Workpiece polishing system, method, and storage medium based on six-axis force sensor

Abstract

The application provides a workpiece polishing system and method based on a six-axis force sensor, and a storage medium, comprising: determining the mechanical response data of a current polishing area; determining the surface roughness of the current polishing area; determining the target polishing parameters of the current polishing area based on the workpiece type, the curvature of the current polishing area, the surface roughness and the target roughness; controlling the industrial robot to perform polishing work based on the target polishing parameters; and cyclically executing the above steps, and in response to the difference between the surface roughness of the current polishing area and the target roughness being less than a difference threshold, controlling the industrial robot to go to the next polishing area to perform polishing work. The application can realize intelligent feedback and self-adaptive adjustment of the polishing process, automatically adjust the polishing parameters for different workpiece areas, effectively solve the problems of excessive polishing or insufficient polishing that are prone to occur when the traditional constant force polishing faces the uneven material hardness or the sharp change of the curvature of the curved surface, and improve the polishing quality and efficiency.

Description

Technical Field

[0001] This specification relates to the field of industrial grinding, and in particular to a workpiece grinding system, method and storage medium based on a six-axis force sensor. Background Technology

[0002] In high-end equipment fields such as aerospace, automobile manufacturing, and precision machinery, the surface grinding quality of complex curved workpieces such as impellers and hubs directly determines the product's operating accuracy, service life, and reliability.

[0003] Traditional workpiece grinding methods have many drawbacks. On the one hand, they often use fixed, preset grinding parameters, which cannot be dynamically adjusted according to changes in the curvature and surface roughness of the workpiece area. This can easily lead to over-grinding or under-grinding when dealing with complex curved surfaces. On the other hand, traditional workpiece grinding methods lack real-time identification and adaptive adjustment capabilities when faced with abnormal working conditions during the grinding process (such as grinding head wear, workpiece hard spots, and stress overload). This can easily lead to workpiece scrapping or equipment damage, making it difficult to cope with diverse workpieces and complex processing scenarios.

[0004] Therefore, it is necessary to provide a workpiece grinding system, method, and storage medium based on a six-axis force sensor to achieve real-time feedback and adaptive control of the grinding process, thereby improving grinding quality and efficiency. Summary of the Invention

[0005] The invention includes a workpiece grinding system based on a six-axis force sensor. The workpiece grinding system includes a management platform and a sensing and control platform. The sensing and control platform includes a six-axis force sensor and an industrial robot. The management platform is configured to: determine the mechanical response data of the current grinding area based on the six-axis force sensor; determine the surface roughness of the current grinding area based on the workpiece type, current grinding parameters, and the mechanical response data; and determine target grinding parameters for the current grinding area using a first model based on the workpiece type, the curvature of the current grinding area, the surface roughness, and the target roughness. The target grinding parameters include target grinding pressure, target feed speed, and target grinding angle. The first model is a machine learning model; the industrial robot is controlled to perform a polishing operation based on the target polishing parameters, and the above steps are executed repeatedly. In response to the difference between the surface roughness of the current polishing area and the target roughness being less than a difference threshold, the industrial robot is controlled to move to the next polishing area to perform the polishing operation. The polishing operation includes: adjusting the joint torque of the industrial robot to apply the target polishing pressure; adjusting the drive power of the industrial robot to match the target feed speed; and adjusting the joint angle of the industrial robot to align with the target polishing angle.

[0006] The invention includes a workpiece grinding method based on a six-axis force sensor. The workpiece grinding method is executed by a management platform of a workpiece grinding system. The method includes: determining the mechanical response data of the current grinding area based on the six-axis force sensor; determining the surface roughness of the current grinding area based on the workpiece type, current grinding parameters, and the mechanical response data; and determining target grinding parameters for the current grinding area using a first model based on the workpiece type, the curvature of the current grinding area, the surface roughness, and the target roughness. The target grinding parameters include target grinding pressure, target feed rate, and target grinding angle. The first model is a machine learning model; the industrial robot is controlled to perform a grinding operation based on the target grinding parameters, and the above steps are executed repeatedly. In response to the difference between the surface roughness of the current grinding area and the target roughness being less than a difference threshold, the industrial robot is controlled to move to the next grinding area to perform the grinding operation. The grinding operation includes: adjusting the joint torque of the industrial robot to apply the target grinding pressure; adjusting the drive power of the industrial robot to match the target feed speed; and adjusting the joint angle of the industrial robot to align with the target grinding angle.

[0007] The invention includes a computer-readable storage medium that stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes a workpiece grinding method.

[0008] The beneficial effects of the invention include, but are not limited to: ① Providing a workpiece grinding system and method based on a six-axis force sensor, which enables intelligent feedback and adaptive adjustment of the grinding process. It automatically adjusts grinding parameters for different workpiece areas, effectively solving the problems of over-grinding or under-grinding that easily occur with traditional constant-force grinding when facing uneven material hardness or drastic changes in surface curvature, thus improving grinding quality and efficiency. ② By using a grinding head made of polyurethane and internally filled with magnetorheological fluid and an integrated electromagnetic coil, the stiffness of the grinding head can be flexibly switched to adapt to different grinding areas, significantly improving grinding efficiency and quality. ③ Gradual adjustment of grinding parameters based on adjustment steps ensures the stability of grinding parameter adjustment, effectively avoiding negative effects such as grinding chatter caused by grinding pressure instability; adjusting the coolant spray flow rate based on target grinding parameters achieves precise temperature control during the grinding process, preventing thermal deformation of the workpiece surface, ensuring processing accuracy, and extending the service life of the grinding head. Attached Figure Description

[0009] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein:

[0010] Figure 1 This is a platform schematic diagram of a workpiece grinding system based on a six-axis force sensor, as shown in some embodiments of this specification. Figure 2 This is an exemplary flowchart of a workpiece grinding method based on a six-axis force sensor, as shown in some embodiments of this specification. Figure 3 These are exemplary schematic diagrams of the first model shown in some embodiments of this specification; Figure 4 These are exemplary schematic diagrams of a second model based on some embodiments shown in this specification; Figure 5 This is an exemplary schematic diagram illustrating the adjustment of coolant injection flow rate based on some embodiments shown in this specification. Detailed Implementation

[0011] To more clearly illustrate the technical solutions of the embodiments in this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this specification. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0012] It should be understood that the terms “system,” “device,” “unit,” and / or “module” used herein are one way to distinguish different components, elements, parts, sections, or assemblies at different levels. However, if other terms can achieve the same purpose, they may be replaced by other expressions.

[0013] As indicated in this specification and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of expressly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0014] Flowcharts are used in this specification to illustrate the operations performed by the system based on embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0015] Figure 1This is a platform schematic diagram of a workpiece grinding system based on a six-axis force sensor, as shown in some embodiments of this specification. The workpiece grinding system 100 is an automated system for performing grinding operations on workpieces.

[0016] In some embodiments, such as Figure 1 As shown, the workpiece grinding system 100 based on a six-axis force sensor may include a user platform 110, a service platform 120, a management platform 130, a sensor network platform 140, and a sensing and control platform 150 that are connected in sequence.

[0017] User platform 110 is an interactive platform for process engineers, maintenance personnel, production schedulers, and other relevant personnel. In some embodiments, user platform 110 may include at least one terminal device, such as a field operation terminal, a remote monitoring interface, or an application program.

[0018] Service platform 120 is a platform for transmitting service requests. In some embodiments, service platform 120 can be configured as a server with communication capabilities to interact with user platform 110 and management platform 130. For example, service platform 120 obtains data such as grinding task requirements, workpiece CAD models, and target grinding parameters uploaded by relevant personnel from user platform 110 and sends them to management platform 130; service platform 120 can also obtain data such as current grinding progress, current surface roughness, and industrial robot operating status parameters (e.g., whether there are faults, anomalies, etc.) from management platform 130 and distribute them to user platform 110 to ensure that process engineers can monitor and adjust grinding parameters and that maintenance personnel can handle abnormal situations in a timely manner.

[0019] The management platform 130 is a control unit used to receive sensor data, perform data analysis and decision-making, and control industrial robots to perform grinding operations.

[0020] In some embodiments, the management platform 130 may be integrated into a processor. The processor is configured to coordinate and manage information transmission and collaboration between platforms and to provide control and management functions for the workpiece grinding system 100. In some embodiments, the processor may include, but is not limited to, a central processing unit (CPU), a graphics processing unit (GPU), a field-programmable gate array (FPGA), etc.

[0021] In some embodiments, the management platform 130 is configured to: determine the mechanical response data of the current grinding area based on a six-axis force sensor; determine the surface roughness of the current grinding area based on the workpiece type, current grinding parameters, and mechanical response data; determine the target grinding parameters of the current grinding area based on the workpiece type, the curvature of the current grinding area, the surface roughness, and the target roughness through a first model, the target grinding parameters including the target grinding pressure, the target feed speed, and the target grinding angle, the first model being a machine learning model; control the industrial robot to perform grinding operations based on the target grinding parameters; and repeatedly execute the above steps, responding to a situation where the difference between the surface roughness of the current grinding area and the target roughness is less than a difference threshold, controlling the industrial robot to move to the next grinding area for grinding operations, wherein the grinding operations include: adjusting the joint torque of the industrial robot to apply the target grinding pressure; adjusting the drive power of the industrial robot to match the target feed speed; and adjusting the joint angle of the industrial robot to align with the target grinding angle.

[0022] The sensor network platform 140 is a platform for sensor communication. In some embodiments, the sensor network platform 140 can be configured as an industrial Ethernet device, a 5G communication module, an edge computing gateway, etc., to interact with the perception and control platform 150 and the management platform 130. For example, the sensor network platform 140 obtains mechanical response data collected by the six-axis force sensor, joint torque, movement speed, joint angle, and other operating status parameters of the industrial robot from the perception and control platform 150, and uploads them to the management platform 130 in real time; the sensor network platform 140 can also obtain target grinding parameters, coolant injection flow rate, and other data from the management platform 130, and send them to the perception and control platform 150 to ensure the efficient and stable operation of the grinding system.

[0023] The sensing and control platform 150 is a hardware execution platform that directly performs the grinding operation. In some embodiments, the sensing and control platform 150 includes a six-axis force sensor 151 and an industrial robot 152.

[0024] A six-axis force sensor is a force sensor that collects and measures mechanical response data. Compared to traditional single-axis force sensors, a six-axis force sensor can simultaneously detect forces or torques in multiple dimensions.

[0025] In some embodiments, the two ends of the six-axis force sensor are connected to the grinding head and the industrial robot, respectively, to detect the dynamic mechanical parameters of the interface between the grinding head and the workpiece during the grinding process.

[0026] Industrial robot 152 is an automated mechanical device for grinding operations, including a robotic arm, a grinding head installed at the end of the robotic arm, and a matching drive mechanism.

[0027] A grinding head is a tool component that directly contacts the workpiece during grinding operations and performs cutting, grinding, or polishing functions. The material type of the grinding head can be determined based on the actual application scenario and requirements.

[0028] In some embodiments, the industrial robot 152 receives parameters such as joint torque, movement speed, and joint rotation angle from the management platform 130 to perform grinding operations. The industrial robot 152 can also feed back data such as current grinding pressure, current feed speed, current joint angle, and current rotation speed of the grinding head to the management platform 130.

[0029] In some embodiments of this specification, the workpiece grinding system 100 based on a six-axis force sensor can form an information operation closed loop between various functional platforms and operate in a coordinated and regular manner under the scheduling of the management platform 130, thereby realizing the intelligent, precise and efficient workpiece grinding process.

[0030] For more information on the functions of the aforementioned platforms, please refer to [link / reference]. Figure 2-5 And its related descriptions.

[0031] It should be noted that the above description of the workpiece grinding system 100 and its platform is for ease of description only and should not be construed as limiting this specification to the scope of the illustrated embodiments. It is understood that those skilled in the art, after understanding the principle of the system, may arbitrarily combine the various platforms or construct subsystems connected to other platforms without departing from this principle. In some embodiments, Figure 1 The user platform 110, service platform 120, management platform 130, sensor network platform 140, and sensing control platform 150 disclosed herein can be different platforms within a single system, or a single platform can implement the functions of two or more of the aforementioned platforms. For example, each platform can share a storage module, or each module can have its own storage module. Such variations are all within the scope of protection of this specification.

[0032] Figure 2 This is an exemplary flowchart of a workpiece grinding method based on a six-axis force sensor, as shown in some embodiments of this specification. Figure 2 As shown, the workpiece grinding method includes S1-S5. In some embodiments, S1-S5 can be executed by a processor.

[0033] S1, based on a six-axis force sensor, determines the mechanical response data of the current grinding area.

[0034] The current grinding area is the local area on the surface of the workpiece that needs to be ground. The current grinding area typically includes defects such as parting lines, welds, and burrs. For example, the current grinding area may include the spoke sidewall of a wheel hub or the blade profile of an impeller.

[0035] In some embodiments, the processor can obtain the current grinding area uploaded by the process engineer from the user platform 110. The processor can also identify the workpiece photo using image recognition algorithms (e.g., U-Net algorithm, YOLO detection algorithm) and determine the area with defects on the workpiece surface as the grinding area.

[0036] Mechanical response data reflects the mechanical forces acting between the grinding head and the workpiece. In some embodiments, the mechanical response data includes linear forces in three degrees of freedom (…). , , ) and torques of the three rotational degrees of freedom ( , , The X and Y axes correspond to the two orthogonal tangential directions of the grinding head along the workpiece surface, respectively. The Z axis is perpendicular to the plane of the X and Y axes and parallel to the normal direction of the workpiece surface in the current grinding area. The processor can acquire mechanical response data through the six-axis force sensor 151. For more information about the six-axis force sensor 151, see [link to documentation]. Figure 1 part.

[0037] S2 determines the surface roughness of the current grinding area based on the workpiece type, current grinding parameters, and mechanical response data.

[0038] The workpiece type is used to reflect the type of workpiece to be ground. In some embodiments, workpiece types include impellers, hubs, gears, supports, etc. The processor can determine the workpiece type using image recognition algorithms (e.g., U-Net algorithm, YOLO detection algorithm).

[0039] Grinding parameters are a set of parameters that control the grinding head to perform grinding operations. Current grinding parameters refer to the grinding parameters used at the current moment.

[0040] In some embodiments, the grinding parameters include grinding pressure, feed rate, and grinding angle.

[0041] Grinding pressure is the force applied by the grinding head along the normal direction of the contact point. Grinding pressure can be expressed as a mechanical response data point. Feed rate is the speed at which the grinding head moves relative to the surface of the workpiece being ground during the grinding process. The feed rate is directly related to the current circumference of the grinding head and its current rotational speed. A faster feed rate results in higher grinding efficiency and greater material removal per unit time. Grinding angle refers to the angle between the grinding head's axis of rotation and the normal to the current grinding area. Different grinding angles affect the contact method between the grinding head and the workpiece (e.g., spot grinding, side grinding, or surface grinding).

[0042] In some embodiments, the processor can determine the grinding pressure based on a six-axis force sensor. The feed rate is determined based on the tachometer at the grinding head and the outer circumference of the grinding head; the grinding angle is determined based on the gyroscope or angle sensor at the grinding head.

[0043] Surface roughness is used to characterize the degree of microscopic undulations on the surface of a workpiece. In some embodiments, surface roughness is expressed as arithmetic mean roughness (Ra), where a lower Ra value indicates a smoother workpiece surface.

[0044] In some embodiments, the processor can determine the surface roughness of the current grinding area in a variety of ways. For example, the processor can consult a pre-built first preset table and obtain the surface roughness of the current grinding area through similarity matching.

[0045] The first preset table includes the correspondence between historical workpiece types, historical grinding parameters, historical mechanical response data, and historical surface roughness. The processor collects a large number of historical grinding records to construct the first preset table. Each historical grinding record includes historical workpiece type, historical grinding parameters, historical mechanical response data, and historical surface roughness measured by a roughness measuring instrument after the historical grinding is completed.

[0046] When querying the first preset table, the processor uses the current workpiece type, real-time grinding parameters, and mechanical response data as matching criteria. It retrieves the historical grinding record with the highest similarity from the first preset table and determines the historical surface roughness corresponding to that historical grinding record as the surface roughness of the current grinding area. The processor can determine the similarity using algorithms such as cosine similarity.

[0047] In some embodiments, the processor can also determine the surface roughness of the current grinding area using a second model. For more information on the second model, see [link to relevant documentation]. Figure 4 And related explanations.

[0048] S3, based on the workpiece type, the curvature of the current grinding area, the surface roughness, and the target roughness, determines the target grinding parameters of the current grinding area through the first model.

[0049] Curvature is used to measure the degree of curvature of the current grinding area of ​​the workpiece. The greater the curvature, the more severe the curvature of the current grinding area, and the more difficult it is to fit the grinding head to the surface. The processor can read the workpiece CAD model and use methods such as Weingarten mapping and Mesh Laplacian method to perform surface fitting and curvature calculation on the current grinding area to determine the curvature.

[0050] The target roughness is the surface roughness that needs to be achieved during the workpiece grinding process. In some embodiments, the target roughness can be preset and determined according to the actual application scenario and requirements. Different grinding tasks (e.g., mirror polishing tasks, deburring tasks) can preset different target roughnesses.

[0051] The target grinding parameters are the grinding parameters used to grind the current grinding area to the target roughness. Since the surface roughness of the current grinding area and the wear of the grinding head are dynamic during the grinding process, the processor needs to dynamically adjust the current grinding parameters to ensure that the target grinding parameters balance grinding efficiency and grinding quality.

[0052] In some embodiments, the target grinding parameters include the target grinding pressure, the target feed rate, and the target grinding angle. For more information on grinding pressure, feed rate, and grinding angle, please refer to the section on grinding parameters above.

[0053] Figure 3 This is an exemplary schematic diagram of a first model based on some embodiments of this specification.

[0054] The first model 350 is a predictive model for predicting the grinding parameters of the target. In some embodiments, the first model can be a machine learning model, such as any one or a combination of neural network (NN) models or other custom model structures.

[0055] In some embodiments, the inputs to the first model 350 include workpiece type 320, curvature 332 of the current grinding area, surface roughness 331, and target roughness 333; the output includes target grinding parameters 360 for the current grinding area. For more information on workpiece type and surface roughness, see S2 and its related description.

[0056] In some embodiments, the inputs to the first model further include the rate of decrease of surface roughness 341 and the degree of wear of the grinding head 342.

[0057] The surface roughness reduction rate 341 is the amount of surface roughness reduction per unit time. The faster the reduction rate, the higher the material removal efficiency of the current grinding process.

[0058] The surface roughness reduction rate 341 can be used to determine whether the current grinding process is operating normally. If the reduction rate is too low, it indicates that there may be hard spots in the current grinding area, or that the grinding head is dull, resulting in insufficient cutting force; if the reduction rate is too fast, it indicates that there may be a risk of over-grinding, and the grinding parameters need to be adjusted in time to avoid damaging the workpiece surface.

[0059] In some embodiments, the processor can continuously detect and acquire surface roughness, and plot a curve showing the gradual decrease of surface roughness over time. The instantaneous slope of the curve at a certain moment represents the rate of decrease of surface roughness at that moment. The processor can use the rate of decrease within a preset time range as input to a first model.

[0060] The wear rating of a grinding head (342) measures the degree of lifespan reduction caused by prolonged use. A higher wear rating indicates more severe material wear and a shorter remaining lifespan. The wear rating can be expressed as the remaining thickness of the grinding head.

[0061] The wear level of the grinding head (342) can be used to judge the processing stability and parameter suitability of the current grinding process. If the wear level is too high, it indicates that the cutting ability of the grinding head may have decreased, resulting in unstable grinding quality. In this case, the grinding head should be replaced or the grinding parameters adjusted in time to compensate for the loss of cutting ability caused by wear.

[0062] In some embodiments, the processor can measure the remaining thickness of the grinding head using an industrial camera or laser rangefinder during grinding operations to determine the degree of wear of the grinding head 342.

[0063] Some embodiments in this specification, by further determining the rate of decrease in surface roughness and the wear degree of the grinding head as inputs to the first model, help the workpiece grinding system to promptly identify abnormalities such as hard spots, over-grinding, and grinding head dulling during the grinding process, thereby optimizing grinding efficiency, extending tool life, and ensuring the stability and consistency of workpiece surface quality.

[0064] In some embodiments, the processor can train a first model based on a large number of first training samples with first labels. The first training samples with first labels can be constructed based on historical grinding records. For example, the processor can acquire multiple sets of historical grinding records and filter them, retaining only the preferred grinding records where the grinding effect is good and the industrial robot operates normally. The processor determines the historical workpiece type, historical curvature, historical surface roughness, historical target roughness, historical descent rate, and historical wear degree of the preferred grinding records as the first training samples, and determines the historical grinding parameters of the preferred grinding records as the first labels. Here, "good grinding effect" means that after the historical grinding process is completed, the difference between the surface roughness of the workpiece and the historical target roughness is less than a preset difference threshold; "normal operation of the industrial robot" means that during the historical grinding process, the industrial robot does not experience overload, error, vibration, or other abnormalities, and always operates normally and stably.

[0065] In some embodiments, the training process of the first model includes: constructing a first training sample set from the multiple first training samples with first labels, and performing multiple rounds of iteration based on the first training sample set. At least one round of iteration includes: selecting one or more first training samples from the training dataset, inputting them into an initial first model to obtain the corresponding model prediction output; substituting the model prediction output and the first label corresponding to the first training sample into a predefined loss function formula to calculate the value of the loss function; iteratively updating the model parameters in the initial first model based on methods such as gradient descent, according to the value of the loss function, until the loss function converges or the number of iterations reaches a preset iteration threshold, etc., to terminate the iteration and obtain the trained first model.

[0066] S4 controls the industrial robot to perform grinding operations based on the target grinding parameters.

[0067] Grinding is a process in which the processor controls the movement and force of an industrial robot to grind, polish, and perform other processing on the surface of a workpiece.

[0068] In some embodiments, the polishing operation includes: adjusting the joint torque of the industrial robot to apply a target polishing pressure; adjusting the drive power of the industrial robot to match the target feed speed; and adjusting the joint angle of the industrial robot to align with the target polishing angle.

[0069] Joint torque refers to the torsional torque output by the drive motor at each rotating joint of the robotic arm of an industrial robot. The greater the joint torque, the greater the pressure that the robotic arm applies to the workpiece through the grinding head, and the greater the target grinding pressure.

[0070] Drive power refers to the output power of the power unit of an industrial robot, which provides power to rotate the grinding head. The greater the drive power, the greater the power the industrial robot provides to the grinding head, and the faster the grinding head rotates.

[0071] Joint angle refers to the bending angle of the robotic arm of an industrial robot. By changing the bending angle of the robotic arm, the contact method and contact area between the grinding head and the current grinding area can be changed. For example, when the axis of the grinding head is perpendicular to the current grinding area, the contact area between the grinding head and the current grinding area is minimal, which is suitable for grinding scattered abrasive protrusions; when the axis of the grinding head is parallel to the current grinding area, the contact area between the grinding head and the current grinding area is maximized, which is suitable for grinding large areas of rough surfaces.

[0072] In some embodiments, the processor can control the industrial robot to perform grinding operations in multiple ways. When the industrial robot supports direct input of target grinding parameters, the processor can directly send the target grinding parameters to the industrial robot, and the industrial robot automatically adjusts the joint torque, drive power, and joint angle through its built-in proportional-integral-derivative (PID) controller. When the industrial robot does not support direct input of target grinding parameters, the processor can determine the direction of fine-tuning (decrease or increase) of the joint torque, drive power, and joint angle based on the relationship between the current grinding parameters and the target grinding parameters.

[0073] The processor, based on the fine-tuning direction, successively corrects the current grinding parameters with a preset step size until the current grinding parameters converge to a preset error range allowed by the target grinding parameters, in order to perform the grinding operation. For example, when the current feed rate in the current grinding parameters is less than the target feed rate, the processor can determine the fine-tuning direction of the drive power as increasing it, and gradually increase the drive power based on a preset step size while monitoring the current feed rate in real time. In response to the current feed rate converging to a preset error range allowed by the target feed rate, the adjustment of the target feed rate is completed.

[0074] The processor can adjust the joint torque of the industrial robot in accordance with the above adjustment method, thereby adjusting the current grinding pressure to the target grinding pressure; and adjust the joint angle of the industrial robot, thereby adjusting the current grinding angle to the target grinding angle.

[0075] S5, repeat S1-S4 in a loop. In response to the difference between the surface roughness of the current grinding area and the target roughness being less than the difference threshold, control the industrial robot to move to the next grinding area to perform grinding operations.

[0076] The difference threshold is the condition for determining whether the current grinding area is finished. It is known that there are two cases where the difference between the surface roughness of the current grinding area and the target roughness is less than the difference threshold: the difference is less than the difference threshold, and the difference is not less than the difference threshold. When the difference is less than the difference threshold, the processor can consider the current grinding area finished. At this time, the processor can control the industrial robot's robotic arm to move the grinding head at the end of the robotic arm to the next grinding area, repeating the grinding operation from S1 to S4.

[0077] In some embodiments, the difference threshold can be preset and determined according to the actual application scenario and requirements.

[0078] This specification provides some embodiments of a workpiece grinding system and method based on a six-axis force sensor, which can realize intelligent feedback and adaptive adjustment of the grinding process, automatically adjust grinding parameters for different workpiece areas, effectively solve the problem of over-grinding or under-grinding that is prone to occur when traditional constant force grinding is faced with uneven material hardness or drastic changes in surface curvature, and improve grinding quality and efficiency.

[0079] Figure 4 This is an exemplary schematic diagram of the second model shown in some embodiments of this specification. In some embodiments, the processor determines vibration features 410 based on mechanical response data 322, the vibration features 410 including frictional vibration features and contact disturbance features; based on workpiece type 320, current grinding parameters 321 and vibration features 410, the processor determines the surface roughness 331 of the current grinding area through the second model 430, the second model 430 being a machine learning model.

[0080] Vibration characteristic 410 refers to the signal fluctuation characteristics extracted from mechanical response data, used to reflect specific patterns or laws of mechanical vibration signals in the time or frequency domain. Vibration characteristics may include friction vibration characteristics and contact disturbance characteristics. For an introduction to mechanical response data, please refer to S1 above and its related explanations.

[0081] Frictional vibration characteristics are used to characterize the tangential force fluctuations when the movement of the grinding head is obstructed. The higher the surface roughness of the workpiece, the more intense the resistance fluctuations the grinding head experiences when moving in the tangential direction, and the greater the frictional vibration characteristics. The tangential direction refers to the direction in which the grinding head moves across the horizontal plane of the current grinding area.

[0082] In some embodiments, the processor can base the linear force on a single time t. and Determine the tangential resultant force Tangential resultant force Linear force at a single time t and The vector composition result. Tangential resultant force. Positively correlated with linear force .

[0083] In some embodiments, the tangential resultant force at a single time t It can be represented as: = .

[0084] In some embodiments, the processor can determine the preset time window in the manner described above. Tangential resultant force within Then to A high-pass filter is applied to remove low-frequency components, and then statistics are collected within a preset time window. The root mean square (RMS) value is used as the characteristic value of frictional vibration within a preset time window. The cutoff frequency of the high-pass filter can be preset based on prior experience.

[0085] Contact disturbance characteristics are used to characterize the torque fluctuations of a grinding head under eccentric force. When the grinding head moves tangentially, upon contacting protruding or recessed defects (e.g., burrs, edges), it instantaneously experiences an additional reaction force, disrupting the original force equilibrium and causing eccentric force on the grinding head, thus triggering abnormal fluctuations in the resultant torque. The greater the number of defects and the greater the degree of protrusion or recess, the more intense the reaction force and the stronger the contact disturbance characteristics.

[0086] In some embodiments, the processor can base its analysis on the torque in the mechanical response data at a single time t. , ,and Determine the resultant torque Resultant torque Torque at a single moment t , ,and The vector synthesis result. The resultant torque. Torque in the mechanical response data positively correlated with a single time t , ,and .

[0087] In some embodiments, the resultant torque at a single time t It can be represented as: = .

[0088] In some embodiments, the processor can determine the preset time window in the manner described above. within Then to A high-pass filter is applied to remove low-frequency components, and then statistics are collected within a preset time window. The root mean square (RMS) value is used as the contact perturbation characteristic for a preset time window. The cutoff frequency of the high-pass filter can be preset based on prior experience.

[0089] The second model 430 is a predictive model for predicting surface roughness. In some embodiments, the second model can be a machine learning model, such as any one or a combination of neural network (NN) models or other custom model structures.

[0090] In some embodiments, the inputs to the second model 430 include workpiece type 320, current grinding parameters 321, and vibration characteristics 410, and the output includes the surface roughness 331 of the current grinding area. For more information on workpiece type 320 and surface roughness 331, see S2 and its related description.

[0091] In some embodiments, the input to the second model 430 further includes frequency domain features 420, whereby the processor determines a power spectral density function based on the mechanical response data; and determines frequency domain features based on the power spectral density function, the frequency domain features including the spectral centroid, spectral entropy, and band energy ratio.

[0092] The power spectral density function is a function that reflects the power distribution of mechanical response data at different frequencies. The horizontal axis of the power spectral density function is frequency (Hz), and the vertical axis is the power per unit frequency bandwidth (N² / Hz).

[0093] In some embodiments, the processor determines a preset time window based on mechanical response data. Tangential resultant force within Then, a high-pass filter is applied to remove low-frequency components, followed by a Fast Fourier Transform (FFT). The squared amplitude of the FFT result is then calculated and normalized to obtain the preset time window. The power spectral density function within. The cutoff frequency of a high-pass filter can be preset based on prior experience. Regarding the determination of the tangential resultant force. For further explanation on the application of high-pass filters, please refer to the relevant instructions above.

[0094] In some embodiments, the processor may also employ spectral analysis techniques such as the Welch method, autoregressive (AR) model method, and wavelet transform to determine the power spectral density function.

[0095] Frequency domain features 420 are used to reflect the distribution of signal energy across different frequency ranges. Frequency domain features may include the spectral centroid, spectral entropy, and band energy ratio.

[0096] The spectral centroid is used to reflect the center frequency of the power concentration region of the entire power spectral density function.

[0097] Spectral entropy reflects the degree of dispersion of power in the frequency dimension of the power spectral density function. The higher the spectral entropy, the more evenly the power is distributed across most frequency ranges; the lower the spectral entropy, the more concentrated the power is in a few specific frequency ranges.

[0098] The band power ratio reflects the weight of power within a specific frequency range in the overall power spectrum. A higher band power ratio for a given frequency range indicates a greater influence of that frequency range on the overall power spectrum.

[0099] In some embodiments, the band energy ratio can reflect the hardness characteristics of the current polishing area. A higher band energy ratio in the high-frequency range indicates a greater proportion of high-frequency disturbances during the polishing process, suggesting higher hardness in the current polishing area and the potential presence of localized abnormal morphologies such as hard spots or burrs. The high-frequency range can be predetermined based on prior experience.

[0100] In some embodiments, the processor can determine frequency domain characteristics in a variety of ways. For example, the processor takes the power spectral density function as input and outputs the spectral centroid using methods such as standard integration and discrete summation; it outputs the spectral entropy using methods such as Shannon spectral entropy and normalized spectral entropy; and it outputs the bandgap energy ratio using methods such as discrete summation and weighted averaging.

[0101] Some embodiments in this specification introduce frequency domain features such as spectral centroid, spectral entropy, and band energy ratio as inputs to a second model, enabling the model to more comprehensively identify the dynamic characteristics of the grinding signal. This significantly improves the accuracy of surface roughness inversion and its adaptability to complex processing conditions, thereby effectively optimizing the quality and efficiency of the grinding process.

[0102] In some embodiments, the processor can train a second model based on a large number of second training samples with second labels. The processor acquires preferred grinding records used to construct the first training samples, and based on the number of interrupted measurements in the preferred grinding records, splits the preferred grinding records into multiple sub-records corresponding to the number of interrupted measurements. In each sub-record, the historical workpiece type, historical grinding parameters, historical friction vibration characteristics, historical contact disturbance characteristics, and historical frequency domain characteristics corresponding to the grinding stage are determined as the second training samples, and the historical surface roughness measured at the end of the grinding stage is determined as the corresponding second label. For an explanation of the preferred grinding records and the acquisition steps, please refer to the relevant content of the first training samples.

[0103] For example, for a preferred grinding record with 3 interrupted measurements, the grinding process can be represented as: Grinding I → Interrupted Measurement I → Grinding II → Interrupted Measurement II → Grinding III → Interrupted Measurement III → Grinding Complete. The processor can divide this preferred grinding record into three sub-records based on the number of interrupted measurements: the first sub-record (Grinding I → Interrupted Measurement I), the second sub-record (Grinding II → Interrupted Measurement II), and the third sub-record (Grinding III → Interrupted Measurement III). The processor uses the inherent historical workpiece type, historical grinding parameters of Grinding I, historical friction vibration characteristics, historical contact disturbance characteristics, and historical frequency domain characteristics as the second training sample for the first sub-record, and the historical surface roughness collected by Interrupted Measurement I as the second label for the first sub-record. The processor can then determine the second training sample with the second label for the second sub-record using the above method. This allows for the construction of multiple second training samples with the second label based on a single preferred grinding record.

[0104] The step of the processor training a primary second model to obtain a second model based on a second training sample with a second label is similar to the step of training a primary first model to obtain a first model based on a first training sample with a first label, as can be seen in the relevant explanation above.

[0105] Some embodiments in this specification, by introducing frictional vibration characteristics and contact disturbance characteristics, can more accurately reflect the dynamic changes in the surface state during grinding, thereby improving the real-time performance and accuracy of surface roughness assessment.

[0106] In some embodiments, the grinding head is made of polyurethane, and the interior of the grinding head is filled with magnetorheological fluid and integrated with an electromagnetic coil.

[0107] In some embodiments, the grinding head is made of polyurethane. For example, polyether polyurethane, polyester polyurethane, polyimide polyurethane, polyurea polyurethane, etc. The polyurethane may be mixed with silicon carbide, alumina and other micro powders, and crosslinking agents such as isocyanate (MDI) may be added to the polyurethane to further improve the wear resistance of the grinding head.

[0108] In some embodiments, the grinding head has an internal cavity filled with magnetorheological fluid and an integrated electromagnetic coil.

[0109] Magnetorheological fluids are fluids that are sensitive to magnetic fields. They are typically composed of a mixture of soft magnetic particles, a carrier fluid, and a stabilizer.

[0110] An electromagnetic coil is a coil device that generates a magnetic field by passing an electric current through it. Electromagnetic coils can be used to generate controllable magnetic fields. Under the influence of a magnetic field, magnetorheological fluids can change from a low-viscosity fluid state to a high-viscosity solid-like state, thus adjusting the viscosity of the magnetorheological fluid. By increasing or decreasing the current in the electromagnetic coil, the viscosity of the magnetorheological fluid can be adjusted, thereby indirectly controlling the stiffness of the grinding head.

[0111] Grinding head stiffness refers to the grinding head's ability to resist deformation. The greater the grinding head stiffness, the harder the grinding head and the higher the grinding efficiency; the lower the grinding head stiffness, the softer the grinding head and the easier it is to conform to the current grinding area.

[0112] The curvature threshold is a critical parameter for determining the magnitude of curvature in the current polishing area. The curvature threshold can be determined based on prior experience.

[0113] In some embodiments, it is known that there are two cases: curvature > curvature threshold and curvature ≤ curvature threshold. In response to curvature > curvature threshold, the processor determines the amount of current reduction in the electromagnetic coil to reduce the stiffness of the grinding head, so that the grinding head softens and conforms to the current grinding area. In response to curvature ≤ curvature threshold, the processor determines the amount of current increase in the electromagnetic coil to increase the stiffness of the grinding head, so that the grinding head hardens and cuts the current grinding area.

[0114] In some embodiments, the processor can determine the current decrease and current increase of the electromagnetic coil in various ways. For example, the processor can preset the current decrease and current increase of the electromagnetic coil based on prior experience; or, the processor can establish a second preset table. The second preset table includes curvature, curvature threshold, and the correspondence between different current decreases or current increases. The processor can determine the current current decrease or current increase by consulting the second preset table and based on the current curvature and curvature threshold. The correspondence can be: the greater the curvature exceeds the curvature threshold, the greater the current decrease; the greater the curvature is less than the curvature threshold, the greater the current increase.

[0115] In some embodiments, the processor determines the current reduction based on curvature and contact perturbation characteristics; and determines the current increase based on curvature and surface roughness. For more information on contact perturbation characteristics, see the related description above.

[0116] In some embodiments, the current reduction is positively correlated with the curvature and contact disturbance characteristics of the current grinding area. A greater curvature in the current grinding area indicates a steeper surface (e.g., deep concave surfaces, sharp bends), resulting in a greater current reduction and lower grinding head stiffness. This allows the grinding head to better adapt to the contours of steep surfaces, avoiding insufficient contact or localized over-grinding due to excessive grinding head stiffness. A greater contact disturbance characteristic in the current grinding area indicates that the grinding head is more susceptible to stronger eccentric torque fluctuations due to factors such as non-uniform contact and edge impacts. The processor can further soften the grinding head, enhancing its ability to buffer instantaneous impact loads, thereby suppressing vibration and sway during the grinding process.

[0117] In some embodiments, the current increase is negatively correlated with the curvature of the current grinding area and positively correlated with the surface roughness. The smaller the curvature of the current grinding area (e.g., a gently sloping surface, a flat surface), the larger the current increase, thereby increasing the rigidity of the grinding head and providing stronger cutting support force, thus improving grinding efficiency while maintaining grinding accuracy. Conversely, the larger the surface roughness of the current grinding area, the larger the current increase, quickly removing surface protrusions and improving grinding efficiency.

[0118] In some embodiments of this specification, a grinding head made of polyurethane and filled with magnetorheological fluid and an integrated electromagnetic coil can be used to flexibly switch the stiffness of the grinding head to adapt to different grinding areas, which significantly improves grinding efficiency and quality.

[0119] Figure 5This is an exemplary schematic diagram illustrating the adjustment of coolant spray flow rate based on some embodiments of this specification. In some embodiments, the processor determines an adjustment step size 520 for the target grinding parameters based on the grinding head stiffness 510 and surface roughness 333; based on the adjustment step size 520, it progressively adjusts the current grinding parameters 321 to the target grinding parameters 360; based on the target grinding parameters 360, it adjusts the coolant spray flow rate 530; and based on the coolant spray flow rate 530, it controls the industrial robot to perform grinding operations while simultaneously spraying coolant 540 into the current grinding area.

[0120] The adjustment step size of 520 is the increment or decrement of each parameter adjustment when gradually correcting the current grinding parameters to the target grinding parameters. The adjustment step size can be expressed as the adjustment speed of grinding pressure, feed rate, and grinding angle.

[0121] In some embodiments, the adjustment step size 520 is positively correlated with the surface roughness of the current grinding area and negatively correlated with the grinding head stiffness. The greater the current grinding head stiffness, the more sensitive the grinding head is to force transmission response. To avoid damaging the workpiece surface, a smaller adjustment step size should be used to achieve a smooth transition of grinding parameters. Conversely, the smaller the current grinding head stiffness, the less responsive the grinding head is to force transmission. A larger adjustment step size can be used to quickly and gradually correct the current grinding parameters to the target grinding parameters, improving the adjustment efficiency of the grinding process. The greater the surface roughness of the current grinding area, the larger the adjustment step size should be used to improve grinding efficiency. Conversely, the smaller the surface roughness of the current grinding area, indicating higher workpiece surface flatness, a smaller adjustment step size should be used to avoid over-grinding and affecting the workpiece surface.

[0122] In some embodiments, after the processor adjusts the stiffness of the grinding head, the adjustment step size needs to be updated in real time to avoid negative effects such as grinding chatter caused by grinding pressure instability.

[0123] In some embodiments, the processor adjusts the current polishing parameters step by step based on the adjustment step size until the current polishing parameters converge to the preset error range allowed by the target parameters in order to perform the polishing operation.

[0124] The spray flow rate 530 reflects the consumption of coolant 540. In some embodiments, frictional heat is generated during the grinding process. If this heat is not dissipated in time, the local temperature of the workpiece surface will rise sharply, causing the workpiece surface to soften and deform, while also accelerating the wear and dulling of the grinding head and shortening its service life. Therefore, during the grinding process, coolant 540 can be sprayed onto the current grinding area to lower the temperature and flush away the dust and debris generated during grinding. Coolant 540 includes, but is not limited to, water and synthetic cutting fluid.

[0125] In some embodiments, the processor can adjust the coolant injection flow rate 530 in various ways. For example, the injection flow rate is positively correlated with the target grinding pressure, the target feed rate, and the target grinding angle. The higher the target grinding pressure, the higher the normal pressure between the grinding head and the current grinding area, the stronger the frictional shearing effect, and the more frictional heat is generated; the faster the target feed rate, the higher the rotational speed of the grinding head, and the more frictional heat is generated; the larger the target grinding angle, the more the contact pattern between the grinding head and the workpiece transitions from surface contact to line contact or point contact, the smaller the contact area between the grinding head and the workpiece, the easier it is to form local hot spots, and it is necessary to increase the coolant injection flow rate.

[0126] In some embodiments, when the grinding operation begins, the processor can control the cooling nozzles integrated on the industrial robot to spray coolant onto the current grinding area at a preset spray flow rate, and adjust the spray flow rate in real time during the grinding operation based on the target grinding parameters.

[0127] Some embodiments in this specification use a step-by-step adjustment of grinding parameters to ensure the stability of the grinding parameter adjustment and effectively avoid negative effects such as grinding chatter caused by grinding pressure instability. Adjusting the coolant spray flow rate based on the target grinding parameters achieves precise temperature control during the grinding process, avoids thermal deformation of the workpiece surface, ensures processing accuracy, and extends the service life of the grinding head.

[0128] In some embodiments, in response to a decrease rate of surface roughness below a decrease rate threshold and mechanical response data exceeding a safety threshold, the processor determines the attitude fine-tuning angle based on the number of cycles of the grinding operation.

[0129] The rate of descent threshold is a criterion for determining whether the rate of descent of surface roughness is abnormal. When the rate of descent of surface roughness is lower than the threshold, it indicates that the contact posture between the grinding head and the workpiece may be unreasonable, or that the grinding head is dulled, resulting in low grinding efficiency. For more information on the rate of descent of surface roughness, please refer to the relevant explanation above.

[0130] The safety threshold is the criterion for judging whether the mechanical response data is abnormal. When the mechanical response data exceeds the safety threshold, it indicates that the load on the grinding head is too large, which can easily lead to serious consequences such as grinding chatter, workpiece scratches, and grinding deformation.

[0131] In some embodiments, mechanical response data exceeding a safety threshold can refer to: tangential resultant force. The absolute value of the resultant force exceeds the threshold of the tangential resultant force, resulting in a resultant torque. The absolute value exceeds the resultant torque threshold. The descent rate threshold, the tangential resultant force threshold, and the resultant torque threshold in the safety threshold can be determined based on the actual application scenario and requirements.

[0132] The attitude fine-tuning angle is the amount of angle adjustment used to adjust the attitude of the grinding head. In some embodiments, when the rate of decrease in surface roughness is lower than a decrease rate threshold and the mechanical response data exceeds a safety threshold, the processor can fine-tune the angle of the grinding head, thereby changing the contact pattern between the grinding head and the workpiece surface and optimizing the stress distribution in the contact area.

[0133] In some embodiments, the current attitude angle of the grinding head is denoted as θ, and the attitude fine-tuning angle is denoted as ψ. The processor can, based on the attitude fine-tuning angle ψ, adjust the joint angles of the industrial robot's end effector to allow the grinding head to move within [θ]. The grinding head oscillates periodically within the angular range of [ψ, θ+ψ]. The oscillation angular velocity of the grinding head can be determined based on prior experience.

[0134] In some embodiments, the processor determines the attitude fine-tuning angle based on the number of grinding cycles completed. The attitude fine-tuning angle is directly related to the number of cycles S1-S4 of the grinding operation. The more cycles S1-S4 have been completed for the current grinding area, the less efficient the grinding is in the current attitude, and even repeated grinding operations will not achieve the desired grinding effect. In this case, the processor can increase the attitude fine-tuning angle, causing the grinding head to oscillate more violently. By exploring new contact postures, the effective cutting contact area between the grinding head and the workpiece surface is optimized, and the load borne by the grinding head is distributed.

[0135] In some embodiments, after determining the attitude fine-tuning angle, the processor can control the grinding head to periodically oscillate based on the attitude fine-tuning angle until the mechanical response data falls back within a safe threshold. The processor continuously monitors the mechanical response data, and stops the periodic oscillation of the grinding head in response to the mechanical response data falling back within the safe threshold.

[0136] In some embodiments of this specification, when the grinding efficiency drops abnormally and there is a risk of overload, the processor determines the attitude fine-tuning angle based on the number of grinding cycles. Based on the attitude fine-tuning angle, the processor controls the grinding head to oscillate periodically until the mechanical response data falls back to within the safe threshold. This helps to dynamically optimize the grinding contact state, alleviate local load abnormalities, and effectively improve grinding stability and safety.

[0137] The basic concepts have been described above. Obviously, for those skilled in the art, the detailed disclosure above is merely illustrative and does not constitute a limitation of this specification. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this specification. Such modifications, improvements, and corrections are suggested in this specification and therefore remain within the spirit and scope of the exemplary embodiments described herein.

[0138] Furthermore, this specification uses specific terms to describe embodiments thereof. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different locations in this specification are not necessarily the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this specification can be appropriately combined.

[0139] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this specification are not intended to limit the order of the processes and methods described herein. Although various examples have been discussed in the foregoing disclosure of some embodiments of the invention that are currently considered useful, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the spirit and scope of the embodiments described herein. For example, while the system components described above can be implemented using hardware devices, they can also be implemented solely using software solutions, such as installing the described system on existing servers or mobile devices.

[0140] Similarly, it should be noted that, in order to simplify the description disclosed herein and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of embodiments in this specification may sometimes combine multiple features into a single embodiment, drawing, or description thereof. However, this method of disclosure does not imply that the subject matter of this specification requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of a single embodiment disclosed above.

[0141] In some embodiments, numbers describing the quantity of components and attributes are used. It should be understood that such numbers used in the description of embodiments are modified in some examples with the terms "approximately," "approximately," or "generally." Unless otherwise stated, "approximately," "approximately," or "generally" indicates that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may be changed based on the characteristics required by individual embodiments. In some embodiments, numerical parameters are to take into account specified significant digits and employ a general method of digit reservation. Although the numerical ranges and parameters used to confirm their breadth of range in some embodiments of this specification are approximate values, in specific embodiments, such values ​​are set as precisely as feasible.

[0142] For each patent, patent application, patent application publication, and other material such as articles, books, specifications, publications, and documents referenced in this specification, the entire contents of which are incorporated herein by reference. This excludes historical application documents that are inconsistent with or conflict with the content of this specification, as well as documents that limit the broadest scope of the claims in this specification (currently or subsequently appended to this specification). It should be noted that in the event of any inconsistency or conflict between the descriptions, definitions, and / or terminology used in the supplementary materials to this specification and the content of this specification, the descriptions, definitions, and / or terminology used in this specification shall prevail.

[0143] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A workpiece grinding system based on a six-axis force sensor, characterized in that, The workpiece grinding system includes a management platform and a sensing and control platform. The sensing and control platform includes a six-axis force sensor and an industrial robot. The management platform is configured as follows: Based on a six-axis force sensor, the mechanical response data of the current grinding area is determined; The surface roughness of the current grinding area is determined based on the workpiece type, current grinding parameters, and the mechanical response data. Based on the workpiece type, the curvature of the current grinding area, the surface roughness, and the target roughness, the target grinding parameters of the current grinding area are determined by a first model. The target grinding parameters include the target grinding pressure, the target feed speed, and the target grinding angle. The first model is a machine learning model. The industrial robot is controlled to perform grinding operations based on the target grinding parameters. as well as, The above steps are executed repeatedly. In response to the difference between the surface roughness of the current grinding area and the target roughness being less than a difference threshold, the industrial robot is controlled to move to the next grinding area to perform the grinding operation. The grinding operation includes: adjusting the joint torque of the industrial robot to apply the target grinding pressure; adjusting the drive power of the industrial robot to match the target feed speed; and adjusting the joint angle of the industrial robot to align with the target grinding angle.

2. The workpiece grinding system according to claim 1, characterized in that, The management platform is further configured as follows: Based on the mechanical response data, vibration characteristics are determined, including frictional vibration characteristics and contact disturbance characteristics; Based on the workpiece type, the current grinding parameters, and the vibration characteristics, the surface roughness of the current grinding area is determined by a second model, which is a machine learning model.

3. The workpiece grinding system according to claim 1, characterized in that, The grinding head is made of polyurethane, and its interior is filled with magnetorheological fluid and integrated with an electromagnetic coil. The management platform is further configured as follows: In response to the curvature being greater than the curvature threshold, the amount of current reduction in the electromagnetic coil is determined to reduce the stiffness of the grinding head, thereby softening the grinding head and conforming it to the current grinding area. In response to the curvature being less than or equal to the curvature threshold, an increase in the current of the electromagnetic coil is determined to increase the stiffness of the grinding head, thereby hardening the grinding head and cutting the current grinding area.

4. The workpiece grinding system according to claim 3, characterized in that, The management platform is further configured as follows: Based on the grinding head stiffness and the surface roughness, the adjustment step size of the target grinding parameters is determined; Based on the adjustment step size, the current polishing parameters are gradually adjusted to the target polishing parameters; Based on the target grinding parameters, adjust the coolant injection flow rate; as well as, Based on the spray flow rate of the coolant, the industrial robot is controlled to spray the coolant into the current grinding area while performing the grinding operation.

5. The workpiece grinding system according to claim 4, characterized in that, The management platform is further configured as follows: In response to the surface roughness decreasing at a rate below a rate of decrease threshold and the mechanical response data exceeding a safety threshold, the attitude fine-tuning angle is determined based on the number of cycles of the grinding operation. Based on the aforementioned attitude fine-tuning angle, the grinding head is controlled to oscillate periodically until the mechanical response data falls back to within the aforementioned safety threshold.

6. A workpiece grinding method based on a six-axis force sensor, characterized in that, The workpiece grinding method is executed by the management platform of the workpiece grinding system, and the workpiece grinding method includes: Based on a six-axis force sensor, the mechanical response data of the current grinding area is determined; The surface roughness of the current grinding area is determined based on the workpiece type, current grinding parameters, and the mechanical response data. Based on the workpiece type, the curvature of the current grinding area, the surface roughness, and the target roughness, the target grinding parameters of the current grinding area are determined by a first model. The target grinding parameters include the target grinding pressure, the target feed speed, and the target grinding angle. The first model is a machine learning model. The industrial robot is controlled to perform grinding operations based on the target grinding parameters. as well as, The above steps are executed repeatedly. In response to the difference between the surface roughness of the current grinding area and the target roughness being less than a difference threshold, the industrial robot is controlled to move to the next grinding area to perform the grinding operation. The grinding operation includes: adjusting the joint torque of the industrial robot to apply the target grinding pressure; adjusting the drive power of the industrial robot to match the target feed speed; and adjusting the joint angle of the industrial robot to align with the target grinding angle.

7. The workpiece grinding method according to claim 6, characterized in that, Determining the surface roughness of the current grinding area includes: Based on the mechanical response data, vibration characteristics are determined, including frictional vibration characteristics and contact disturbance characteristics; Based on the workpiece type, the current grinding parameters, and the vibration characteristics, the surface roughness of the current grinding area is determined by a second model, which is a machine learning model.

8. The workpiece grinding method according to claim 6, characterized in that, The grinding head is made of polyurethane, and its interior is filled with magnetorheological fluid and integrated with an electromagnetic coil. The workpiece grinding method further includes: In response to the curvature being greater than the curvature threshold, the amount of current reduction in the electromagnetic coil is determined to reduce the stiffness of the grinding head, thereby softening the grinding head and conforming it to the current grinding area. In response to the curvature being less than or equal to the curvature threshold, an increase in the current of the electromagnetic coil is determined to increase the stiffness of the grinding head, thereby hardening the grinding head and cutting the current grinding area.

9. The workpiece grinding method according to claim 8, characterized in that, The workpiece grinding method further includes: Based on the grinding head stiffness and the surface roughness, the adjustment step size of the target grinding parameters is determined; Based on the adjustment step size, the current polishing parameters are gradually adjusted to the target polishing parameters; Based on the target grinding parameters, adjust the coolant injection flow rate; and, Based on the spray flow rate of the coolant, the industrial robot is controlled to spray the coolant into the current grinding area while performing the grinding operation.

10. A computer-readable storage medium, characterized in that, The storage medium stores computer instructions. When the computer reads the computer instructions in the storage medium, the computer executes the workpiece grinding method as described in claim 6.

Images