An environment perception-based device error intelligent detection system and method

Abstract

This invention discloses an intelligent equipment error detection system and method based on environmental perception, relating to the field of precision machining technology. The system includes: at the end of the rough grinding process, acquiring residual deviation data of several sampling points relative to the theoretical contour, recording the coordinates of at least three reference feature points in the rough grinding machine coordinate system, and encapsulating this data into a workpiece-following data package; based on the discrete residual deviation data, fitting and generating a continuous error topology surface describing the spatial distribution of residual errors on the workpiece's machined surface; after transferring the workpiece to a fine grinding machine and completing clamping, re-measuring the coordinates of at least three reference feature points in the fine grinding machine coordinate system, calculating the coordinate transformation relationship of the workpiece before and after the process transfer based on the two sets of reference feature point coordinates, and obtaining an inherited error field; and compensating and correcting the cutting depth of each machining point on the fine grinding trajectory based on the inherited error field. This significantly improves the cross-process machining accuracy and the controllability of process flow.

Description

Technical Field

[0001] This invention relates to the field of precision machining technology, specifically to an intelligent equipment error detection system and method based on environmental perception. Background Technology

[0002] Grinding is typically performed in two steps: coarse grinding and fine grinding, to balance processing efficiency and surface finish. In actual production, after the coarse grinding step, the workpiece inevitably has a certain amount of residual error. This residual error is spatially non-uniformly distributed, manifested as different deviation values ​​and directions in different areas of the workpiece's machined surface.

[0003] In existing technologies, rough grinding and fine grinding each constitute an independent closed-loop control system. The fine grinding process uses the theoretical profile as the sole processing reference, assuming the initial state of the workpiece entering fine grinding is ideal, and completely ignoring the residual error distribution information left over from the rough grinding process. This approach has the following drawbacks: Firstly, fine grinding adds its own processing error on top of the residual error of coarse grinding, causing the errors of the two processes to accumulate randomly, and the final processing accuracy is constrained by the systematic deviation of coarse grinding.

[0004] Secondly, when the workpiece is transferred from coarse grinding to fine grinding, it needs to be re-clamped. The positional deviation introduced by the clamping is superimposed with the residual error of coarse grinding. The fine grinding control system cannot effectively distinguish and independently compensate for the two, which further exacerbates the uncontrollability of the error.

[0005] Third, existing technologies lack quantitative assessment methods for the solvability of residual errors in coarse grinding. It is impossible to determine at the process handover stage whether the residual errors in coarse grinding exceed the compensation capacity of the fine grinding process, resulting in defective workpieces flowing directly into the fine grinding process, causing a waste of processing resources, and the final product accuracy cannot be guaranteed. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent detection system and method for equipment errors based on environmental perception, so as to solve the problems raised in the prior art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an intelligent detection method for equipment errors based on environmental perception, specifically comprising the following steps: Step 100: At the end of the rough grinding process, the surface of the workpiece after rough grinding is measured and sampled to obtain residual deviation data of several sampling points relative to the theoretical contour, and the coordinates of at least three reference feature points on the workpiece in the coordinate system of the rough grinding machine are recorded. The residual deviation data and the coordinates of the reference feature points are encapsulated into a workpiece accompanying data package. Step 200: Based on the discrete residual deviation data in the accompanying data package, fit and generate a continuous error topology surface that describes the spatial distribution of residual error on the workpiece machining surface; Step 300: After transferring the workpiece to the fine grinding machine and completing the clamping, remeasure the coordinates of the at least three reference feature points in the coordinate system of the fine grinding machine. Calculate the coordinate transformation relationship of the workpiece before and after the process transfer based on the coordinates of the two sets of reference feature points. Transform the error topology surface to the coordinate system of the fine grinding machine to obtain the inherited error field. Step S400: Based on the inherited error field, compensate and correct the cutting depth of each processing point on the fine grinding trajectory to eliminate the residual error left over from the rough grinding process.

[0008] Furthermore, in step S100, the measurement and sampling of the workpiece surface after rough grinding to obtain residual deviation data of several sampling points relative to the theoretical contour specifically involves: At the end of the rough grinding process, while the workpiece is still in the clamping state, a laser displacement sensor is driven to scan along the workpiece's machining surface according to a preset grid path. The grid path covers the entire machining area of ​​the workpiece, and the spacing between adjacent scanning lines and the sampling interval along the line are determined based on the workpiece's machining area and the preset sampling density. For each sampling point, the difference between the workpiece surface height value measured by the laser displacement sensor and the theoretical contour height value corresponding to that point is obtained to obtain the residual deviation value and deviation direction of that sampling point. A positive value indicates that the material residue is too high, and a negative value indicates that the material removal is too high. The planar coordinates of all sampling points and the corresponding residual deviation values ​​together constitute the residual deviation dataset. The residual deviation dataset and the coordinates of at least three reference feature points are encapsulated into the workpiece accompanying data package.

[0009] Furthermore, in step 200, based on the discrete residual deviation data in the accompanying data packet, a continuous error topological surface describing the spatial distribution of residual errors on the workpiece machining surface is fitted and generated, specifically as follows: Using the planar coordinates of each sampling point in the residual deviation dataset as the independent variable and the corresponding residual deviation value as the dependent variable, a thin plate spline interpolation method is used for surface fitting. The thin plate spline interpolation takes minimizing the overall bending energy of the surface as the constraint objective, and generates a continuous error topological surface S(x,y) that smoothly transitions between sampling points, provided that the interpolation conditions of all sampling points are met. The minimum bending energy constraint makes the shape of the fitted surface consistent with the physical characteristics of the continuous and gradual change of the actual machining error, avoiding the introduction of spurious error fluctuations between sampling points.

[0010] Furthermore, in step S300, the coordinates of the at least three reference feature points in the fine grinding machine coordinate system are remeasured. Based on the coordinates of the two sets of reference feature points, the coordinate transformation relationship of the workpiece before and after the process transfer is calculated. The error topology surface is then transformed to the fine grinding machine coordinate system to obtain the inherited error field, specifically: After the fine grinding machine tool is clamped and stabilized, the machine tool probe is driven to contact the at least three reference feature points in sequence, and the measured coordinates of each reference feature point in the fine grinding machine tool coordinate system are recorded. Using the coordinates of the reference feature points in the coarse grinding machine tool coordinate system as the source point set and the corresponding coordinates in the fine grinding machine tool coordinate system as the target point set, the rigid body coordinate transformation relationship of the workpiece before and after the process transfer is analytically solved by minimizing the sum of squared Euclidean distances between the two sets of corresponding points. The rigid body transformation relationship consists of rotational and translational components, reflecting the actual pose deviation of the workpiece after reclamping. The planar coordinates of each discrete sampling point in the residual deviation data set are mapped to the fine grinding coordinate system through rigid body transformation to obtain a new discrete point coordinate set. Then, thin plate spline interpolation fitting is performed again on the new discrete point coordinate set to generate the inherited error field S′(x,y). The inherited error field is strictly aligned with the fine grinding machining coordinate system in space to eliminate the interference of pose deviation introduced by reclamping on error propagation.

[0011] Furthermore, before compensating and correcting the fine grinding trajectory based on the inherited error field, a step is included to assess the resolvability of the residual error from the coarse grinding, specifically: Step S301: Based on the inherited error field S′(x,y), perform area integration on the absolute amount of error in the entire processing area of ​​the workpiece, and compare the obtained integral value with the product of the workpiece processing area and the tolerance band width of the fine grinding process to obtain the cross-process error inheritance coefficient Γ, where Γ represents the proportion of the total residual error of the rough grinding process to the available compensation capacity of the fine grinding process. Step S302: Preset lower threshold Γ low With upper limit threshold Γ high When Γ < Γ low If the residual error is determined to be self-resolving within the natural machining allowance range of fine grinding, the workpiece proceeds to fine grinding according to the conventional process without compensation correction. low ≤Γ<Γ high When residual error is determined, proactive intervention is needed to compensate and correct the cutting depth before proceeding to fine grinding; when Γ≥Γ high If the residual error in the coarse grinding process exceeds the upper limit of the resolving capacity of the fine grinding process, the workpiece is returned to the coarse grinding process for re-grinding. After the re-grinding is completed, steps 100 to 400 are repeated until Γ decreases to Γ. high the following.

[0012] Furthermore, in step S400, based on the inherited error field, the cutting depth of each machining point on the fine grinding trajectory is compensated and corrected, specifically as follows: For each machining point on the fine grinding trajectory, the error value of the inherited error field S′(x,y) at that point is extracted. Based on the ratio of the available process allowance at that point to the maximum absolute error value of the inherited error field in the entire machining area, the error digestion coefficient λ at that point is determined. The value of λ is in the range of (0,1]. A larger value is taken where the allowance is sufficient to actively eliminate the inherited error, and a smaller value is taken where the allowance is tight to prevent overcutting. The product of the error digestion coefficient and the error value of the inherited error field at that point is added to the nominal cutting depth at that point to obtain the actual compensated cutting depth at that point. A fine grinding compensation machining path is generated according to the actual compensated cutting depth of each machining point, driving the fine grinding machine to perform dynamic cutting depth correction, so that the fine grinding process can actively offset the residual error left by the rough grinding process while eliminating its own machining error.

[0013] An intelligent equipment error detection system based on environmental perception, the system includes a surface residual error acquisition module, a reference feature point measurement module, a workpiece following data management module, an error topology surface construction module, a coordinate registration module, an error genetic evaluation module, a compensation path generation module, and a fine grinding CNC execution module; The surface residual error acquisition module and the reference feature point measurement module's rough grinding side probes are both connected to the workpiece accompanying data management module via industrial Ethernet. The residual deviation dataset and the coordinates of the rough grinding side reference feature points are uploaded to the workpiece accompanying data management module and encapsulated into a workpiece accompanying data package. The workpiece data management module is connected to the error topology surface construction module via industrial Ethernet. When the workpiece enters the fine grinding process, it sends a read completion signal to the error topology surface construction module, triggering the subsequent processing flow to start sequentially. The error topology surface construction module, coordinate registration module, error genetic evaluation module and compensation path generation module are integrated and run on the same industrial control computer. The four modules sequentially transmit processing results through software data interfaces, and complete surface fitting, coordinate registration, error evaluation and path generation in turn. The fine grinding side probe of the reference feature point measurement module is connected to the industrial control computer via industrial Ethernet, and sends the coordinates of the fine grinding side reference feature points directly to the coordinate registration module. The error genetic evaluation module is connected to the CNC system of the rough grinding machine and the compensation path generation module via industrial Ethernet. When it is determined that the machine needs to be returned for re-grinding, it sends a re-grinding command to the CNC system of the rough grinding machine. When it is determined that the machine can proceed to fine grinding, it sends a process flow decision command to the compensation path generation module. The compensation path generation module is connected to the fine grinding CNC execution module via an industrial Ethernet, and sends the generated compensation machining path to the fine grinding CNC execution module in the form of CNC code. After the fine grinding CNC execution module completes the machining, it sends a machining completion signal to the industrial control computer via an industrial Ethernet. The industrial control computer archives the machining data to the workpiece accompanying data management module, thus completing the cross-process compensation closed loop.

[0014] Furthermore, the surface residual error acquisition module includes a laser displacement sensor and its drive control unit. The laser displacement sensor is mounted on the side of the spindle of the rough grinding machine or on a dedicated measuring bracket, and its measurement direction is consistent with the normal direction of the workpiece machining surface. The drive control unit controls the scanning movement of the laser displacement sensor according to a preset grid path. The scanning line spacing and sampling interval along the grid path are determined according to the workpiece machining area and the preset sampling density. The drive control unit calculates the difference between the actual height value of the sensor at each sampling point and the theoretical contour height value of the corresponding point read by the CNC system in real time to generate a signed residual deviation dataset. After sampling, the residual deviation dataset is uploaded to the workpiece accompanying data management module via industrial Ethernet.

[0015] Furthermore, the workpiece accompanying data management module includes a workpiece identification unit and an accompanying data storage unit. The workpiece identification unit includes an RFID reader or QR code scanner respectively installed at the rough grinding machine's unloading station and the fine grinding machine's loading station. The RFID reader or QR code scanner is connected to the accompanying data storage unit via an industrial Ethernet. When the workpiece arrives at each station, it automatically reads the workpiece's unique identifier and triggers the writing or reading operation of the corresponding accompanying data packet. The accompanying data storage unit is a storage server that establishes bidirectional industrial Ethernet communication connections with the surface residual error acquisition module, the reference feature point measurement module's rough grinding side probe, and the industrial control computer. When the workpiece enters the fine grinding process, it sends a reading completion signal carrying the workpiece's unique identifier to the industrial control computer. The industrial control computer retrieves the corresponding accompanying data packet from the accompanying data storage unit based on the identifier and triggers the error topology surface construction module to start.

[0016] Furthermore, the error genetic evaluation module also includes a grinding feedback unit. When the cross-process error genetic coefficient is not lower than a preset upper limit threshold, the grinding feedback unit sends a grinding instruction to the CNC system of the rough grinding machine via industrial Ethernet. The grinding instruction contains the planar position information of the out-of-tolerance area in the inherited error field and the corresponding residual deviation. The rough grinding machine performs targeted grinding on the out-of-tolerance area according to the grinding instruction. After the grinding is completed, the CNC system of the rough grinding machine sends a grinding completion signal to the industrial control computer via industrial Ethernet. The industrial control computer triggers the surface residual error acquisition module to re-scan and sample the workpiece surface, update the workpiece accompanying data packet, and re-trigger the error topology surface construction module, coordinate registration module, and error genetic evaluation module to perform evaluation in sequence until the cross-process error genetic coefficient drops below the preset upper limit threshold. Then, a process flow decision instruction allowing fine grinding is output.

[0017] Furthermore, after completing the fine grinding process, the fine grinding CNC execution module sends a processing completion signal carrying a unique workpiece identifier to the industrial control computer via industrial Ethernet. After receiving the processing completion signal, the industrial control computer archives the inherited error field parameters, error inheritance coefficients, error digestion coefficients of each processing point, and compensation processing path data used in this processing to the accompanying data storage unit of the workpiece accompanying data management module, indexed by the unique workpiece identifier. The archived data serves as the error compensation record for the entire process of the workpiece, for subsequent quality traceability and batch error pattern analysis of similar workpieces.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention constructs a continuous error topological surface on the workpiece surface at the end of the rough grinding stage and allows it to flow with the workpiece. After coordinate registration, the residual error field of the rough grinding is transferred to the fine grinding process, driving the fine grinding process to actively offset the residual error left by the rough grinding while eliminating its own processing error. This achieves the directional elimination of cross-process errors rather than passive accumulation. At the same time, the introduction of a cross-process error inheritance coefficient quantifies the solvability of residual errors, enabling the active interception of unqualified workpieces at the process handover stage, thereby significantly improving the cross-process processing accuracy and the controllability of process flow. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating an intelligent detection method for equipment errors based on environmental perception, according to the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Example: Figure 1 As shown, the present invention provides a technical solution. Step 100: At the end of the rough grinding process, the surface of the workpiece after rough grinding is measured and sampled to obtain residual deviation data of several sampling points relative to the theoretical contour, and the coordinates of at least three reference feature points on the workpiece in the coordinate system of the rough grinding machine are recorded. The residual deviation data and the coordinates of the reference feature points are encapsulated into a workpiece accompanying data package. Step 200: Based on the discrete residual deviation data in the accompanying data package, fit and generate a continuous error topology surface that describes the spatial distribution of residual error on the workpiece machining surface; Step 300: After transferring the workpiece to the fine grinding machine and completing the clamping, remeasure the coordinates of the at least three reference feature points in the coordinate system of the fine grinding machine. Calculate the coordinate transformation relationship of the workpiece before and after the process transfer based on the coordinates of the two sets of reference feature points. Transform the error topology surface to the coordinate system of the fine grinding machine to obtain the inherited error field. Step S400: Based on the inherited error field, compensate and correct the cutting depth of each processing point on the fine grinding trajectory to eliminate the residual error left over from the rough grinding process.

[0022] Furthermore, in step S100, the measurement and sampling of the workpiece surface after rough grinding to obtain residual deviation data of several sampling points relative to the theoretical contour specifically involves: At the end of the rough grinding process, while the workpiece is still in the clamping state, a laser displacement sensor is driven to scan along the workpiece's machining surface according to a preset grid path. The grid path covers the entire machining area of ​​the workpiece, and the spacing between adjacent scanning lines and the sampling interval along the line are determined based on the workpiece's machining area and the preset sampling density. For each sampling point, the difference between the workpiece surface height value measured by the laser displacement sensor and the theoretical contour height value corresponding to that point is obtained to obtain the residual deviation value and deviation direction of that sampling point. A positive value indicates that the material residue is too high, and a negative value indicates that the material removal is too high. The planar coordinates of all sampling points and the corresponding residual deviation values ​​together constitute the residual deviation dataset. The residual deviation dataset and the coordinates of at least three reference feature points are encapsulated into the workpiece accompanying data package.

[0023] In this study, n sampling points are obtained during the final stage of rough grinding. The planar coordinates of the i-th sampling point in the rough grinding machine coordinate system are Pi=(xi,yi), and the corresponding residual deviation value is δi. The error topological surface S(x,y) is fitted using the thin plate spline interpolation method, and its expression is as follows: S(x,y)=a0+a1x+a2y+∑ i=1 n wi·U(||P-Pi||); Where P=(x,y) are the planar coordinates of any point on the surface, and U(r)=r 2 lnr is the radial basis function of the thin plate spline, a0, a1, a2 are polynomial coefficients, and wi is the weight coefficient corresponding to the i-th sampling point; By ensuring that the surface satisfies the interpolation condition S(xi,yi)=δi (i=1,2,…,n) and the natural boundary condition, ∑ i= 1 n wi=0; ∑ i=1 n wixi=0;∑ i=1 n wiyi=0; Solve the system of equations to obtain the coefficients a0, a1, a2 and the weight coefficient wi, and obtain the uniquely determined error topological surface S(x,y).

[0024] Furthermore, in step 200, based on the discrete residual deviation data in the accompanying data packet, a continuous error topological surface describing the spatial distribution of residual errors on the workpiece machining surface is fitted and generated, specifically as follows: Using the planar coordinates of each sampling point in the residual deviation dataset as the independent variable and the corresponding residual deviation value as the dependent variable, a thin plate spline interpolation method is used for surface fitting. The thin plate spline interpolation takes minimizing the overall bending energy of the surface as the constraint objective, and generates a continuous error topological surface S(x,y) that smoothly transitions between sampling points, provided that the interpolation conditions of all sampling points are met. The minimum bending energy constraint makes the shape of the fitted surface consistent with the physical characteristics of the continuous and gradual change of the actual machining error, avoiding the introduction of spurious error fluctuations between sampling points.

[0025] Furthermore, in step S300, the coordinates of the at least three reference feature points in the fine grinding machine coordinate system are remeasured. Based on the coordinates of the two sets of reference feature points, the coordinate transformation relationship of the workpiece before and after the process transfer is calculated. The error topology surface is then transformed to the fine grinding machine coordinate system to obtain the inherited error field, specifically: After the fine grinding machine tool is clamped and stabilized, the machine tool probe is driven to contact the at least three reference feature points in sequence, and the measured coordinates of each reference feature point in the fine grinding machine tool coordinate system are recorded. Taking the coordinates of the reference feature points in the coarse grinding machine tool coordinate system as the source point set and the corresponding coordinates in the fine grinding machine tool coordinate system as the target point set, the rigid body coordinate transformation relationship of the workpiece before and after the process transfer is analytically solved by minimizing the sum of squared Euclidean distances between the two sets of corresponding points. The rigid body transformation relationship consists of rotational and translational components, reflecting the actual pose deviation of the workpiece after reclamping. After mapping the coordinates of all points in the error topology surface S(x,y) to the fine grinding machine tool coordinate system through the rigid body transformation, the inherited error field S′(x,y) in the fine grinding coordinate system is refitted to obtain the inherited error field S′(x,y) in the fine grinding coordinate system. The inherited error field is strictly aligned with the fine grinding machining coordinate system in space, eliminating the interference of the pose deviation introduced by reclamping on the error transmission.

[0026] Let A be the column vector of the coordinates of the three reference feature points on the workpiece in the coordinate system of the rough grinding machine. k =(x k A ,y k A ) T (k=1,2,3) The coordinates obtained by re-measuring in the coordinate system of the fine grinding machine tool are B. k =(x k B ,y k B ) T (k=1,2,3); where A k Let x represent the column vector of coordinates of the k-th reference feature point in the coordinate system of the rough grinding machine. k A Represents its x-coordinate, y k A B represents its ordinate; k Let x represent the column vector of coordinates of the k-th reference feature point in the coordinate system of the fine grinding machine. k B Represents its x-coordinate, y k B Indicate its ordinate; Calculate the column vector of the centroid coordinates of the three reference feature points in the coordinate system of the rough grinding machine: A'=(1 / 3)*Σ k= 1 3 A k Similarly, the column vector B' of the centroid coordinates of the three reference feature points in the coordinate system of the fine grinding machine is obtained. Remove the centroid from the coordinates of each feature point, and let A k '=A k -A',B k '=B k-B';A k 'and B k 'These represent the coordinates of the feature points after centroid removal; Construct the covariance matrix: H = Σ k=1 3 A k 'B k ' T ; where singular value decomposition is performed on H, H = U∧V T Analytical solution of the rotation matrix R=VU T The translation vector t = B' - RA'; where U and V represent two-dimensional orthogonal matrices; it should be noted that ∧ represents a diagonal matrix, whose diagonal elements are singular values; The planar coordinates of each discrete sampling point in the original residual deviation dataset are mapped to the fine grinding coordinate system through rigid body transformation B=R·A+t to obtain a new set of discrete point coordinates. Then, thin plate spline interpolation fitting is performed again on the new coordinate set to generate the inherited error field S′(x,y).

[0027] It should be noted that in the analytical solution of the rotation matrix R, it is usually necessary to verify whether det(R) = +1 is satisfied, indicating pure rotation and excluding reflection. In the two-dimensional case, if det(VU) T If ) = -1, then the result is a reflection transformation rather than a rotation transformation. Therefore, in this case, .

[0028] Furthermore, before compensating and correcting the fine grinding trajectory based on the inherited error field, a step is included to assess the resolvability of the residual error from the coarse grinding, specifically: Step S301: Based on the inherited error field S′(x,y), perform area integration on the absolute amount of error in the entire processing area of ​​the workpiece, and compare the obtained integral value with the product of the workpiece processing area and the tolerance band width of the fine grinding process to obtain the cross-process error inheritance coefficient Γ, where Γ represents the proportion of the total residual error of the rough grinding process to the available compensation capacity of the fine grinding process. Step S302: Preset lower threshold Γ low With upper limit threshold Γ high When Γ < Γ low If the residual error is determined to be self-resolving within the natural machining allowance range of fine grinding, the workpiece proceeds to fine grinding according to the conventional process without compensation correction. low ≤Γ<Γ high When residual error is determined, proactive intervention is needed to compensate and correct the cutting depth before proceeding to fine grinding; when Γ≥Γ highIf the residual error in the coarse grinding process exceeds the upper limit of the resolving capacity of the fine grinding process, the workpiece is returned to the coarse grinding process for re-grinding. After the re-grinding is completed, steps 100 to 400 are repeated until Γ decreases to Γ. high the following.

[0029] Furthermore, in step S400, based on the inherited error field, the cutting depth of each machining point on the fine grinding trajectory is compensated and corrected, specifically as follows: For each machining point on the fine grinding trajectory, the error value of the inherited error field S′(x,y) at that point is extracted. Based on the ratio of the available process allowance at that point to the maximum absolute error value of the inherited error field in the entire machining area, the error digestion coefficient λ at that point is determined. The value of λ is in the range of (0,1]. A larger value is taken where the allowance is sufficient to actively eliminate the inherited error, and a smaller value is taken where the allowance is tight to prevent overcutting. The product of the error digestion coefficient and the error value of the inherited error field at that point is added to the nominal cutting depth at that point to obtain the actual compensated cutting depth at that point. A fine grinding compensation machining path is generated according to the actual compensated cutting depth of each machining point, driving the fine grinding machine to perform dynamic cutting depth correction, so that the fine grinding process can actively offset the residual error left by the rough grinding process while eliminating its own machining error.

[0030] In this embodiment, the planar coordinates of the j-th machining point on the fine grinding trajectory are (xj, yj), and the nominal cutting depth of this point is d. j nom The error value of the inherited error field at this point is S′(xj,yj). An error digestion coefficient λ is defined, which is determined based on the available process allowance Δj at the current processing point in the fine grinding process and the maximum absolute error value of the inherited error field, Smax′=max|S′(x,y)|. λ=min(1,η·Δj / Smax′); where η is the allowance retention coefficient, η∈(0,1), used to ensure that each point still retains a process allowance of not less than (1-η)Δj after compensation, to prevent overcutting; The actual cutting depth at the j-th machining point is corrected to: d j act =d j nom +λ·S′(xj,yj); Generate a fine grinding compensation path according to the djact method, and perform dynamic correction of the cutting depth on the workpiece machining surface.

[0031] Based on the inherited error field S′(x,y), the cross-process error inheritance coefficient Γ is calculated within the workpiece machining region Ω: Γ=∫∫ Ω |S′(x,y)|dxdy / (A·δtol); where A=∫∫ Ωdxdy is the area of ​​the workpiece machining region, and δtol is the tolerance zone width of the fine grinding process; Preset lower threshold Γ low With upper limit threshold Γ high , satisfying 0<Γ low <Γ high <1; when Γ < Γ low If the residual error is determined to be self-resolving within the natural machining allowance range of fine grinding, the workpiece proceeds to fine grinding according to the conventional process without compensation correction. low ≤Γ<Γ high When determining residual error, proactive intervention is needed. After compensating and correcting the cutting depth, fine grinding can proceed.

[0032] When Γ≥Γ high If the residual error in the coarse grinding process exceeds the upper limit of the resolving capacity of the fine grinding process, the workpiece is returned to the coarse grinding process for re-grinding. After the re-grinding is completed, steps 100 to 400 are repeated until Γ decreases to Γ. high the following.

[0033] An intelligent equipment error detection system based on environmental perception, the system includes a surface residual error acquisition module, a reference feature point measurement module, a workpiece following data management module, an error topology surface construction module, a coordinate registration module, an error genetic evaluation module, a compensation path generation module, and a fine grinding CNC execution module; The surface residual error acquisition module and the reference feature point measurement module's rough grinding side probes are both connected to the workpiece accompanying data management module via industrial Ethernet. The residual deviation dataset and the coordinates of the rough grinding side reference feature points are uploaded to the workpiece accompanying data management module and encapsulated into a workpiece accompanying data package. The workpiece data management module is connected to the error topology surface construction module via industrial Ethernet. When the workpiece enters the fine grinding process, it sends a read completion signal to the error topology surface construction module, triggering the subsequent processing flow to start sequentially. The error topology surface construction module, coordinate registration module, error genetic evaluation module and compensation path generation module are integrated and run on the same industrial control computer. The four modules sequentially transmit processing results through software data interfaces, and complete surface fitting, coordinate registration, error evaluation and path generation in turn. The fine grinding side probe of the reference feature point measurement module is connected to the industrial control computer via industrial Ethernet, and sends the coordinates of the fine grinding side reference feature points directly to the coordinate registration module. The error genetic evaluation module is connected to the CNC system of the rough grinding machine and the compensation path generation module via industrial Ethernet. When it is determined that the machine needs to be returned for re-grinding, it sends a re-grinding command to the CNC system of the rough grinding machine. When it is determined that the machine can proceed to fine grinding, it sends a process flow decision command to the compensation path generation module. The compensation path generation module is connected to the fine grinding CNC execution module via an industrial Ethernet, and sends the generated compensation machining path to the fine grinding CNC execution module in the form of CNC code. After the fine grinding CNC execution module completes the machining, it sends a machining completion signal to the industrial control computer via an industrial Ethernet. The industrial control computer archives the machining data to the workpiece accompanying data management module, thus completing the cross-process compensation closed loop.

[0034] Furthermore, the surface residual error acquisition module includes a laser displacement sensor and its drive control unit. The laser displacement sensor is mounted on the side of the spindle of the rough grinding machine or on a dedicated measuring bracket, and its measurement direction is consistent with the normal direction of the workpiece machining surface. The drive control unit controls the scanning movement of the laser displacement sensor according to a preset grid path. The scanning line spacing and sampling interval along the grid path are determined according to the workpiece machining area and the preset sampling density. The drive control unit calculates the difference between the actual height value of the sensor at each sampling point and the theoretical contour height value of the corresponding point read by the CNC system in real time to generate a signed residual deviation dataset. After sampling, the residual deviation dataset is uploaded to the workpiece accompanying data management module via industrial Ethernet.

[0035] Furthermore, the workpiece accompanying data management module includes a workpiece identification unit and an accompanying data storage unit. The workpiece identification unit includes an RFID reader or QR code scanner respectively installed at the rough grinding machine's unloading station and the fine grinding machine's loading station. The RFID reader or QR code scanner is connected to the accompanying data storage unit via an industrial Ethernet. When the workpiece arrives at each station, it automatically reads the workpiece's unique identifier and triggers the writing or reading operation of the corresponding accompanying data packet. The accompanying data storage unit is a storage server that establishes bidirectional industrial Ethernet communication connections with the surface residual error acquisition module, the reference feature point measurement module's rough grinding side probe, and the industrial control computer. When the workpiece enters the fine grinding process, it sends a reading completion signal carrying the workpiece's unique identifier to the industrial control computer. The industrial control computer retrieves the corresponding accompanying data packet from the accompanying data storage unit based on the identifier and triggers the error topology surface construction module to start.

[0036] Furthermore, the error genetic evaluation module also includes a grinding feedback unit. When the cross-process error genetic coefficient is not lower than a preset upper limit threshold, the grinding feedback unit sends a grinding instruction to the CNC system of the rough grinding machine via industrial Ethernet. The grinding instruction contains the planar position information of the out-of-tolerance area in the inherited error field and the corresponding residual deviation. The rough grinding machine performs targeted grinding on the out-of-tolerance area according to the grinding instruction. After the grinding is completed, the CNC system of the rough grinding machine sends a grinding completion signal to the industrial control computer via industrial Ethernet. The industrial control computer triggers the surface residual error acquisition module to re-scan and sample the workpiece surface, update the workpiece accompanying data packet, and re-trigger the error topology surface construction module, coordinate registration module, and error genetic evaluation module to perform evaluation in sequence until the cross-process error genetic coefficient drops below the preset upper limit threshold. Then, a process flow decision instruction allowing fine grinding is output.

[0037] Furthermore, after completing the fine grinding process, the fine grinding CNC execution module sends a processing completion signal carrying a unique workpiece identifier to the industrial control computer via industrial Ethernet. After receiving the processing completion signal, the industrial control computer archives the inherited error field parameters, error inheritance coefficients, error digestion coefficients of each processing point, and compensation processing path data used in this processing to the accompanying data storage unit of the workpiece accompanying data management module, indexed by the unique workpiece identifier. The archived data serves as the error compensation record for the entire process of the workpiece, for subsequent quality traceability and batch error pattern analysis of similar workpieces.

[0038] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for intelligent detection of equipment errors based on environmental perception, characterized in that: Specifically, the steps include the following: Step 100: At the end of the rough grinding process, the surface of the workpiece after rough grinding is measured and sampled to obtain residual deviation data of several sampling points relative to the theoretical contour, and the coordinates of at least three reference feature points on the workpiece in the coordinate system of the rough grinding machine are recorded. The residual deviation data and the coordinates of the reference feature points are encapsulated into a workpiece accompanying data package. Step 200: Based on the discrete residual deviation data in the accompanying data package, fit and generate a continuous error topology surface that describes the spatial distribution of residual error on the workpiece machining surface; Step 300: After transferring the workpiece to the fine grinding machine and completing the clamping, remeasure the coordinates of the at least three reference feature points in the coordinate system of the fine grinding machine. Calculate the coordinate transformation relationship of the workpiece before and after the process transfer based on the coordinates of the two sets of reference feature points. Transform the error topology surface to the coordinate system of the fine grinding machine to obtain the inherited error field. Step S400: Based on the inherited error field, compensate and correct the cutting depth of each machining point on the fine grinding trajectory.

2. The intelligent equipment error detection method based on environmental perception according to claim 1, characterized in that: In step S100, the measurement and sampling of the workpiece surface after rough grinding to obtain residual deviation data of several sampling points relative to the theoretical profile specifically involves: At the end of the rough grinding process, while the workpiece is still in the clamping state, a laser displacement sensor is driven to scan along the workpiece processing surface according to a preset grid path. The grid path covers the entire processing area of ​​the workpiece. The spacing between adjacent scanning lines and the sampling interval along the line are determined according to the workpiece processing area and the preset sampling density. For each sampling point, the difference between the workpiece surface height value measured by the laser displacement sensor and the theoretical contour height value corresponding to that point is used to obtain the residual deviation value and deviation direction of that sampling point. The planar coordinates of all sampling points and the corresponding residual deviation values ​​together constitute the residual deviation dataset. The residual deviation dataset and the coordinates of at least three reference feature points are encapsulated into the workpiece accompanying data package.

3. The intelligent detection method for equipment errors based on environmental perception according to claim 1, characterized in that: In step 200, based on the discrete residual deviation data in the accompanying data packet, a continuous error topological surface describing the spatial distribution of residual errors on the workpiece machining surface is fitted and generated, specifically as follows: Using the planar coordinates of each sampling point in the residual deviation dataset as independent variables and the corresponding residual deviation values ​​as dependent variables, a thin plate spline interpolation method is used for surface fitting. The thin plate spline interpolation takes minimizing the overall bending energy of the surface as the constraint objective, and generates a continuous error topological surface S(x,y) that smoothly transitions between sampling points under the premise of satisfying the interpolation conditions of all sampling points.

4. The intelligent detection method for equipment errors based on environmental perception according to claim 3, characterized in that: In step S300, the coordinates of the at least three reference feature points in the fine grinding machine coordinate system are remeasured. Based on the coordinates of the two sets of reference feature points, the coordinate transformation relationship of the workpiece before and after the process transfer is calculated. The error topology surface is then transformed to the fine grinding machine coordinate system to obtain the inherited error field, specifically: After the fine grinding machine tool is clamped and stabilized, the machine tool probe is driven to contact the at least three reference feature points in sequence, and the measured coordinates of each reference feature point in the fine grinding machine tool coordinate system are recorded. Taking the coordinates of the reference feature points in the coarse grinding machine tool coordinate system as the source point set and the corresponding coordinates in the fine grinding machine tool coordinate system as the target point set, the rigid body coordinate transformation relationship of the workpiece before and after the process transfer is analytically solved by minimizing the sum of squared Euclidean distances between the two sets of corresponding points. The planar coordinates of each discrete sampling point in the residual deviation data set are mapped to the fine grinding coordinate system through rigid body transformation to obtain a new discrete point coordinate set. Then, thin plate spline interpolation fitting is performed again on the new discrete point coordinate set to generate the inherited error field S′(x,y).

5. The intelligent detection method for equipment errors based on environmental perception according to claim 4, characterized in that: Before compensating and correcting the fine grinding trajectory based on the inherited error field, the process also includes an assessment of the resolvability of the residual error from the coarse grinding, specifically: Step S301: Based on the inherited error field S′(x,y), perform area integration on the absolute amount of error in the entire processing area of ​​the workpiece, and compare the obtained integral value with the product of the workpiece processing area and the tolerance band width of the fine grinding process to obtain the cross-process error inheritance coefficient Γ, where Γ represents the proportion of the total residual error of the rough grinding process to the available compensation capacity of the fine grinding process. Step S302: Preset lower threshold Γ low With upper limit threshold Γ high When Γ < Γ low If the residual error is determined to be self-resolving within the natural machining allowance range of fine grinding, the workpiece proceeds to fine grinding according to the conventional process without compensation correction. low ≤Γ<Γ high In such cases, determining residual error requires proactive intervention. When Γ≥Γ high If the residual error in the coarse grinding process exceeds the upper limit of the resolving capacity of the fine grinding process, the workpiece is returned to the coarse grinding process for re-grinding. After the re-grinding is completed, steps 100 to 400 are repeated until Γ decreases to Γ. high the following.

6. The intelligent detection method for equipment errors based on environmental perception according to claim 5, characterized in that: In step S400, the cutting depth of each machining point on the fine grinding trajectory is compensated and corrected according to the inherited error field, specifically as follows: For each processing point on the fine grinding trajectory, the error value of the inherited error field S′(x,y) at that point is extracted. Based on the ratio of the available process allowance at that point to the maximum absolute error value of the inherited error field in the entire processing area, the error digestion coefficient λ at that point is determined. The value range of λ is (0,1]. The product of the error digestion coefficient and the error value of the inherited error field at that point is superimposed on the nominal cutting depth at that point to obtain the actual compensated cutting depth at that point; a fine grinding compensation machining path is generated according to the actual compensated cutting depth of each machining point, and the fine grinding machine tool is driven to perform dynamic correction of the cutting depth, so that the fine grinding process can actively offset the residual error left by the rough grinding process while eliminating its own machining error.

7. An intelligent equipment error detection system based on environmental perception, employing the intelligent equipment error detection method based on environmental perception as described in any one of claims 1-6, characterized in that: include: The surface residual error acquisition module is installed on the rough grinding machine and is used to scan and sample the workpiece surface to obtain the residual deviation data of each sampling point relative to the theoretical profile. The reference feature point measurement module is set on the rough grinding machine and the fine grinding machine respectively. It is used to measure the coordinates of at least three reference feature points on the workpiece in their respective machine coordinate systems at the end of the rough grinding stage and after the fine grinding clamping is stable. The workpiece accompanying data management module is used to encapsulate residual deviation data and reference feature point coordinates into a workpiece accompanying data package using the workpiece's unique identifier as an index, and to realize the binding and transmission of the data package and the workpiece entity during the process transfer. The error topology surface construction module is used to read the residual deviation data in the workpiece accompanying data package and fit and generate a continuous error topology surface that describes the spatial distribution of residual errors on the workpiece machining surface. The coordinate registration module is used to analytically solve the rigid body transformation relationship of the workpiece before and after the process transfer based on the coordinates of two sets of reference feature points, register the error topology surface to the coordinate system of the fine grinding machine tool, and generate the inherited error field. The error genetic evaluation module is used to calculate the cross-process error genetic coefficient based on the inherited error field, and output process flow decision instructions based on the relationship between the coefficient and a preset threshold. The compensation path generation module is used to receive the process flow decision instruction and adaptively generate a fine grinding compensation processing path based on the error values ​​of each point in the inherited error field and the available process allowance. The fine grinding CNC execution module is used to receive the compensated machining path and drive the fine grinding machine to perform machining according to the actual compensated cutting depth at each machining point.

8. The intelligent equipment error detection system based on environmental perception according to claim 7, characterized in that: The surface residual error acquisition module includes a laser displacement sensor and its drive control unit. The laser displacement sensor is installed on the side of the spindle of the rough grinding machine or on a dedicated measuring bracket, and the measuring direction is consistent with the normal direction of the workpiece surface. The drive control unit controls the laser displacement sensor to scan and sample the workpiece surface according to a preset grid path, calculates the difference between the measured height value of each sampling point and the theoretical contour height value of the corresponding point, generates a signed residual deviation dataset, and uploads the residual deviation dataset to the workpiece accompanying data management module via industrial Ethernet.

9. The intelligent equipment error detection system based on environmental perception according to claim 7, characterized in that: The workpiece accompanying data management module includes a workpiece identification unit and an accompanying data storage unit. The workpiece identification unit includes RFID readers installed at the rough grinding machine's unloading station and the fine grinding machine's loading station, respectively, which are used to automatically read the workpiece's unique identifier when the workpiece arrives at each station, triggering the writing or reading operation of the corresponding accompanying data packet. When the workpiece enters the fine grinding process, the accompanying data storage unit sends a reading completion signal carrying the workpiece's unique identifier to the industrial control computer, triggering the start of the error topology surface construction module.

10. The intelligent equipment error detection system based on environmental perception according to claim 7, characterized in that: The error genetic evaluation module also includes a grinding feedback unit. When the cross-process error genetic coefficient is not lower than a preset upper limit threshold, the grinding feedback unit sends a grinding instruction to the CNC system of the rough grinding machine via industrial Ethernet. The grinding instruction includes the location information of the out-of-tolerance area in the inherited error field and the corresponding residual deviation. After the grinding is completed, the CNC system of the rough grinding machine sends a grinding completion signal to the industrial control computer, triggering the surface residual error acquisition module to re-execute the scanning sampling and subsequent evaluation process until the cross-process error genetic coefficient drops below the preset upper limit threshold, and outputs a process flow decision instruction that allows fine grinding to proceed.

Images