finally, transient energy analysis is used to calculate dynamic confidence weights, and the dual-channel signals are adaptively weighted and fused, with a current loop feedforward injected. This scheme achieves rapid compensation for millisecond-level cutting impacts while shielding mechanical noise. This invention employs fast and slow dual-channel filtering based on an inverse dynamics model, combined with transient energy difference calculation of dynamic confidence weights, to achieve adaptive fusion compensation of disturbance signals.

Description

Technical Field

[0001] This invention relates to the field of flying saw servo control, and specifically to a servo synchronous tracking control method and system for cold-cutting flying saws. Background Technology

[0002] The cold-cutting flying saw servo system is a key piece of equipment in modern metallurgical pipe production lines. It utilizes a high-speed rotating metal circular saw blade to synchronously track and mill continuously moving pipes to a fixed length. During this process, to overcome the severe load disturbances generated at the moment of sawing contact, existing core control technologies typically incorporate a Disturbance Observer (DOB) based on an inverse dynamics model. This algorithm primarily establishes an inverse dynamics model of the system, combines motor speed and current feedback, calculates an estimate of the load torque, and performs feedforward compensation to maintain the system's dynamic stability.

[0003] However, in real-world high-speed dynamic tracking scenarios, conventional DOB (Digital Obligation) suffers from a significant contradiction between bandwidth and noise suppression. On one hand, the cutting impact generated when the saw blade cuts into the pipe has an extremely short rise time, constituting an effective high-frequency signal. This requires the Q-filter in the observer to be configured with a high cutoff frequency to quickly capture millisecond-level impacts. On the other hand, the inherent high-frequency interference in the mechanical transmission chain is also located in the high-frequency band. If high bandwidth is simply pursued, the aforementioned invalid noise will be amplified indiscriminately by the observer and fed back to the current loop, easily leading to servo system whistling, motor overheating, or even divergent oscillations. Conversely, if the filter bandwidth is reduced to suppress noise, it will cause a significant phase lag in the observer's observation of the actual cutting impact, making it impossible for the control system to respond to load changes in a timely manner, ultimately failing to effectively solve the technical problem of motor speed drop during sawing. Summary of the Invention

[0004] To address the issues of high bandwidth causing servo system howling and reduced filtering bandwidth leading to observer phase lag, this invention proposes a synchronous tracking control method for a cold-cutting flying saw servo in a first aspect. The method includes: synchronously acquiring the angular velocity signal and torque-current setpoint signal of the motor in each control cycle of the servo driver; processing the angular velocity signal and torque-current setpoint signal based on preset motor nominal parameters and according to the inverse operation logic of Newton's second law to obtain an original disturbance estimate containing full-bandwidth information; performing high-bandwidth filtering and low-bandwidth filtering on the original disturbance estimate in parallel to obtain high-dynamic disturbance observations and steady-state disturbance observations, respectively; calculating the amplitude deviation between the high-dynamic disturbance observations and the steady-state disturbance observations, comparing the amplitude deviation with a preset noise floor threshold to obtain a dynamic confidence weight characterizing the reliability of the high-dynamic disturbance observations; performing weighted fusion of the high-dynamic disturbance observations and the steady-state disturbance observations under the dynamic confidence weight to obtain a final disturbance compensation command, which is then superimposed onto the current loop feedforward input of the servo driver.

[0005] This invention resolves the contradiction in the bandwidth setting of disturbance observers in traditional servo control: increasing bandwidth improves dynamic response but introduces high-frequency noise, while decreasing bandwidth suppresses noise but causes phase lag. By performing high- and low-bandwidth filtering in parallel and calculating dynamic confidence weights, the control strategy can be adaptively adjusted according to actual working conditions: when a severe impact is detected at the moment the flying saw contacts the material, the weight of high-frequency components is automatically increased to achieve rapid compensation; during steady-state operation, the low-bandwidth observation is automatically biased to filter noise. Compared to a single bandwidth observer, this method significantly improves the synchronous tracking accuracy and system stability of the flying saw under abrupt cutting conditions.

[0006] Further, calculating the original disturbance estimate includes: ;in Represents the torque constant of the motor; The sampled torque current; The nominal value of the system's moment of inertia; The angular acceleration is calculated based on the angular velocity signal; The angular velocity signal; The set viscous friction coefficient; This represents the original disturbance estimate.

[0007] By combining real-time acquired torque current, angular velocity, and angular acceleration, and using the inverse operation of Newton's second law, the total original disturbance, including load mutations and friction, can be accurately calculated without adding an expensive external torque sensor. This provides a real, full-band data foundation for subsequent filtering and fusion processing.

[0008] Furthermore, both the high-bandwidth filtering and the low-bandwidth filtering are implemented using a second-order Butterworth low-pass filter.

[0009] A second-order Butterworth low-pass filter is selected, which has a steeper cutoff characteristic compared to a first-order filter and a flatter passband amplitude-frequency characteristic compared to filters such as Chebyshev. It can effectively filter out high-frequency measurement noise while preserving the true phase and amplitude information of the disturbance signal to the maximum extent, and avoid excessive signal distortion caused by filtering.

[0010] Furthermore, obtaining the dynamic confidence weights includes: ; This represents the observed value of the high dynamic disturbance; This represents the observed steady-state disturbance; The noise floor threshold; The set gain coefficient; The dynamic confidence weight is denoted as .

[0011] Compared to simple threshold hard switching, the relationship between deviation and noise floor is used to achieve a smooth transition of weights, avoiding abrupt changes or oscillations in the control signal during control mode switching, and ensuring the smoothness of the servo system when transitioning from steady state to cutting state.

[0012] Furthermore, the disturbance compensation command specifically includes: ;in This indicates the disturbance compensation command; This represents the observed value of the high dynamic disturbance; This represents the observed steady-state disturbance; This represents the dynamic confidence weight.

[0013] By using dynamic weights to complement and fuse high-dynamic and steady-state observations, the signal output to the current loop feedforward is ensured to contain both the high-frequency energy required to counteract cutting impacts and to eliminate unnecessary background noise, thus achieving precise torque compensation.

[0014] In a second aspect, the present invention provides a servo synchronous tracking control system for a cold-cutting flying saw, comprising: a synchronous data acquisition and preprocessing unit, used to synchronously acquire the motor angular velocity and torque current in each current loop update cycle, and obtain time-domain aligned angular acceleration signals through detrending term processing and differential operation; a raw torque estimation unit, connected to the synchronous data acquisition and preprocessing unit, used to calculate raw disturbance estimates containing full-frequency components based on the inverse Newtonian dynamics model using the angular velocity, torque current, and angular acceleration; and a dual-channel verification filtering unit, connected to the raw torque estimation unit, used to perform parallel processing on the raw disturbance estimates, dividing... The system outputs high-dynamic disturbance observations that retain high-frequency impact characteristics and steady-state disturbance observations that suppress mechanical noise. A dynamic confidence analysis unit, connected to the dual-channel verification and filtering unit, calculates the instantaneous deviation energy between the high-dynamic disturbance observations and the steady-state disturbance observations, and obtains the dynamic confidence weight based on the relationship between the instantaneous deviation energy and the noise threshold. A torque compensation generation and execution unit, connected to both the dual-channel verification and filtering unit and the dynamic confidence analysis unit, performs linear weighted fusion of the high-dynamic disturbance observations and the steady-state disturbance observations according to the dynamic confidence weight, obtains the disturbance compensation command, and injects it into the current loop feedforward input.

[0015] Compared to general servo control systems, this system integrates synchronous data acquisition, detrending term processing, and dual-channel verification and filtering units. From the system architecture level, it ensures the time-domain alignment of input signals and the parallel execution of processing logic, which can meet the stringent hardware real-time requirements of cold-cutting flying saws for real-time capture and suppression of the extremely short-term impact load generated when the saw blade contacts the steel pipe.

[0016] Furthermore, the original torque estimation unit calculates the original disturbance estimate using the following method: ;in Represents the torque constant of the motor; The sampled torque current; The angular velocity is mentioned. The angular acceleration is calculated based on the angular velocity. This represents the original disturbance estimate.

[0017] Furthermore, the dual-channel verification filter unit includes a first channel and a second channel operating in parallel; the first channel is configured with a low-pass filter with a cutoff frequency of 400Hz for outputting the high dynamic disturbance observation value; the second channel is configured with a low-pass filter with a cutoff frequency of 35Hz for outputting the steady-state disturbance observation value.

[0018] Furthermore, the dynamic confidence analysis unit is configured to calculate the dynamic confidence weights using the following method: ; This represents the observed value of the high dynamic disturbance; This represents the observed steady-state disturbance; The noise floor threshold; This is the gain coefficient; The dynamic confidence weight is denoted as .

[0019] Furthermore, the torque compensation generation and execution unit is configured to calculate the disturbance compensation command by means of the following method: ;in This indicates the disturbance compensation command; This represents the observed value of the high dynamic disturbance; This represents the observed steady-state disturbance; This represents the dynamic confidence weight.

[0020] The technical effects of this invention are as follows:

[0021] This invention proposes a servo control technology for cold-cutting flying saws based on dual-channel verification filtering and dynamic confidence fusion. Its core innovation lies in breaking through the limitation of the single bandwidth of traditional disturbance observers. By processing high-bandwidth and low-bandwidth observations in parallel and generating dynamic confidence weights by comparing their deviation with the noise floor, adaptive fusion is achieved. This resolves the contradiction between high dynamic response and low noise suppression, enabling millisecond-level precise compensation of the flying saw at the moment of cutting impact, significantly improving synchronous tracking performance under complex working conditions. Attached Figure Description

[0022] Figure 1 This is a schematic flowchart illustrating a servo synchronous tracking control method for a cold-cutting flying saw according to an embodiment of the present invention;

[0023] Figure 2 This is a schematic diagram showing a comparison of the amplitude-frequency response curves of the dual-channel filter in an embodiment of the present invention;

[0024] Figure 3 This is a schematic diagram illustrating the time-domain response separation characteristics of the fast-channel and slow-channel observations at the moment of cutting impact in an embodiment of the present invention.

[0025] Figure 4 This is a schematic comparison of a local magnified view of the improved dual-channel fusion algorithm and the traditional single-channel algorithm in the cutting impact region in an embodiment of the present invention.

[0026] Figure 5 This is a schematic bar chart illustrating the normalized comparison of the improved algorithm with the traditional algorithm in terms of three key performance indicators: steady-state noise, peak impulse response, and response time.

[0027] Figure 6 This is a schematic block diagram illustrating the structure of a servo synchronous tracking control system for a cold-cutting flying saw according to an embodiment of the present invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0030] An embodiment of a servo synchronous tracking control method for cold-cutting flying saws:

[0031] like Figure 1 As shown, a servo synchronous tracking control method for a cold-cutting flying saw according to the present invention includes:

[0032] S1. Perform detrending term processing based on synchronous data acquisition link to achieve time-domain phase alignment of discrete signals.

[0033] In this embodiment, the hardware platform of the cold-cutting flying saw servo system exemplarily adopts a high-performance servo driver based on an FPGA+DSP dual-core architecture, wherein the DSP main frequency is not less than 300MHz and the current loop update cycle is set to 62.5. To address the observation errors caused by signal timing misalignment in conventional interference observers, this embodiment first establishes the following synchronous data acquisition link in the underlying hardware logic.

[0034] In this embodiment, the FPGA module of the servo driver is configured to interrupt each current loop cycle ( Within this timeframe, the analog-to-digital converter is synchronously triggered to sample. The system acquires the mechanical angular velocity fed back from the motor encoder in real time. and the torque current setpoint inside the current loop. To eliminate the impact of sensor temperature drift or zero-point offset on subsequent calculations, the controller has an internal capacity... A first-in-first-out (FIFO) circular buffer is used to calculate the moving average of the signal as a DC baseline in real time, and the baseline is subtracted from the original signal to complete the detrending term processing.

[0035] Subsequently, regarding the angular velocity signal The system employs a five-point center difference algorithm for discrete differential operations to obtain high-precision angular acceleration signals. This step ensures strict alignment of the acceleration signal and the torque current signal in the time domain, with the maximum phase deviation controlled within... Within this range, a precise input vector is provided for subsequent calculations.

[0036] S2. Calculate the original torque estimate containing the full frequency band components using the nominal physical parameters combined with the inverse Newtonian dynamics model.

[0037] After completing the timing alignment of the signals, the system constructs an inverse dynamics model using pre-calibrated physical parameters, mapping kinematic quantities to dynamic quantities. In this embodiment, the nominal value of the system's moment of inertia is pre-identified offline. Set as Viscous friction coefficient Set as Next, based on the inverse logic of Newton's second law, the original torque estimate containing information about the total disturbance is calculated. As shown in the following formula:

[0038] ;

[0039] By eliminating the known components used to drive load acceleration and overcome inherent friction from the total electromagnetic torque through the torque balance equation, the remaining part is reduced to the original disturbance term.

[0040] in The torque constant representing the motor is set to [value] in this embodiment. ; The sampled torque current; The calculated angular acceleration; This is the raw, unfiltered disturbance estimate. Since the above calculation process does not incorporate any low-pass filtering, the obtained... The full frequency range of the signal is preserved. This means that both the actual high-frequency impact when the saw blade cuts into the steel pipe (the effective signal) and the high-frequency vibrations generated by gearbox meshing (the ineffective noise) are completely mapped onto the signal. The numerical fluctuations.

[0041] S3. Determine the dual-channel verification to perform differential filtering on the original torque to separate steady-state and transient characteristics.

[0042] To resolve the conflict between bandwidth and noise suppression, this embodiment establishes a parallel dual-channel filter after the signal reconstruction layer.

[0043] The first channel is a fast channel, configured as a high-bandwidth path. In this embodiment, a cutoff frequency is used. Second-order Butterworth low-pass filter This channel allows high-frequency abrupt changes to pass through rapidly, and its output is denoted as the high-dynamic perturbation observation. The channel is designed to cut at the moment of contact, typically lasting less than [a certain duration]. It can reproduce the impact waveform without phase lag.

[0044] The second channel is a slow channel and is configured as a high signal-to-noise ratio path. In this embodiment, a cutoff frequency is used. Second-order Butterworth low-pass filter The channel depth attenuation is higher than... The mechanical vibration component, whose output is denoted as the steady-state disturbance observation. This channel serves as the system's baseline verification signal, used to characterize the average load level during the steady-state sawing process.

[0045] like Figure 2 As shown, this is the amplitude-frequency response curve of the dual-channel filter constructed in this embodiment. The cutoff frequency of the first channel is configured as follows: This channel maintains 0dB gain in the low to mid-high frequency range (<400Hz), allowing high-frequency abrupt signals to pass through without loss; the cutoff frequency of the second channel is configured as follows: The curve shows significant attenuation characteristics for frequency components above 35Hz. (Through...) Figure 2 It can be seen that the two channels constitute a significant bandwidth difference in the frequency domain, which provides a physical basis for the subsequent separation of steady-state load and transient impact.

[0046] S4. Introduce a transient energy analysis mechanism to calculate the deviation amplitude and combine it with the background noise threshold to obtain the dynamic confidence weight.

[0047] In this embodiment, the system also utilizes the transient energy analysis method in signal information theory to calculate the difference between the fast and slow dual-channel output values ​​in real time and convert it into confidence weights. First, the instantaneous deviation energy is calculated. Then the dynamic confidence function is calculated. The calculation formula is as follows:

[0048] ;

[0049] The above calculation method is used to solve the problem of feature identification of non-periodic shocks and periodic clutter, and a soft-switching gating mechanism is realized through a nonlinear mapping function. Indicates the amplitude intensity of high-frequency components; As the noise floor threshold, in this embodiment, it can be taken as the peak value of the background noise during no-load operation. times; The gain coefficient can be set to [value] in this embodiment. , used to adjust the steepness of the function curve; The dynamic confidence weight of the fast channel signal at the current moment, with a value range of [value range missing]. .

[0050] When the system is under no-load or steady-state sawing, the dual-channel deviation is small. As the exponent term approaches infinity, it leads to... Approaching At this point, the system determines that the high-frequency component is mechanical noise and suppresses it.

[0051] When the saw blade contacts the pipe and causes a violent impact, the deviation amplitude instantaneously exceeds the threshold, i.e. The exponential term tends to ,lead to rapidly approaching At this point, the system determines that a real load surge has been detected and fully trusts the fast-channel data.

[0052] like Figure 3 As shown, the time-domain response characteristics of the dual-channel output values ​​are displayed at the instant of the cutting impact, i.e., 500ms (marked in the figure). At the instant of the impact, from 500ms to 502ms, the high-dynamic disturbance observation responds rapidly and deviates significantly from the baseline, while the steady-state disturbance observation responds with a lag and a smaller amplitude due to the low-pass filtering effect. The system calculates the deviation between the two in real time. When this difference exceeds the set noise floor threshold, it can accurately determine the occurrence of the actual load impact, thereby triggering dynamic adjustment of the confidence weight.

[0053] S5. The dual-channel signals are linearly weighted and fused according to the confidence level weights and injected into the current loop feedforward input terminal to achieve compensation execution.

[0054] Finally, the processor calculates the confidence weights in real time. The dual-channel signals are weighted and fused to obtain the final disturbance compensation command. :

[0055] ;

[0056] The calculated results This is directly superimposed on the feedforward input of the servo driver's current loop. Based on the above calibration parameter settings, when the cold-cutting flying saw is in the non-cutting idling and retraction phase, since the background noise is usually low, the weighting... It also remained at a low level, with the system's main output... It effectively shields the high-frequency noise of gear meshing, ensuring smooth motor operation without any whistling.

[0057] At the moment the cutter head contacts the pipe, the impact torque generated by the cutting resistance is usually high and far exceeds the set threshold. , making exist The internal step jumps to a higher level. The system instantly switches to... The fast-channel mode with high bandwidth generates a reverse electromagnetic torque that can respond rapidly within a millisecond-level time window.

[0058] like Figure 4 As shown, this intuitively demonstrates the comparison between the embodiment of the present invention and the traditional single-channel algorithm in terms of time-domain response. At the moment of the 500ms cutting impact: the traditional algorithm exhibits significant phase lag in its response, and the peak value only reaches approximately 65 N·m, failing to fully reproduce the true impact torque; while the improved algorithm of the embodiment of the present invention shows no significant phase lag, and the captured impact peak value reaches approximately 193 N·m. Furthermore, the rapid callback and jitter after the peak value indicate that the improved algorithm can sensitively follow high-frequency load changes without being masked by the filter smoothing, verifying the system's millisecond-level response capability to transient impacts.

[0059] like Figure 5 As shown, this embodiment demonstrates normalized comparison results across three key performance indicators. First, in terms of steady-state noise, this solution reduces it to 0.12, meaning that during the non-cutting phase, the system reduces background noise by approximately 88% through the intervention of the slow channel, effectively resolving the motor whine problem. Second, in terms of peak impact response, this solution improves it by 2.96 times, demonstrating its ability to reproduce real loads. Finally, in terms of response time, this solution shortens it to 0.12, representing a nearly one-order-of-magnitude improvement in response speed. These data strongly support the significant technological advancements of this invention in resolving the contradiction between bandwidth and noise suppression.

[0060] An embodiment of a servo synchronous tracking control system for cold-cutting flying saws:

[0061] On the other hand, the present invention also provides a servo synchronous tracking control system for cold-cutting flying saws. For example... Figure 6 As shown, the system includes the following interconnected and cooperative units:

[0062] Synchronous Data Acquisition and Preprocessing Unit: This unit is configured in the FPGA module of the servo driver and is used to construct a strictly synchronous data acquisition link. This unit operates during each current loop interrupt cycle (e.g., Within this timeframe, the analog-to-digital converter is synchronously triggered to sample and acquire the mechanical angular velocity fed back by the motor encoder in real time. and the torque current setpoint inside the current loop. This unit also includes a first-in-first-out (FIFO) circular buffer (e.g., capacity N=100) for real-time calculation of the signal's moving average as a DC baseline and for detrending processing to eliminate sensor temperature drift or zero-point offset. Furthermore, this unit employs a five-point center difference algorithm to discretize and differentiate the detrended angular velocity signal, obtaining a high-precision angular acceleration signal. This ensures that the acceleration signal and the torque current signal are strictly aligned in the time domain.

[0063] Raw torque estimation unit: This unit is connected to the synchronous data acquisition and preprocessing unit and is used to calculate the raw torque estimate containing full-frequency components using pre-calibrated physical parameters combined with the inverse Newtonian dynamics model. This unit calculates the raw torque estimate based on the formula... Calculate the unfiltered raw disturbance estimate .in, The full-frequency components of the signal are preserved, including the actual high-frequency impact of the saw blade cutting in and the high-frequency noise generated by gearbox meshing. In this embodiment, the nominal value of the system's moment of inertia is... Set as Viscous friction coefficient Set as torque constant Set as .

[0064] Dual-channel verification filter unit: This unit is connected to the original torque estimation unit, constructing a parallel dual-channel topology for evaluating the original torque. Perform differential filtering. This unit specifically includes:

[0065] Fast-channel subunit: configured as a high-bandwidth path, using a cutoff frequency. Second-order Butterworth low-pass filter Used to output observations with high dynamic disturbances The impact waveform is reproduced without phase lag.

[0066] Slow channel subunit: configured as a high signal-to-noise ratio path, using a cutoff frequency. Second-order Butterworth low-pass filter Used to output steady-state disturbance observations It deeply attenuates the mechanical vibration components and uses them as the system's reference verification signal.

[0067] Dynamic Confidence Analysis Unit: This unit is connected to the dual-channel verification filter unit and is used to introduce a transient energy analysis mechanism to calculate the difference between the fast and slow dual-channel output values ​​in real time and generate dynamic confidence weights. This unit first calculates the instantaneous deviation energy. Then, a variant of the Sigmoid function is used to calculate the confidence weight of the fast-channel signal at the current time. The calculation logic is as follows: ,in The noise floor threshold, This is the gain coefficient. This unit implements soft-switching gating through this logic: when the deviation amplitude is less than a threshold, it is considered noise. The deviation approaches 0; when the deviation amplitude exceeds the threshold, it is determined to be a load shock. It rapidly approaches 1.

[0068] Torque compensation generation and execution unit: This unit is connected to the dual-channel verification filtering unit and the dynamic confidence analysis unit, respectively, and is used to perform weighted fusion of the dual-channel signals according to the confidence weight. This unit is based on the formula... Output the final disturbance compensation command. The instruction is then directly superimposed onto the feedforward input of the servo driver current loop to achieve dynamic torque compensation.

[0069] A cold-cutting flying saw servo synchronous tracking control system also includes other components well known to those skilled in the art, such as communication interfaces. Their settings and functions are known in the art and will not be described in detail here.

[0070] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented using computer-readable / executable instructions stored or otherwise maintained by such a computer-readable medium.

Claims

1. A cold cut flying saw servo synchronization tracking control method, characterized in that, The method includes: synchronously acquiring the angular velocity signal and torque current command signal of the motor in each control cycle of the servo driver; Based on the preset motor nominal parameters, the angular velocity signal and torque current given signal are processed according to the inverse operation logic of Newton's second law to obtain the original disturbance estimate containing full-frequency information. The original disturbance estimates are processed in parallel with high-bandwidth filtering and low-bandwidth filtering to obtain high-dynamic disturbance observations and steady-state disturbance observations, respectively. Calculate the amplitude deviation between the high-dynamic disturbance observation and the steady-state disturbance observation, and compare the amplitude deviation with a preset noise floor threshold to obtain a dynamic confidence weight characterizing the reliability of the high-dynamic disturbance observation, including: , This represents the observed value of the high dynamic disturbance. This represents the steady-state disturbance observation value. The noise floor threshold, This is the gain coefficient. The dynamic confidence weight; The high-dynamic disturbance observations and steady-state disturbance observations are weighted and fused under the dynamic confidence weight to obtain the final disturbance compensation command, which is then superimposed on the current loop feedforward input of the servo driver.

2. The servo synchronous tracking control method for a cold-cutting flying saw according to claim 1, characterized in that, Calculating the original disturbance estimate includes: ； in Represents the torque constant of the motor; The sampled torque current; The angular acceleration is calculated based on the angular velocity signal; The nominal value of the system's moment of inertia; The set viscous friction coefficient; The angular velocity signal; This is the original disturbance estimate.

3. The servo synchronous tracking control method for a cold-cutting flying saw according to claim 1, characterized in that, Both the high-bandwidth filtering and low-bandwidth filtering are implemented using a second-order Butterworth low-pass filter.

4. The servo synchronous tracking control method for a cold-cutting flying saw according to claim 1, characterized in that, The disturbance compensation command is specifically as follows: ;in This indicates the disturbance compensation command.

5. A servo synchronous tracking control system for a cold-cutting flying saw, characterized in that, The system includes: A synchronous data acquisition and preprocessing unit is used to synchronously acquire the motor angular velocity and torque current in each current loop update cycle, and obtain the time-domain aligned angular acceleration signal through detrending term processing and differential operation. A raw torque estimation unit, connected to the synchronous data acquisition and preprocessing unit, is used to calculate the raw disturbance estimate containing full-frequency components based on the inverse Newtonian dynamics model using the angular velocity, torque current, and angular acceleration. A dual-channel verification filtering unit, connected to the raw torque estimation unit, is used to process the raw disturbance estimate in parallel and output high-dynamic disturbance observations that retain high-frequency impact characteristics. The system includes a dynamic confidence analysis unit, connected to the dual-channel verification and filtering unit, for calculating the instantaneous deviation energy between the high dynamic disturbance observation and the steady-state disturbance observation, and obtaining the dynamic confidence weight based on the relationship between the instantaneous deviation energy and the preset noise floor threshold; and a torque compensation generation and execution unit, connected to both the dual-channel verification and filtering unit and the dynamic confidence analysis unit, for performing linear weighted fusion of the high dynamic disturbance observation and the steady-state disturbance observation according to the dynamic confidence weight, obtaining the disturbance compensation command and injecting it into the current loop feedforward input. Obtaining dynamic confidence weights includes: , This represents the observed value of the high dynamic disturbance. This represents the steady-state disturbance observation value. The noise floor threshold, This is the gain coefficient. The dynamic confidence weight is denoted as .

6. The servo synchronous tracking control system for a cold-cutting flying saw according to claim 5, characterized in that, The method by which the original torque estimation unit calculates the original disturbance estimate is as follows: ； in Represents the torque constant of the motor; The sampled torque current; The angular acceleration is calculated based on the angular velocity. The nominal value of the system's moment of inertia; The set viscous friction coefficient; The angular velocity is mentioned. This is the original disturbance estimate.

7. The servo synchronous tracking control system for a cold-cutting flying saw according to claim 5, characterized in that, The dual-channel verification and filtering unit includes a first channel and a second channel that operate in parallel. The first channel is equipped with a low-pass filter with a cutoff frequency of 400Hz, which is used to output the high dynamic disturbance observation value; The second channel is equipped with a low-pass filter with a cutoff frequency of 35Hz, which is used to output the steady-state disturbance observation value.

8. The servo synchronous tracking control system for a cold-cutting flying saw according to claim 5, characterized in that, The specific method by which the torque compensation generation and execution unit calculates the disturbance compensation command is as follows: ;in This indicates the disturbance compensation command.

Images