Self-adaptive polishing method and system for outer wall of sampling needle, product and medium

Abstract

The invention discloses a sampling needle outer wall self-adaptive polishing method and system, a product and a medium, and relates to the field of edge calculation. The method comprises the following steps: acquiring force and vibration signals to calculate an energy dissipation index; judging a cutting edge area, a sound window area or a transition area according to the index gradient; in the cutting edge area, a force signal is stabilized by controlling the feeding speed, and the rotating speed of a main shaft is adjusted, so that the kurtosis value tends to be minimum; in the sound window area, the feeding speed is controlled to stabilize the kurtosis target value; and in the transition area, the feeding rotation speed is synchronously adjusted according to the relative position, so that the force and the kurtosis track the dynamically changing target. By implementing the technical scheme provided by the invention, the polishing precision of the sampling needle is improved.

Description

Technical Field

[0001] This application relates to the field of edge computing, and more particularly to an adaptive polishing method, system, product, and medium for the outer wall of a sampling needle. Background Technology

[0002] Currently, with the development of medical and in vitro diagnostic technologies, sampling needles are increasingly widely used in the biomedical field. The surface quality of the outer wall and end face of the sampling needle directly affects puncture performance, biocompatibility, and sample residue; therefore, surface polishing of the outer wall is crucial.

[0003] In related technologies, a force-controlled adaptive robotic polishing method is commonly used. This method integrates a highly sensitive force sensor at the end of the polishing system to monitor the contact pressure between the polishing wheel and the needle surface in real time. When the pressure deviates from the preset value due to changes in the needle profile or clamping errors, the feed position of the polishing wheel is adjusted to maintain a globally constant polishing force, thereby aiming to achieve a uniform material removal rate and surface finish across the entire needle.

[0004] However, in related technologies, since the aim is to eliminate all surface contour undulations to maintain constant contact force, when processing workpieces with discontinuous changes in surface micro-morphology, the instantaneous surge in cutting resistance at the contact point when the grinding wheel enters the micro-rough region from the smooth region will be introduced into the force control closed loop as a high-frequency disturbance, challenging the stability margin of the control system, easily causing small oscillations and overshoot of the polishing force, which in turn leads to uneven and unpredictable material removal in the morphology boundary region, impairing the polishing accuracy. Summary of the Invention

[0005] This application provides an adaptive polishing method, system, product, and medium for the outer wall of a sampling needle to improve the accuracy of the sampling needle polishing process.

[0006] The first aspect of this application provides an adaptive polishing method for the outer wall of a sampling needle, the method comprising: The system collects force and vibration signals generated at the contact point between the grinding wheel and the sampling needle during polishing. It calculates the kurtosis value of the real-time time series data of the vibration signal and weights it with the real-time force signal value according to preset weights to obtain an energy dissipation index sequence. Based on the energy dissipation gradient sequence, it determines the functional region to which the real-time polishing position belongs. When the real-time polishing position is in the cutting edge region, the feed speed is adjusted to stabilize the real-time force signal value at the first preset force value. Simultaneously, based on the real-time feedback of kurtosis value changes, the spindle speed is continuously adjusted to make the real-time kurtosis value approach a local minimum. When the real-time polishing position is in the acoustic window region, the feed speed is adjusted to stabilize the real-time kurtosis value at the first preset kurtosis target value, while monitoring that the peak value of the real-time force signal value does not exceed the second preset force value. When the real-time polishing position is in the transition region, the feed speed and spindle speed are adjusted synchronously according to the position weight coefficient, so that the real-time force signal value and real-time kurtosis value respectively track the dynamic target force value and dynamic target kurtosis value.

[0007] In the above embodiments, an energy dissipation index is constructed by fusing force signals reflecting macroscopic forces with vibration kurtosis values ​​characterizing microscopic morphology. Using the gradient of this index, functional regions with different processing requirements, such as the cutting edge region and the acoustic window region, are autonomously identified. Furthermore, specific control strategies can be applied to different regions, especially actively maintaining the target roughness (kurtosis) in the acoustic window region. This identification-adaptation control mode avoids control oscillations and overshoot caused by target conflicts when a single force control strategy crosses morphological boundaries, ensuring improved accuracy of the sampling needle polishing process across different functional regions and boundaries.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, when the real-time polishing position is in the cutting edge region, by adjusting the feed speed to stabilize the real-time force signal value at a first preset force value, and simultaneously adjusting the direction of increase or decrease of the spindle speed based on the real-time change feedback of the kurtosis value, after the real-time kurtosis value approaches a local minimum, the method further includes: When the real-time kurtosis value is lower than the third preset kurtosis threshold and the real-time force signal is higher than the third preset force value, the control of adjusting the feed rate and adjusting the direction of increase / decrease of the spindle speed is paused, a preset high-frequency sinusoidal disturbance is applied to the spindle speed, and the real-time disturbance force signal value is monitored; when the real-time disturbance force signal value falls from the third preset force value to the fourth preset force value, the application of the high-frequency sinusoidal disturbance is stopped, and the control of adjusting the feed rate and adjusting the direction of increase / decrease of the spindle speed is resumed.

[0009] In the above embodiment, by identifying the contradictory signal of low kurtosis and high force, the grinding wheel clogging problem, which is difficult to detect by conventional control, is diagnosed. Once the diagnosis is confirmed, conventional control is paused and a high-frequency sinusoidal disturbance is applied to the spindle speed. The resulting micro-vibration is used to actively clear the grinding wheel clogging online until the force signal is monitored to return to normal levels, indicating that the grinding wheel's cutting performance has been restored. This avoids machining defects such as burn-in and abrasion caused by grinding wheel passivation, ensuring that the control system always makes decisions based on the actual cutting state, thereby improving the final surface forming quality and geometric accuracy of the sampling needle.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, after applying a preset high-frequency sinusoidal disturbance to the spindle speed, the method further includes: Calculate the root mean square value of the vibration signal; when the root mean square value exceeds the fifth preset threshold, keep the amplitude of the high-frequency sinusoidal disturbance unchanged, and switch the fixed preset frequency to a dynamic frequency that is randomly modulated within the preset frequency bandwidth; monitor the real-time disturbance force signal value until the real-time disturbance force signal value drops from the third preset force value to the fourth preset force value, and then stop applying the dynamic frequency.

[0011] In the above embodiments, during online chip removal using high-frequency disturbances, the root mean square value of the vibration signal is calculated in real time to monitor and identify resonance states that may damage the workpiece. Once resonance is detected, chip removal is not stopped; instead, while maintaining the disturbance amplitude, the fixed frequency is switched to a dynamic random frequency. This disrupts the stable conditions for resonance establishment, thereby achieving efficient chip removal while actively suppressing harmful vibrations. This avoids secondary damage to the sampling needle itself during the repair of the grinding wheel, ultimately ensuring the machining accuracy and surface integrity of the sampling needle.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, the functional region to which the real-time polishing position belongs is determined based on the energy dissipation gradient sequence, specifically including: When the value of the energy dissipation gradient sequence first exceeds the sixth preset gradient threshold, the current position is marked as a potential boundary point. From the potential boundary point, continue polishing forward along the axis for a preset verification distance. Within the verification distance, monitor the value of the energy dissipation index sequence. When the value of the energy dissipation index sequence has fallen back below the normal baseline value at the end of the verification distance, ignore this boundary judgment, and at the same time, reduce the first preset force value by the preset compensation value within the subsequent preset compensation distance. When the value of the energy dissipation index sequence continues to stabilize at a new platform value different from the normal baseline value at the end of the verification distance, the potential boundary point is confirmed as the true boundary.

[0013] In the above embodiments, when the energy dissipation gradient first exceeds the limit, the boundary is not directly determined. Instead, it is first marked as a potential boundary point. Then, a verification distance is polished forward, and the subsequent behavior of the signal within this interval is analyzed to distinguish whether the point is a false boundary caused by occasional material anomalies, where the signal will quickly fall back, or a true boundary where the physical properties have continuously changed, and the signal will stabilize on the new platform. This closed-loop decision-making logic of "hypothesis-verification" can filter out misjudgments caused by internal material defects or noise, ensuring that the exclusive processing strategies for different functional areas are applied to the correct physical location, thereby improving the final contour and surface processing accuracy of the sampling needle.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, after monitoring the values ​​of the energy dissipation index sequence within the verification distance, the method further includes: Calculate the sequence variance of the energy dissipation index sequence; when the value of the energy dissipation index sequence does not fall below the normal baseline value and the sequence variance exceeds the preset unstable variance threshold at the end of the verification distance, determine the potential boundary point as the boundary of the heat-affected zone, and reduce the first preset force value and spindle speed to the preset low energy processing parameters until the sequence variance falls below the unstable variance threshold.

[0015] In the above embodiments, within the hypothesis-verification framework, the ability to identify "unstable" states is achieved by introducing the statistical dimension of the sequence variance of energy dissipation indices. This not only distinguishes between stable plateaus and falling signals but also identifies physically uneven and fragile heat-affected zones, automatically switching to dedicated low-energy processing parameters for a safe transition, avoiding damage such as microcracks in these areas. This ability to identify and adaptively process complex boundaries ensures the processing integrity of special areas of the workpiece and improves the overall forming accuracy of the sampling needle.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, when the real-time polishing position is within the acoustic window region, after adjusting the feed speed to stabilize the real-time kurtosis value at a first preset kurtosis target value, and simultaneously monitoring that the peak value of the real-time force signal does not exceed a second preset force value, the method further includes: When the negative rate of change of the real-time kurtosis value exceeds the preset rate of change threshold and the decrease in value within the preset time window exceeds the preset amplitude threshold, the adjustment of the feed speed is paused, and the direction of increase or decrease of the spindle speed is continuously adjusted according to the real-time change feedback of the kurtosis value; until the real-time kurtosis value rises back to the first preset kurtosis target value, the adjustment of the feed speed is resumed.

[0017] In the above embodiment, the kurtosis drop caused by entering a locally overly smooth region is identified by using both the rate of change and amplitude of the kurtosis value. At this point, the normal feed rate adjustment is paused, preventing the controller from making drastic speed increases to compensate for errors, which could lead to scratches on the workpiece surface. Instead, a gentler method of fine-tuning the spindle speed is adopted to actively repair the surface texture. This "freeze-repair" intelligent intervention transforms potential machining defects into an online self-healing process, ultimately improving the surface machining accuracy of the sampling needle.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, when the real-time polishing position is in the transition zone, after synchronously adjusting the feed rate and spindle speed according to the position weighting coefficient so that the real-time force signal value and the real-time kurtosis value respectively track the dynamic target force value and the dynamic target kurtosis value, the method further includes: When the force deviation and kurtosis deviation both exceed their respective preset error limits, the interpolation control logic of the transition zone is terminated; the first preset kurtosis target value and the second preset force value of the acoustic window zone are used as the current target value.

[0019] In the above embodiments, by setting dual monitoring of force deviation and kurtosis deviation, abrupt changes in the physical property boundaries within the transition zone that are steeper than the preset model are identified. Once identified, the currently inapplicable interpolation control is stopped, and a stable control strategy for the subsequent acoustic window region is switched in advance. This avoids oscillations and overshoots that may occur when the control system tracks a failed target, prevents processing defects at the boundaries of critical functional areas, and thus ensures the overall forming accuracy of the sampling needle.

[0020] In a second aspect, embodiments of this application provide an adaptive polishing system for the outer wall of a sampling needle. The adaptive polishing system for the outer wall of the sampling needle includes: one or more processors and a memory; the memory is coupled to the one or more processors and is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the adaptive polishing system for the outer wall of the sampling needle to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on an adaptive polishing system for the outer wall of a sampling needle, cause the adaptive polishing system for the outer wall of the sampling needle to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on an adaptive polishing system for the outer wall of a sampling needle, cause the adaptive polishing system for the outer wall of the sampling needle to perform the method described in the first aspect and any possible implementation thereof.

[0023] Understandably, the adaptive polishing system for the outer wall of the sampling needle provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the adaptive polishing method for the outer wall of the sampling needle provided in the embodiments of this application. Therefore, the beneficial effects that can be achieved can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. This application constructs an energy dissipation index by fusing force signals reflecting macroscopic forces with vibration kurtosis values ​​characterizing microscopic morphology. Using the gradient of this index, functional regions with different processing requirements, such as the cutting edge region and the acoustic window region, are autonomously identified. Furthermore, specific control strategies can be applied to different regions, especially actively maintaining the target roughness (kurtosis) in the acoustic window region. This identification-adaptation control mode avoids control oscillations and overshoot caused by target conflicts when a single force control strategy crosses morphological boundaries, ensuring improved accuracy of the sampling needle polishing process across different functional regions and boundaries.

[0025] 2. This application diagnoses grinding wheel clogging problems that are difficult to detect by conventional control by identifying the contradictory signal of low kurtosis and high force values ​​coexisting. Once the diagnosis is confirmed, conventional control is paused and a high-frequency sinusoidal disturbance is applied to the spindle speed. The resulting micro-vibration actively and online clears the grinding wheel clogging until the force signal is monitored to return to normal levels, indicating that the grinding wheel's cutting performance has been restored. This avoids machining defects such as burn-in and abrasion caused by grinding wheel passivation, ensuring that the control system always makes decisions based on the actual cutting conditions, thereby improving the final surface finish and geometric accuracy of the sampling needle.

[0026] 3. This application, when applying high-frequency disturbances for online chip removal, monitors and identifies potential workpiece-damaging resonance states by calculating the root mean square value of the vibration signal in real time. Once resonance is detected, chip removal is not stopped; instead, while maintaining the disturbance amplitude, the fixed frequency is switched to a dynamic random frequency. This disrupts the stability conditions for resonance establishment, thereby achieving efficient chip removal while actively suppressing harmful vibrations. This avoids secondary damage to the sampling needle itself during the repair of the grinding wheel, ultimately ensuring the machining accuracy and surface integrity of the sampling needle. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating an adaptive polishing method for the outer wall of the sampling needle in an embodiment of this application. Figure 2 This is another flowchart illustrating the adaptive polishing method for the outer wall of the sampling needle in this application embodiment; Figure 3 This is an exemplary hardware structure diagram of an adaptive polishing system for the outer wall of the sampling needle in an embodiment of this application. Detailed Implementation

[0028] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0029] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0030] In related technologies, force-controlled adaptive robotic polishing methods are commonly used. This method integrates a highly sensitive force sensor at the end of the polishing system to monitor the contact pressure between the grinding wheel and the needle surface in real time, aiming to maintain a globally constant polishing force across the entire workpiece surface to achieve uniform material removal. However, for workpieces like sampling needles with specific functional zones, their surfaces are not homogeneous but consist of regions with different microstructures and processing objectives (such as areas requiring sharp cutting edges and areas requiring specific roughness, like acoustic windows). Related technologies treat all morphological changes indiscriminately as disturbances to the constant force. When the grinding wheel crosses the boundaries of functional zones, the force signal fluctuates drastically due to abrupt changes in material properties, forcing the control system to overcompensate. This leads to oscillations, overshoot, and even processing defects at critical boundaries, making it difficult to achieve refined processing of different functional areas.

[0031] In this embodiment, the vibration kurtosis value characterizing the micro-morphology is weighted and fused with the force signal reflecting the macroscopic force to construct an energy dissipation index that more comprehensively reflects the contact state. Based on the gradient change of this index, different functional regions such as the cutting edge region, acoustic window region, and transition region can be identified online autonomously, thus abandoning the one-size-fits-all control approach in related technologies. Furthermore, a dedicated composite control strategy is applied to different regions: in the cutting edge region, sharpness is ensured by force control combined with kurtosis optimization; in the acoustic window region, kurtosis control is the main method to actively maintain a specific texture; and in the transition region, interpolation control is used to achieve a smooth transition. This intelligent zoning and adaptive control paradigm solves the inherent contradictions and instabilities of single-target control at the morphology boundary, ensuring that the processing of each functional region can meet specific design requirements, achieving high-precision, customized polishing.

[0032] Figure 1 This is a flowchart illustrating the adaptive polishing method for the outer wall of the sampling needle in this application embodiment, including the following steps: S101. Collect the force and vibration signals generated at the contact point between the grinding wheel and the sampling needle during the polishing process.

[0033] Among them, the force signal refers to the force collected by the force sensor, which represents the normal pressure and tangential friction force between the grinding wheel and the sampling needle during the contact and material removal process; the vibration signal refers to the high-frequency mechanical vibration at the contact point caused by factors such as surface unevenness, material cutting, and abrasive grain shedding.

[0034] Specifically, to capture instantaneous changes during the polishing process, sensors are typically deployed closest to the point of application. For example, a multi-dimensional force sensor is integrated under the fixture that holds the sampling needle, or an accelerometer is mounted on the spindle of the polishing wheel. These sensors operate at a sampling rate much higher than the control system's response frequency, ensuring complete capture of all high-frequency dynamic information, including impacts and chatter. Before being sent to the controller, the raw electrical signal typically undergoes preprocessing such as amplification and filtering by signal conditioning circuitry to remove environmental electromagnetic interference and irrelevant mechanical noise, thereby improving the signal-to-noise ratio.

[0035] S102. Calculate the kurtosis value of the real-time time series data of the vibration signal, and then perform a weighted summation of the kurtosis value and the real-time value of the force signal according to a preset weight to obtain the energy dissipation index sequence.

[0036] Among them, kurtosis value refers to the ratio of the fourth central moment to the square of the variance of the vibration signal time series data points. It is a dimensionless statistical quantity that represents the steepness or flatness of the distribution shape of the sequence. In this scenario, it is related to the micro-roughness of the sampling needle surface. A high kurtosis value usually indicates that there are a large number of impact protrusions on the surface. The preset weight is a value determined in advance based on a large amount of process experimental data and optimization algorithm. It is used to balance the importance of macroscopic force and micro-vibration in the final evaluation index. The energy dissipation index sequence refers to the composite index time series that characterizes the total energy consumption of the processing area per unit time after fusing information from the two different physical dimensions of force and kurtosis.

[0037] Specifically, energy dissipation during polishing manifests in two main aspects: the energy consumed by the grinding wheel applying pressure to the workpiece to induce plastic deformation and material removal, primarily characterized by the magnitude of the force signal; and the energy consumed by the impact and scraping between the abrasive grains on the grinding wheel and the microscopic peaks and valleys on the workpiece surface, characterized by the impact characteristics (i.e., kurtosis) of the vibration signal. During calculation, a fixed-length sliding time window is established, and the kurtosis of the vibration signal within the window is calculated in real time. Simultaneously, the average or instantaneous value of the force signal within the time window is extracted. Finally, the index value at the current moment is calculated using the formula: Energy Dissipation Index = α * Kurtosis Value + β * Force Signal Value (where α and β are preset weights). As the time window slides, a continuous sequence of energy dissipation indices is formed.

[0038] The core technical idea behind this step is that a single sensor signal cannot fully describe the physical process of polishing, necessitating the effective fusion of information from different dimensions. Specifically, the force signal primarily reflects the energy consumed by macroscopic plastic deformation and material removal, while the kurtosis value characterizes the energy consumed by the impact and scratching of abrasive particles and surface peaks and valleys at the microscopic level. Combining the two allows for a more comprehensive depiction of the total energy dissipation state at the contact interface. If only the force signal is used as feedback, it is difficult to distinguish between two drastically different processing states: a lower force acting on a rough surface (efficient cutting); and a higher force acting on a smooth surface (potentially causing "scorching" or "rubbing"). These two situations may produce similar force readings, but their physical nature and processing effects are vastly different. Similarly, if only the kurtosis value is used, while surface texture can be perceived, the true material removal rate is difficult to ascertain, potentially leading to dimensional deviations in the macroscopic profile.

[0039] Therefore, by constructing an energy dissipation index sequence and fusing information from two different physical dimensions—force and kurtosis—the resulting composite index can more robustly characterize the total energy consumption of the processing area per unit time. In the specific calculation, a fixed-length sliding time window is opened, and the kurtosis of the vibration signal within the window is calculated in real time. Simultaneously, the average or instantaneous value of the force signal within this time window is extracted. Finally, the index value at the current moment is calculated using the formula: Energy dissipation index = α * kurtosis value + β * force signal value (where α and β are preset weights). As the time window continuously slides, an energy dissipation index sequence that accurately reflects changes in the processing state is formed.

[0040] The preset weights α and β are parameters used to balance the relative importance of macroscopic forces and microscopic morphological information in boundary identification. The method for determining these weights includes the following steps: selecting a batch of standard sampling needles with clear and known functional region boundary locations as calibration samples; setting a set of α and β combinations covering different weight ratios for testing, where α + β = 1; using each weight combination, calculating the sensor data of the calibration samples during real or simulated polishing processes to generate corresponding energy dissipation index sequences and gradient sequences; comparing and analyzing the gradient sequences under each weight set, and selecting the set of α and β that maximizes the signal-to-noise ratio of the gradient peak at the known functional region boundary (i.e., maximizing the distinction between the boundary peak and the background gradient fluctuations in non-boundary areas), as the final preset weights. This process aims to maximize the saliency of boundary features and ensure the accuracy of subsequent identification.

[0041] In some embodiments, real-time computation can be achieved in a variety of ways: Optionally, a calculation scheme based on a digital signal processor (DSP) is adopted: the collected force and vibration signal data stream is sent into a dedicated DSP chip; the high-speed multiply-accumulate unit inside the DSP is used to realize the pipelined and efficient calculation of statistics (mean, variance, fourth-order central moment) within the sliding window; the weighted summation is completed inside the DSP, and the final energy dissipation index sequence is sent to the upper motion controller through the bus.

[0042] It is understandable that other methods can be used to generate the energy dissipation index in this step, such as using software algorithms on a high-performance multi-core CPU, which is not limited here.

[0043] S103. Determine the functional region to which the real-time polishing position belongs based on the energy dissipation gradient sequence.

[0044] Among them, the energy dissipation gradient sequence refers to the sequence formed by the rate of change of the energy dissipation index values ​​along the axial position of the sampling needle. A large gradient value indicates that the physical properties of the polished surface have changed drastically. The functional area refers to the specific sections on the sampling needle with different surface precision and morphology requirements according to medical use, including the cutting edge area, the acoustic window area, and the transition area. The cutting edge area refers to the area at the needle tip that requires extremely high smoothness and sharpness. The acoustic window area refers to the rough area on the needle shaft that is specially processed to enhance the ultrasonic imaging effect. The transition area refers to the intermediate zone connecting the smooth area and the rough area.

[0045] Specifically, this step enables the identification of the current region by sensing changes in surface physical properties, rather than relying on preset absolute coordinates, thereby achieving self-adaptation to workpiece installation errors and batch variations. The energy dissipation index I and the corresponding axial position P are recorded. The gradient G can be calculated differentially: G = (I_current - I_previous) / (P_current - P_previous). Moving from a smooth needle bar (low and stable energy dissipation value, gradient close to zero) to a rough acoustic window region (high energy dissipation value and large fluctuations) generates a significant positive gradient pulse. Conversely, leaving the acoustic window region generates a negative gradient pulse. By setting a gradient threshold (the gradient threshold is a reliable criterion value for distinguishing between real boundaries and background fluctuations, determined by preprocessing experiments on multiple batches of standard sampling needles, statistically analyzing the typical amplitude of energy dissipation gradient pulses generated at the functional region boundaries, and combining this with the signal noise level during processing), once a gradient pulse exceeds the gradient threshold, the functional region state is switched.

[0046] In some embodiments, online determination of functional areas can be achieved in multiple ways: Optionally, a threshold-based finite state machine (FSM) scheme can be adopted: a finite state machine is established that includes states such as "edge region", "transition region", and "acoustic window region"; different positive and negative gradient thresholds are set as triggering conditions for state transitions. For example, when the state is in the "transition region" state and the energy dissipation gradient is detected to exceed the positive threshold T1, the state machine transitions to the "acoustic window region" state and loads the control strategy for that region.

[0047] It is understandable that other methods can be used to achieve adaptive recognition of the functional area in this step, such as using a one-dimensional convolutional neural network (1D-CNN) to extract features and classify gradient sequences, which is not limited here.

[0048] S104. When the real-time polishing position is in the cutting edge area, the feed speed is adjusted to stabilize the real-time force signal value at the first preset force value. At the same time, based on the real-time change feedback of the kurtosis value, the direction of increase or decrease of the spindle speed is continuously adjusted so that the real-time kurtosis value approaches the local minimum value.

[0049] The cutting edge zone refers to the part at the very tip of the sampling needle responsible for piercing, and its quality directly determines the smoothness of piercing and the size of the wound. The first preset force value is the optimal polishing pressure reference value pre-set through process experiments based on the material properties of the cutting edge zone and the desired material removal rate. It is determined through process experiments (Design of Experiments, DOE), with the final surface roughness (Ra) and material removal rate (MRR) of the cutting edge zone as optimization targets. By changing the polishing force, the range of force values ​​that can simultaneously obtain the lowest surface roughness and the highest material removal efficiency is found, and a stable value is selected from it. The local minimum value refers to the relatively lowest kurtosis value that can be achieved by adjusting the spindle speed under the current combination of processing parameters, representing the limit of surface finish under the polishing pressure.

[0050] Specifically, when S103 determines that the current area is a cutting edge zone, it aims to achieve the highest surface finish (lowest kurtosis) and stable material removal rate (constant force value). During execution, there are two parallel control loops: Force control: The real-time force signal value is compared with a first preset force value, and the feed axis speed is adjusted continuously in real time using closed-loop control algorithms such as PID (proportional-integral-derivative). If the force value is too small, the feed is increased; if the force value is too large, the feed is decreased, thereby achieving constant control of the polishing pressure.

[0051] Kurtosis optimization: Instead of directly controlling the kurtosis to a fixed value, the spindle speed is adjusted exploratoryly. For example, a small increment is given to the spindle speed, and the trend of kurtosis changes over a subsequent period of time is observed. If the kurtosis decreases, it indicates that the adjustment is in the right direction, and the speed will be increased again in the next cycle; if the kurtosis increases instead, it indicates that the adjustment is in the wrong direction, and the adjustment will be reversed (the speed will be decreased) in the next cycle. Through continuous trial-and-error feedback-adjustment, the speed that makes the surface smoothest (lowest kurtosis) is dynamically found.

[0052] In some embodiments, the dual-objective control in this step can be implemented in a variety of ways: Optionally, a decoupled PID control scheme can be adopted: design a force feedback PID controller, whose input is the force error (real-time force - target force) and output is the adjustment amount of the feed speed; design an independent kurtosis optimization logic module, whose input is the real-time rate of change of the kurtosis value and output is the increase / decrease command of the spindle speed; the two modules run in parallel, with the force control PID running at a high frequency and the kurtosis optimization logic running at a slightly lower frequency to wait for the kurtosis change trend to stabilize.

[0053] It is understandable that other methods can be used to achieve coordinated control in this step, such as using fuzzy control or neural networks to establish a complex nonlinear model between force, kurtosis, feed, and speed, which is not limited here.

[0054] In some embodiments, in the edge area polishing control of S104, in order to deal with the pseudo-smooth state caused by the coexistence of low kerf and high force due to grinding wheel blockage, a variable frequency high-frequency disturbance can be actively applied to the spindle speed to remove chips online and suppress potential resonance, thereby restoring the grinding wheel cutting performance and ensuring the final machining accuracy of the edge.

[0055] Specifically, this supplementary step aims to address a hidden problem in the polishing stage of the cutting edge area: grinding wheel clogging. Under the S104 control strategy, the goal is to minimize the kurtosis value. However, when the micropores of the grinding wheel become clogged with chips, the cutting ability decreases, turning into rubbing or burning, which also leads to reduced vibration and impact, further lowering the kurtosis value. Simultaneously, due to the surge in friction, the force required to maintain the feed abnormally increases. At this point, one might mistakenly believe that the ideal polishing effect has been achieved, but in reality, the machining quality is deteriorating. To solve this problem: First, continuously monitor whether two key indicators simultaneously meet the trigger conditions: the real-time kurtosis value is lower than the third preset kurtosis threshold, and the real-time force signal is higher than the third preset force value. The third preset kurtosis threshold is a critical value set based on extensive experimental data, characterizing that the cutting edge surface has entered an ultra-precision machining state; the third preset force value is the upper limit that the normal cutting force should not exceed at this kurtosis level, calibrated through process experiments. When both conditions are met simultaneously, it is highly probable that the grinding wheel has become clogged.

[0056] Once triggered, the normal PID control is paused, and the system enters active chip removal mode: a preset high-frequency sinusoidal disturbance is superimposed on the spindle's reference speed. This disturbance essentially causes the grinding wheel to vibrate at high frequencies and amplitudes, acting like a miniature dust shaker. It uses inertial force to dislodge the chips clogging the abrasive grains, thus restoring the grinding wheel's sharpness. During this period, the fluctuation value of the force signal generated after the superimposed disturbance is closely monitored; this is the real-time disturbance force signal value, and its magnitude directly reflects the severity of the blockage.

[0057] However, applying a fixed high-frequency vibration to a slender sampling needle carries the risk of inducing structural resonance, potentially causing severe vibration and damage to the workpiece. Therefore, this step also includes a safety vibration suppression mechanism: synchronously calculating the root mean square (RMS) value of the vibration signal to characterize the total vibration energy. When this RMS value exceeds a fifth preset threshold (the fifth preset threshold is a critical energy value pre-set based on modal analysis of the sampling needle or experimental impact testing, used to characterize that the workpiece has entered a dangerous resonance state), resonance is determined to have occurred. At this point, the disturbance amplitude remains unchanged, but the fixed disturbance frequency is switched to a dynamic frequency that is randomly or swept within a preset frequency bandwidth. This rapid, irregular frequency change disrupts the stable conditions for resonance establishment, actively suppressing harmful vibrations of the workpiece without stopping the chip removal operation.

[0058] Finally, regardless of whether the system executes fixed-frequency or dynamic variable-frequency disturbance, the completion criterion is consistent: continuously monitor the real-time disturbance force signal value. When this value drops from above the third preset force value to below the fourth preset force value (the fourth preset force value is experimentally calibrated and represents a state where the grinding wheel is clean and the rotational resistance is minimal), the chip removal is considered complete. At this point, all disturbances are immediately stopped, and the system seamlessly reverts to the original control logic of S104 for adjusting the feed speed and spindle speed, continuing polishing.

[0059] The above technical steps diagnose and resolve grinding wheel clogging issues online by identifying low-kurtosis and high-force pseudo-smoothness signal characteristics. The dynamic frequency conversion disturbance strategy, while efficiently removing chips, avoids the risk of workpiece resonance, ensuring that the control system always makes decisions based on the actual cutting state, thus improving the final forming accuracy and consistency of the cutting edge profile.

[0060] S105. When the real-time polishing position is in the acoustic window area, the feed speed is adjusted to stabilize the real-time kurtosis value at the first preset kurtosis target value, while monitoring that the peak value of the real-time force signal does not exceed the second preset force value.

[0061] The acoustic window region refers to the rough surface on the sampling needle specifically manufactured to enhance ultrasonic wave reflection. The first preset kurtosis target value is an ideal kurtosis value determined experimentally based on the required ultrasonic imaging clarity. It directly corresponds to a specific surface roughness and is calibrated according to the functional requirements of the acoustic window region (such as ultrasonic imaging clarity). By preparing a series of samples with different kurtosis values ​​(corresponding to different roughnesses) and conducting functional tests (such as ultrasonic imaging tests), the sample with the best imaging effect is found, and the surface kurtosis value is calibrated as the first preset kurtosis target value. The second preset force value is a safety upper limit threshold, which is preset based on the structural stiffness and material strength of the sampling needle. It is used to prevent the needle from deforming or being damaged due to excessive force when manufacturing the rough surface. The safety threshold is determined through mechanical analysis or step-by-step loading experiments on the sampling needle. This value is significantly lower than the ultimate load that would cause plastic deformation or structural damage to the sampling needle, and leaves sufficient safety margin.

[0062] Specifically, when S103 determines that the machine has entered the acoustic window zone, it creates and maintains a specific roughness. Therefore, the core of the control switches to kurtosis control. The real-time kurtosis value is compared with a first preset kurtosis target value, and the feed rate is adjusted using closed-loop algorithms such as PID. If the real-time kurtosis is too low (the surface is too smooth), the feed rate is increased, making the interaction between the abrasive grains and the workpiece more intense; if the real-time kurtosis is too high (the surface is too rough), the feed rate is slowed down, making the interaction smoother. At the same time, the force signal is continuously collected and compared with a second preset force value. Once the instantaneous peak value of the force exceeds this safety threshold, a protection mechanism is immediately triggered, such as pausing the feed or alarm shutdown, to ensure the safety of the machining process.

[0063] In some embodiments, master-slave control in this step can be implemented in a variety of ways: Optionally, a model predictive control (MPC) scheme is adopted: a dynamic model is established that can predict the impact of different feed rate sequences on kurtosis and force values ​​over a short period of time in the future; in each control cycle, an optimization problem is solved: under the premise of satisfying the peak force constraint (not exceeding the second preset force value), the optimal feed rate control sequence is found to minimize the error between the predicted kurtosis value and the target value; the first value of the optimized control sequence is applied to the actuator.

[0064] It is understandable that other methods can be used to achieve the constrained single-objective control in this step, such as using a fuzzy logic controller with safety monitoring, which is not limited here.

[0065] S106. When the real-time polishing position is in the transition zone, the feed speed and spindle speed are adjusted synchronously according to the position weight coefficient so that the real-time force signal value and the real-time kurtosis value track the dynamic target force value and the dynamic target kurtosis value respectively.

[0066] The transition zone is the middle area connecting the cutting edge / smooth needle bar and the acoustic window zone. Its function is to avoid stress concentration caused by abrupt changes in surface characteristics. The position weight coefficient refers to the relative position of the current polishing point within the entire length of the transition zone, and is a normalized value from 0 to 1. The dynamic target force value and dynamic target kurtosis value are target curves that change smoothly with position by linear interpolation of the target value at the starting point of the transition zone (such as the first preset force value and the minimum kurtosis of the cutting edge zone) and the target value at the ending point (such as the safety force value of the acoustic window zone and the first preset kurtosis target value).

[0067] Specifically, when S103 determines that the position is in the transition zone, a smooth switching control strategy is executed to ensure that the transition from a stable operating state (such as the cutting edge zone control mode) to another stable operating state (such as the sound window zone control mode) is gradual and shock-free. During execution, firstly, based on the current axial position read from the encoder, the position weight coefficient k is calculated (e.g., k = (current position - transition zone start point) / (transition zone end point - transition zone start point)). Then, the control targets of the two zones are linearly interpolated using the coefficient k to generate the dynamic target value for the current position: dynamic target force = (1-k) * start point target force + k * end point target force; dynamic target kurtosis = (1-k) * start point target kurtosis + k * end point target kurtosis. Subsequently, two independent PID control loops are simultaneously activated. The force control loop tracks the dynamically changing target force value by adjusting the feed rate, and the kurtosis control loop tracks the dynamically changing target kurtosis value by adjusting the spindle speed, thereby achieving synchronous and smooth transition of the two key physical quantities.

[0068] In some embodiments, to address the control conflict caused by the mismatch between the smooth transition control model of S106 and the physically sharp regional boundaries, the tracking error of the force and kurtosis dual channels can be monitored, and the interpolation can be stopped and the control mode of the next region can be transitioned to in advance when the error synchronization exceeds the limit, thereby protecting the geometric accuracy of the critical boundary.

[0069] Specifically, this aims to address a unique but critical problem in the fabrication of high-end sampling probes. The linear interpolation transition strategy described in S106 is effective only if the physical boundaries of the sampling probe's functional areas are gradual and smooth. However, for acoustic windows manufactured using advanced processes such as laser etching, the boundaries may be microscopically cliffs with almost no physical transition. When the smooth control model of S106 encounters such a physical abrupt change, a catastrophic control conflict occurs: the actual physical interaction changes instantaneously the moment the grinding wheel touches the edge of the cliff. The real-time force signal generates a huge positive peak due to the impact, while the real-time kurtosis value also spikes instantaneously due to scraping the rough surface. However, since the position is still in the early stage of the transition zone, the dynamic target force and dynamic target kurtosis values ​​calculated by interpolation remain at a low level.

[0070] The loss of control is determined when two error values—force deviation (the difference between the real-time force signal and the dynamic target force value) and kurtosis deviation (the difference between the real-time kurtosis value and the dynamic target kurtosis value)—simultaneously and instantaneously become extremely large, far exceeding the fluctuation range under normal control. This dual loss of control phenomenon is used as the criterion. The preset error limit is not an empirical physical threshold, but rather a boundary value defined by simulation and experimentation based on its own performance, used to define its stable tracking capability. Any deviation exceeding this limit is considered as the control loop being unable to effectively track its target.

[0071] Once the force deviation and kurtosis deviation are detected to exceed their respective preset error limits, the exception handling procedure for model failure will be triggered: First, the interpolation control logic in the transition zone is immediately suspended. This means that the dynamic target value is no longer calculated based on the position weight coefficient, and the smooth transition scheme of S106 is abandoned.

[0072] Then, skipping all remaining transition steps, the first preset kurtosis target value and the second preset force value of the acoustic window region are directly used as the current target values ​​and forcibly loaded into the control system. This is equivalent to allowing the system to transition to the next stable operating state in advance (i.e., the acoustic window region control mode of S105). The advantage of doing this is that it no longer attempts to fit an unfittable physical reality with an incorrect model, avoiding severe oscillations in the control system (especially the PID integral term) due to huge tracking errors. This prevents the grinding wheel from being chipped, rounded, or fluttering at critical boundaries, protecting the sharpness that the acoustic window boundary should have.

[0073] The above technical steps use the dual runaway state inside the control system as the identification criterion. By abandoning the mismatched smooth transition model and jumping ahead, control oscillations and processing defects caused by multi-target tracking conflicts at sharp boundaries are avoided. This maximizes the protection of the geometric integrity of the microscopic sharp boundaries manufactured by advanced technology and improves the processing accuracy of high-end sampling needles.

[0074] In the above embodiments, an energy dissipation index is constructed by fusing force signals reflecting macroscopic forces with vibration kurtosis values ​​characterizing microscopic morphology. Using the gradient of this index, functional regions with different processing requirements, such as the cutting edge region and the acoustic window region, are autonomously identified. Furthermore, specific control strategies can be applied to different regions, especially actively maintaining the target roughness (kurtosis) in the acoustic window region. This identification-adaptation control mode avoids control oscillations and overshoot caused by target conflicts when a single force control strategy crosses morphological boundaries, ensuring improved accuracy of the sampling needle polishing process across different functional regions and boundaries.

[0075] In some other embodiments of this application, during the polishing process, when encountering occasional hard points or residual stress release zones within the material, these transient signal pulses may be misinterpreted as true functional area boundaries, leading to premature switching of processing strategies and resulting in surface damage. The adaptive polishing method for the outer wall of the sampling needle provided in this application can distinguish between transient anomalies and true boundaries by initiating a verification process and analyzing the persistence of subsequent signals, thereby filtering out interference from false boundaries.

[0076] like Figure 2 The diagram shown is another flowchart illustrating the adaptive polishing method for the outer wall of the sampling needle provided in this application embodiment, including the following steps: S201. Collect the force and vibration signals generated at the contact point between the grinding wheel and the sampling needle during the polishing process.

[0077] S202. Calculate the kurtosis value of the real-time time series data of the vibration signal, and then perform a weighted summation of the kurtosis value and the real-time value of the force signal according to a preset weight to obtain the energy dissipation index sequence.

[0078] S203. Determine the functional region to which the real-time polishing position belongs based on the energy dissipation gradient sequence.

[0079] Steps S201-S213 and Figure 1 Steps S101-S103 in the illustrated embodiment are similar and can be found in the descriptions of steps S101-S103, which will not be repeated here.

[0080] S204. When the value of the energy dissipation gradient sequence exceeds the sixth preset gradient threshold for the first time, mark the current position as a potential boundary point.

[0081] The sixth preset gradient threshold is a gradient critical value pre-set based on a large number of process experiments. The setting principle is that this value should be higher than the signal background noise gradient in normal processing, but lower than the minimum effective gradient generated at the boundary of the real functional area. It is used to screen out possible boundary locations. Potential boundary points refer to candidate boundary locations that have been initially identified but not yet finally confirmed. They are temporary markers that need to be verified in subsequent steps.

[0082] Specifically, the continuous execution after S203 calculates the energy dissipation gradient sequence is the triggering step for the "hypothesis-verification" boundary identification logic. The application scenario is to address the difficulty of simple gradient thresholding methods in distinguishing between "real, continuous functional zone boundaries" and "instantaneous, isolated physical anomalies." These anomalies may originate from tiny hard phases within the material, the instantaneous release of local residual stress during processing, etc., all of which generate gradient pulses similar to real boundaries, leading to misjudgments. The purpose of this step is to, when the first matching gradient pulse is detected, not immediately make a final decision on region switching, but first mark the point as a suspect, thereby initiating the subsequent verification procedure. During execution, the latest value of the gradient sequence is compared in real time with the sixth preset gradient threshold. Once the value is detected to exceed the threshold for the first time, the axial position encoder reading of the current sampling needle is stored in the register as a potential boundary point.

[0083] S205. Starting from the potential boundary point, continue polishing forward along the axial direction for the preset verification distance.

[0084] Here, the axial direction refers to the direction along the geometric center line of the sampling needle; the preset verification distance is a short distance travel value pre-set based on materials science knowledge and process experiments. The length is usually set to be greater than the size of typical isolated anomalies in the material (such as hard phase particles), but much smaller than the minimum length of any functional region.

[0085] Specifically, its core function is to proactively perform short-distance reconnaissance polishing. The purpose is to collect surface characteristic data within a short distance after the potential boundary point, thus providing a basis for determining the authenticity of the boundary point. Ignoring the original machining path program, it executes independent motion commands: starting from the marked potential boundary point, it moves forward axially by a distance equal to the preset verification distance, while maintaining the current polishing parameters (such as feed rate and spindle speed). This process must ensure the stability of the machining parameters to ensure that changes in subsequent energy dissipation indicators are solely due to changes in the workpiece surface properties, thereby eliminating interference from changes in control parameters.

[0086] S206. Within the verification distance, monitor the values ​​of the energy dissipation index sequence.

[0087] Specifically, this step occurs synchronously with S205 and is part of the data acquisition process. It is executed throughout the entire verification distance movement. When the verification run of S205 begins, a synchronization signal to start recording is simultaneously sent to the data acquisition system. The data acquisition system begins storing the energy dissipation index value calculated for each sampling period into a dedicated data buffer (such as a first-in-first-out queue (FIFO) or a fixed-size array). When the verification run ends, the controller sends a synchronization signal to stop recording; at this point, the buffer contains the historical record of the energy dissipation index changes with position over the verification distance.

[0088] In some embodiments, to address complex signal patterns encountered during boundary verification that are neither transient anomalies nor stable platforms, the sequence variance of energy dissipation indices can be introduced as a third criterion to identify and employ dedicated low-energy modes to safely process unstable boundaries such as heat-affected zones, thereby avoiding processing damage.

[0089] Specifically, this step aims to solve the more complex and subtle boundary identification problem, upgrading the binary decision logic ("true" or "false") of S207 and S208 to a ternary decision model. In some advanced manufacturing processes (such as laser processing), the boundary of a functional area is not a sharp step, but may be a "heat-affected zone" (HAZ) with extremely heterogeneous physical properties. Due to incomplete phase transitions, the microstructure, hardness, and internal stress in this region are extremely unstable. When the polishing wheel enters this region, the energy dissipation signal is characterized by neither a rapid decline like that of an isolated hard point, nor a stabilization on a new platform like that of entering a new functional area, but rather a continuous, irregular, and violent fluctuation.

[0090] To address this situation, this step performs the following operations in parallel while monitoring the S206 data: Calculate sequence variance: Calculate the sequence variance of the collected energy dissipation index sequence within the verification distance in real time. Variance is a statistic that measures the degree of data dispersion. Low variance means that the signal is stable on a certain plateau (regardless of whether it is new or old), while high variance directly quantifies the "instability" or "severity of fluctuation" of the signal.

[0091] A ternary decision-making mechanism is established: at the end of the verification distance, a more complete decision logic is executed: if the indicator falls back to the baseline (S207), it is determined to be a "false boundary"; if the indicator does not fall back and the sequence variance is lower than the preset unstable variance threshold, it is determined to be a "true boundary" (S208). The unstable variance threshold is set based on statistical analysis of a large amount of experimental data from normal processing areas and known heat-affected zones, and is used to distinguish whether the signal is stable or fluctuates violently; if the indicator does not fall back and the sequence variance exceeds the unstable variance threshold, it is determined that the encountered boundary is a heat-affected zone boundary.

[0092] Once a heat-affected zone is identified, the current processing parameters are temporarily suspended, and the first preset force value and spindle speed are simultaneously reduced to preset low-energy processing parameters. These low-energy processing parameters are a set of safety parameters determined through process experiments, designed to minimize impact and thermal damage to the fragile structures in the heat-affected zone while ensuring the lowest possible material removal rate. Polishing continues using these parameters, with continuous monitoring of the sequence variance of the energy dissipation index, until this variance value falls back below the unstable variance threshold. This indicates that the entire heat-affected zone has been safely traversed, at which point the normal processing control mode is resumed.

[0093] The above steps, by introducing sequence variance, enable the control system to identify unstable states, thereby upgrading boundary identification from a binary judgment to a ternary decision model capable of distinguishing true boundaries, false boundaries, and unstable regions. This allows for the matching of specific low-energy safe processing strategies for material-sensitive and easily damaged areas such as heat-affected zones, avoiding quality defects such as microcracks and microstructural degradation that may result from processing with conventional parameters, and improving the processing yield and reliability of complex workpieces.

[0094] S207. When the value of the energy dissipation index sequence has fallen back below the normal baseline value at the end of the verification distance, the boundary judgment is ignored, and the first preset force value is reduced by the preset compensation value within the subsequent preset compensation distance.

[0095] The normal baseline value is a reference value obtained by averaging or filtering the energy dissipation index data in the stable processing area before the potential boundary point marked by S204, representing the normal or original surface state; the preset compensation distance and preset compensation value are safety correction parameters that are preset by experiments based on material hardness and processing safety, and are used for short-term protective processing after encountering hard points.

[0096] Specifically, the grinding wheel encounters isolated hard points (such as carbide particles in the material) or stress concentration points smaller than the verification distance during its movement. At the moment of contact with this point, the energy dissipation index spikes dramatically, triggering boundary prediction in S204. However, once the grinding wheel passes this point, it returns to the same substrate material, causing the energy dissipation index to drop rapidly. During execution, the end data in the verification data buffer is analyzed. If the value has fallen back to the normal baseline or lower, the trigger is considered a false alarm. The potential boundary point markers are then cleared, the boundary judgment is ignored, and the control flow returns to the previous regular polishing mode. To address any potential impact on the grinding wheel or needle tip, the target force value (first preset force value) is temporarily reduced by a preset compensation value over the next preset compensation distance, handling the recently traversed area more gently—this is an adaptive protection mechanism.

[0097] S208. When the energy dissipation index sequence continues to stabilize at a new plateau value that is different from the normal baseline value at the end of the verification distance, the true boundary of the potential boundary point is confirmed.

[0098] Among them, the new platform value refers to the new stable value range to which the energy dissipation index sequence converges within the verification distance and has a significant statistical difference from the normal baseline value, representing a continuous change in the physical properties of the workpiece surface; the true boundary refers to the actual starting position of the potential boundary point as verified as the functional zoning.

[0099] Specifically, this step is another branch of the decision-making logic. The grinding wheel has indeed moved from one functional area to another, for example, from a smooth needle bar to a rough acoustic window. Because the surface characteristics (such as roughness) of the new area are different from the previous one, the energy dissipation index will rise (or fall) to a new stable level that matches the surface characteristics after crossing the boundary, and will remain at this new level throughout the entire verification distance. During execution, after analyzing the verification data buffer, if it is found that the value has not fallen back, but has stabilized at a new plateau value (which can be judged by calculating the mean and variance of the data in the buffer, i.e., the mean is significantly different from the baseline value, and the variance is less than a certain stable threshold), the potential boundary point marked S204 is determined to be the true boundary. After confirmation, the axial position of the point will be officially recorded, and based on this, the region state switch of the main control program will be triggered, for example, from the cutting edge mode to the transition zone mode, or from the transition zone mode to the acoustic window mode, thereby loading the processing strategy that should be applied to the subsequent functional areas. This "hypothesis-verification" closed loop is thus completed, and a highly robust boundary identification has been performed.

[0100] S209. When the real-time polishing position is in the cutting edge area, the feed speed is adjusted to stabilize the real-time force signal value at the first preset force value. At the same time, based on the real-time change feedback of the kurtosis value, the direction of increase or decrease of the spindle speed is continuously adjusted so that the real-time kurtosis value approaches the local minimum value.

[0101] S210. When the real-time polishing position is in the acoustic window area, the feed speed is adjusted to stabilize the real-time kurtosis value at the first preset kurtosis target value, while monitoring that the peak value of the real-time force signal does not exceed the second preset force value.

[0102] Steps S209-S210 and Figure 1 Steps S104-S105 in the illustrated embodiment are similar and can be found in the descriptions of steps S104-S105, which will not be repeated here.

[0103] In some embodiments, during the processing of the S210 acoustic window area, in order to cope with the sharp drop in kurtosis value caused by encountering a local ultra-smooth area, the main feed control can be temporarily frozen and the spindle speed optimization can be used to actively repair the surface texture, so as to avoid the control system over-compensating and causing processing defects.

[0104] Specifically, the core of S210 is to maintain a high kurtosis target value by adjusting the feed rate. However, if the grinding wheel accidentally passes over a naturally occurring or microscopically smooth area left over from the previous process, the real-time kurtosis value will momentarily and significantly drop below the target value. At this time, the standard PID controller will determine that a huge negative error has occurred and command the feed rate to increase sharply, attempting to compensate for this error by roughening. This excessive compensation behavior may cause scratches, nicks, or irreversible surface burns on the smooth area, thus creating new defects. This step solves this problem by introducing an exception handling procedure.

[0105] First, this type of steep cliff is identified using a dual criterion: The negative rate of change of the real-time kurtosis value exceeds a preset rate of change threshold: this is used to capture the speed at which the kurtosis value decreases. The preset rate of change threshold is a critical rate that can distinguish between normal fluctuations and abnormal drops, obtained from statistical analysis of a large amount of normal processing data.

[0106] The value decreases by more than a preset amplitude threshold within a preset time window: this is used to capture the depth of the kurtosis value's decline. Together with the former, it constitutes a joint judgment on sharp drops, eliminating the interference of transient spikes with insufficient depth caused by a single noise point.

[0107] When both conditions are met, the system is determined to have entered a dangerous pathological chase state. At this point, the following repair logic is executed: First, pausing the adjustment of the feed rate is equivalent to freezing the S210's main controller, forcibly preventing the execution of any erroneous instructions to increase the feed rate, thereby avoiding potential damage to the workpiece.

[0108] Next, the control strategy is temporarily switched, and the direction of increase or decrease of the spindle speed is continuously adjusted based on the real-time feedback of the kurtosis value. The kurtosis optimization logic of the cutting edge region (S104) is borrowed, but with the opposite goal. By fine-tuning the spindle speed (increasing or decreasing) and observing the upward trend of the kurtosis value, the speed at which the surface texture can be roughened again is found.

[0109] Finally, this process continues until the real-time kurtosis value rises back to the first preset kurtosis target value, indicating that the surface texture has been successfully restored. At this point, the automatic resumption of feed rate adjustments returns to the original normal control flow of S210.

[0110] The above steps, by monitoring the dynamic rate of change of kurtosis value, suspend the main control loop and temporarily enable auxiliary control logic for repair, avoid permanent defects on the workpiece surface caused by overcompensation of the PID controller, and transform potential processing failure points into an intelligent online self-repair process.

[0111] S211. When the real-time polishing position is in the transition zone, the feed speed and spindle speed are adjusted synchronously according to the position weight coefficient so that the real-time force signal value and the real-time kurtosis value track the dynamic target force value and the dynamic target kurtosis value respectively.

[0112] Step S211 and Figure 1 Step S106 in the illustrated embodiment is similar and can be found in the description of step S106, which will not be repeated here.

[0113] In the above embodiments, potential boundary points are initially marked by exceeding a threshold energy dissipation gradient instead of direct judgment. A short verification distance is actively polished and the signal sequence during this period is monitored. This verification mechanism can distinguish between two situations: if the signal falls back to the baseline value, it is determined to be an instantaneous pseudo-boundary caused by material internal stress release and is ignored; if the signal stabilizes at the new plateau value, it is confirmed as the true functional area boundary. Upgrading the threshold trigger to a hypothesis-verification closed-loop decision logic improves the robustness of boundary identification, can cope with batch differences in workpieces or occasional anomalies within the material, and ensures the accuracy of functional area division.

[0114] The following describes an exemplary adaptive polishing system 300 for the outer wall of a sampling needle provided in an embodiment of this application. Figure 3 This is an exemplary hardware structure diagram of the adaptive polishing system 300 for the outer wall of the sampling needle provided in this application embodiment.

[0115] In some embodiments, the adaptive polishing system 300 for the outer wall of the sampling needle is a computer device or includes a computer device in the adaptive polishing system 300 for the outer wall of the sampling needle. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores data. The network interface of the computer device is used to communicate with other external terminals or servers via a network connection. In some embodiments, the network interface can be a wired network interface; in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, it implements the methods in the embodiments of this application.

[0116] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0117] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application 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 scope of the technical solutions of the embodiments of this application.

[0118] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0119] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0120] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. An adaptive polishing method for the outer wall of a sampling needle, characterized in that, include: Collect force and vibration signals generated at the contact point between the grinding wheel and the sampling needle during the polishing process; The kurtosis value of the real-time time series data of the vibration signal is calculated, and the kurtosis value and the real-time value of the force signal are weighted and summed according to a preset weight to obtain the energy dissipation index sequence; the kurtosis value is the ratio of the fourth central moment to the square of the variance of the time series data points of the vibration signal. The functional region to which the real-time polishing position belongs is determined based on the energy dissipation gradient sequence; the energy dissipation gradient sequence is obtained by calculating the rate of change of the energy dissipation index sequence along the axial position of the sampling needle; the functional region includes the cutting edge region, the acoustic window region, and the transition region. When the real-time polishing position is in the cutting edge area, the feed speed is adjusted to stabilize the real-time force signal value at the first preset force value. At the same time, based on the real-time change feedback of the kurtosis value, the direction of increase or decrease of the spindle speed is continuously adjusted so that the real-time kurtosis value approaches the local minimum value. When the real-time polishing position is in the acoustic window area, the feed speed is adjusted to stabilize the real-time kurtosis value at the first preset kurtosis target value, while monitoring that the peak value of the real-time force signal does not exceed the second preset force value. When the real-time polishing position is in the transition zone, the feed rate and spindle speed are adjusted synchronously according to the position weighting coefficient, so that the real-time force signal value and the real-time kurtosis value track the dynamic target force value and the dynamic target kurtosis value respectively; the position weighting coefficient is calculated based on the relative position of the current axial position in the transition zone; the dynamic target force value and the dynamic target kurtosis value are calculated by linear interpolation of the first preset force value and the first preset kurtosis target value using the position weighting coefficient.

2. The method according to claim 1, characterized in that, When the real-time polishing position is in the cutting edge area, the feed speed is adjusted to stabilize the real-time force signal value at a first preset force value. Simultaneously, based on the real-time change feedback of the kurtosis value, the direction of increase or decrease of the spindle speed is continuously adjusted so that the real-time kurtosis value approaches a local minimum. The process further includes: When the real-time kurtosis value is lower than the third preset kurtosis threshold and the real-time force signal is higher than the third preset force value, the control of adjusting the feed speed and adjusting the direction of increase or decrease of the spindle speed is suspended, a preset high-frequency sinusoidal disturbance is applied to the spindle speed, and the real-time disturbance force signal value is monitored. When the real-time disturbance force signal value drops from the third preset force value to the fourth preset force value, the application of the high-frequency sinusoidal disturbance is stopped, and the control of adjusting the feed speed and adjusting the direction of increase or decrease of the spindle speed is resumed.

3. The method according to claim 2, characterized in that, After applying a preset high-frequency sinusoidal disturbance to the spindle speed, the method further includes: Calculate the root mean square value of the vibration signal; When the root mean square value exceeds the fifth preset threshold, the amplitude of the high-frequency sinusoidal disturbance remains unchanged, while the fixed preset frequency is switched to a dynamic frequency that is randomly modulated within the preset frequency bandwidth; the fifth preset threshold is used to characterize the workpiece resonance. The application of the dynamic frequency is stopped when the real-time disturbance force signal value drops from the third preset force value to the fourth preset force value.

4. The method according to claim 1, characterized in that, The step of determining the functional region to which the real-time polishing position belongs based on the energy dissipation gradient sequence specifically includes: When the value of the energy dissipation gradient sequence exceeds the sixth preset gradient threshold for the first time, the current position is marked as a potential boundary point; the sixth preset gradient threshold is used to identify potential boundaries. Starting from the potential boundary point, continue polishing forward along the axial direction for a predetermined verification distance; Within the verification distance, monitor the values ​​of the energy dissipation index sequence; When the verification distance ends and the value of the energy dissipation index sequence has fallen below the normal baseline value, the boundary judgment is ignored, and the first preset force value is reduced by a preset compensation value within the subsequent preset compensation distance; the normal baseline value is the average data before reaching the potential boundary point. When the verification distance ends, if the value of the energy dissipation index sequence remains stable at a new plateau value different from the normal baseline value, then the true boundary of the potential boundary point is confirmed.

5. The method according to claim 4, characterized in that, Within the verification distance, after monitoring the values ​​of the energy dissipation index sequence, the method further includes: Calculate the sequence variance of the energy dissipation index sequence; When the verification distance ends, if the value of the energy dissipation index sequence does not fall below the normal baseline value and the sequence variance exceeds the preset unstable variance threshold, the potential boundary point is determined to be the boundary of the heat-affected zone, and the first preset force value and the spindle speed are reduced to preset low-energy processing parameters until the sequence variance falls below the unstable variance threshold.

6. The method according to claim 1, characterized in that, When the real-time polishing position is within the acoustic window region, after adjusting the feed speed to stabilize the real-time kurtosis value at a first preset kurtosis target value, and simultaneously monitoring that the peak value of the real-time force signal does not exceed a second preset force value, the method further includes: When the negative rate of change of the real-time kurtosis value exceeds a preset rate of change threshold and the decrease in value within a preset time window exceeds a preset amplitude threshold, the adjustment of the feed speed is paused, and the direction of increase or decrease of the spindle speed is continuously adjusted based on the real-time change feedback of the kurtosis value. The adjustment of the feed rate will resume once the real-time kurtosis value rises back to the first preset kurtosis target value.

7. The method according to claim 1, characterized in that, When the real-time polishing position is in the transition zone, after synchronously adjusting the feed rate and spindle speed according to the position weighting coefficient so that the real-time force signal value and real-time kurtosis value respectively track the dynamic target force value and dynamic target kurtosis value, the method further includes: When both the force deviation and kurtosis deviation exceed their respective preset error limits, the interpolation control logic of the transition zone is terminated; the force deviation is the deviation between the real-time force signal value and the dynamic target force value, and the kurtosis deviation is the deviation between the real-time kurtosis value and the dynamic target kurtosis value. The first preset kurtosis target value and the second preset force value of the sound window area are used as the current target value.

8. An adaptive polishing system for the outer wall of a sampling needle, characterized in that, The adaptive polishing system for the outer wall of the sampling needle includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the adaptive polishing system for the outer wall of the sampling needle to perform the method as described in any one of claims 1-7.

9. A computer program product containing instructions, characterized in that, When the computer program product is run on the adaptive polishing system of the outer wall of the sampling needle, the adaptive polishing system of the outer wall of the sampling needle performs the method as described in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instruction is executed on the adaptive polishing system of the outer wall of the sampling needle, the adaptive polishing system of the outer wall of the sampling needle performs the method as described in any one of claims 1-7.

Images