Aero-engine super-air-cooled blade silicon-based ceramic core conformal constraint sizing method

Abstract

The application discloses an aero-engine super-air-cooled blade silicon-based ceramic core conformal constraint and correction method, and belongs to the technical field of ceramic core post-processing and correction for precision casting of aero-engines. The method establishes a transfer process reference when the blank is still supported by the printing platform, forms a geometric deviation and a structure risk partition through three-dimensional measurement registration, implements differential normal correction on both sides of the paired block conformal unit, and limits the sum of the contact load; under the boundary of normal keeping and tangential limited range sliding, cleaning, drying and two-side secondary photocuring are completed, and after stabilization, hierarchical release and re-measurement are performed. Thus, the surface warping, leading and trailing edge twisting, thin-wall depression, hole position and wall distance excess are inhibited, and the indentation, crack and connecting rib damage risk are reduced.

Description

Technical Field

[0001] This invention relates to the field of post-processing and calibration technology for ceramic cores used in precision casting of aero-engines, and particularly to a conformal constraint calibration method for silicon-based ceramic cores for super-air-cooled blades of aero-engines. Background Technology

[0002] Photopolymerization molding can manufacture silicon-based ceramic cores for super-air-cooled blades of aero-engines, featuring double walls, film pores, and complex internal cavities. Before secondary curing, the polymer network and particle skeleton of the core preform are not yet fully stable. Platform separation, removal of printing supports, and subsequent cleaning and drying processes alter its load boundaries and liquid state. The combined effects of polymerization shrinkage, solvent evaporation, capillary pressure, and self-weight can easily cause surface warping, leading and trailing edge twisting, thin-wall depressions, film pore axis misalignment, and changes in the distance between the double walls.

[0003] Existing rigid fixtures typically clamp the blank with a few positioning points or fixed surfaces, making it difficult to continuously fit complex curved surfaces. They also struggle to distribute different constraint forces for thin-walled areas, unsupported spans, connecting ribs, and areas adjacent to air film holes. Insufficient clamping force cannot offset deformation caused by post-processing shrinkage, while excessive clamping force may result in indentations, microcracks, or broken connecting ribs. Repeated disassembly and reassembly of the fixture before and after cleaning, drying, or secondary curing can also lead to loss of process references and secondary springback.

[0004] Therefore, there is a need for a conformal constraint correction method for silicon-based ceramic cores of super-air-cooled blades for aero-engines that can overcome the shortcomings of the existing technologies. Summary of the Invention

[0005] One objective of this invention is to propose a conformal constraint correction method for silicon-based ceramic cores of super-cooled blades for aero-engines. Addressing the issues that complex surface and internal structural dimensional distortions easily occur in photocurable silicon-based ceramic core blanks during platform removal, support disassembly, cleaning, drying, and secondary curing, and that traditional rigid fixtures struggle to simultaneously ensure conformal support, correction capability, and local structural safety, this invention proposes a method that establishes a transport process benchmark under platform support, registers the measured geometry, and partitions structural risks. It uses the relative normal displacement of paired conformal units on both sides as the local correction amount and the sum of the contact loads on both sides as the safety constraint amount. Furthermore, it ensures that normal constraint and tangential limit release penetrate through cleaning, drying, and secondary photocuring on both sides. After the quality, shrinkage, and curing dosage stabilize, the constraint is gradually released and retested. This approach reduces the risk of indentation, cracking, and damage to connecting ribs while maintaining the accuracy of hole positions, wall distances, and surface profiles.

[0006] This invention provides a conformal constraint correction method for silicon-based ceramic cores of supercooled blades for aero-engines, comprising: S1. With the blank supported by the printing platform, the process positioning part of the blank is positioned and connected to the transfer frame, and the positioning connection is maintained to complete the separation of the platform and the removal of the printing support, so as to obtain a blank to be calibrated that maintains the process reference. S2. Measure the three-dimensional geometry of the blank to be corrected and register it with the nominal model to obtain geometric deviation data. Based on the local structural risk, divide the blank to be corrected into a correction zone, a support zone, and a clearance zone. S3. Based on the correction area, support area and avoidance area, corresponding segmented conformal units are arranged on both sides of the blank to be corrected. Relative normal displacement is applied according to the geometric deviation data, and the sum of the contact loads of the pair of segmented conformal units is limited to within the upper limit of the contact load to obtain a constrained blank. S4. Maintain the alignment position of the process reference and the constrained blank, and perform cleaning, drying and secondary photocuring on both sides in sequence. The segmented conformal unit maintains its position along the local normal of the contact interface and allows the non-process reference area to slide within a limited stroke along the tangent of the contact interface. S5. After the constrained preform reaches the conditions of quality stability, shrinkage stability and secondary photocuring dosage, the segmented conformal unit is released in stages, and the geometric deviation data and damage state are re-measured to obtain the calibration result.

[0007] Optionally, S1 includes: the process positioning part is at least one of a core head, a tenon, and a preset process block; The transfer frame establishes a process reference through a positioning surface and locking components that cooperate with the process positioning part; The platform separation and printing support removal are carried out in the order of extending from the area adjacent to the transfer frame to the unsupported span area, and the transfer frame bears the load of the separated blank.

[0008] Optionally, S2 includes: the geometric deviation data includes the surface normal deviation, the torsion of the leading and trailing edges, the offset of the air film hole axis, and the double wall spacing deviation; The same batch of raw blank samples were divided into test groups according to different local wall thicknesses, unsupported spans, distances to connecting ribs, and distances to the edge of the air film hole. With the other structural parameters and contact conditions fixed, the load was applied step by step to obtain the relationship between the damage initiation load and the corresponding structural parameters of each test group. The structural parameters when the damage initiation load in the relationship is equal to the target correction load were determined as the wall thickness limit, the unsupported span limit, the adjacent distance limit of the connecting rib, and the distance limit of the air film hole edge, respectively. The target correction load was determined by the contact load when the geometric deviation of the same batch of raw blank samples with preset geometric deviations entered the nominal model size tolerance. The locations where the local wall thickness is less than the wall thickness limit, the unsupported span is greater than the unsupported span limit, the distance to the connecting rib is less than the adjacent distance limit of the connecting rib, or the distance to the edge of the air film hole is less than the edge distance limit of the air film hole are determined as avoidance zones. The locations where the geometric deviation exceeds the dimensional tolerance and can be contacted by both sides of the core are determined as correction zones. The remaining contactable locations used to disperse the contact load are determined as support zones.

[0009] Optionally, S3 includes: each of the segmented conformal units includes a partially conformal bearing block, a compliant surface contact pad disposed on the contact side of the partially conformal bearing block, a partially normal guide that defines the movement direction of the partially conformal bearing block, a displacement stop that defines the amount of advance, and a load limiter that detects and limits the contact load; Two segmented conformal units located in the correction zone are arranged opposite to each other. The contact area of ​​the segmented conformal unit located in the support zone is larger than the contact area of ​​the adjacent segmented conformal unit in the correction zone. The segmented conformal unit located in the avoidance zone is spaced apart from the surface of the blank. For the correction zone, two relatively arranged segmented conformal units are advanced respectively. When the contact load of each segmented conformal unit first reaches the corresponding preset unit probe load, the initial contact position of the segmented conformal unit is determined. The sum of the probe loads of the two preset units is less than the upper limit of the contact load of the pair of segmented conformal units, and the sum of the contact loads of the two units during the probe process does not exceed the upper limit of the contact load. For the support area, the segmented conformal unit is advanced with a probe load less than the corresponding allowable support load, and the position where the contact load first reaches the probe load is determined as the initial contact position. Furthermore, the determination of the upper limit of the contact load includes: using samples of the same batch of materials, loading them with the same contact pad material, contact area and loading direction as the corresponding conformal unit, to obtain the damage initiation load when the first indentation, crack, thin-wall indentation or connecting rib damage occurs; The damage initiation load is reduced based on the local wall thickness at the corresponding location, the unsupported span, and the distance from the connecting ribs and air film holes. For the correction zone, the reduced load is determined as the upper limit of the contact load of the pair of segmented conformal elements; for the support zone, the reduced load is determined as the allowable support load of the corresponding support zone segmented conformal elements. Furthermore, for each pair of segmented conformal elements, the relative normal displacement of the two elements is used as the local correction amount, and the sum of the contact loads of the two elements is used as the local safety constraint amount. When the geometric deviation does not reach the corresponding target range and the local safety constraint reaches the upper limit of the contact load, the advancement of the pair of segmented conformal units is stopped, and the support load of the adjacent support segmented conformal units is increased within the range not exceeding the allowable support load. When an adjacent support zone block conformal unit reaches the allowable support load, the load is shared by another adjacent support zone block conformal unit that has not reached the allowable support load, or the corresponding correction path is stopped and the geometric deviation of the unreached target section is recorded. To counteract the torsion of the leading and trailing edges, at least two pairs of segmented conformal elements are staggered along the length of the core, and relative normal displacements in opposite directions are applied to adjacent pairs of segmented conformal elements to form a torsional couple. The target range is determined by the dimensional tolerances of the nominal model.

[0010] Optionally, S4 includes: a tangential micro-slip layer is provided between the compliant surface contact pad and the local conformal bearing block, or a limit planar micro-movement pair is provided between the local normal guide and the base on which the local normal guide is installed; The displacement stop defines a local normal position, and the tangential micro-slip layer or the limiting plane micro-movement pair defines a tangential stroke limit, so that the tangential relative displacement between the process positioning part and the positioning connection of the transfer frame is zero, and the non-process reference area outside the positioning connection slides within the tangential stroke limit as the blank shrinks. Furthermore, the compliant surface contact pad and the locally conformal bearing block are respectively provided with cleaning and drainage holes, drying airflow holes and secondary curing light-receiving areas that communicate with the surface of the blank; The cleaning fluid is input from the cleaning drain hole located on the upper side along the direction of gravity and discharged from the cleaning drain hole located on the lower side. Drying gas is alternately introduced into both sides of the core through the drying airflow holes; The secondary curing light irradiates the core from both sides in a time-division manner through the secondary curing process. During the cleaning, drying and secondary curing processes, the sum of the contact loads of each pair of segmented conformal units is continuously monitored, and the sum of the contact loads is ensured not to exceed the upper limit of the contact load of the pair of segmented conformal units.

[0011] Optionally, S5 includes: acquiring the green blank quality and feature point spacing at preset time intervals; When the change in mass within two consecutive preset time intervals is not greater than the mass change threshold, it is determined that the mass stability condition has been met. When the change in the distance between feature points within two consecutive preset time intervals is not greater than the shrinkage change threshold, it is determined that the shrinkage stability condition has been met. The condition for secondary photocuring dose is determined to be met if the cumulative irradiation at each measuring point on both sides of the green blank is not less than the irradiation required for the corresponding material sample in the same batch to reach the target curing degree. The mass change threshold, shrinkage change threshold, and preset time interval are determined by the stable segment of time between the mass and the distance between characteristic points of the same batch of green blank samples. Furthermore, the graded removal of the segmented conformal unit includes first reducing the contact load of the correction segmented conformal unit, then reducing the support load of the support segmented conformal unit, and re-measuring the geometric deviation data and damage status after each level of load reduction. When the change in the retest deviation relative to the deviation before release does not exceed the springback threshold determined by the nominal model size tolerance, the retest deviation does not exceed the corresponding size tolerance, and no indentation, crack, thin-walled depression or damage to the connecting rib is detected, the retest is deemed to have passed, and the positioning connection between the process positioning part and the transfer frame is released after each level of retest has passed. When the damage is detected, regardless of whether the change exceeds the rebound threshold, the removal process is stopped and a correction result containing a failure flag, residual geometric deviation data, and damage status is output. When no damage is detected and the change exceeds the rebound threshold, the abnormal recovery branch is entered. The load before the release of this stage is restored and the drying or secondary light curing time is extended. The quality stability condition, shrinkage stability condition and secondary light curing dose condition are re-evaluated. If the condition is met, the release and retest of this stage are performed again. If the condition is not met, the constraint is maintained and drying or secondary light curing continues. When no damage is detected, the change does not exceed the rebound threshold, but the retest deviation exceeds the corresponding dimensional tolerance, the release is stopped and a calibration result containing a failure flag, residual geometric deviation data, and damage status is output. When the number of abnormal recovery attempts or the cumulative processing time reaches the stopping condition determined by the same batch of green blank samples, the stop is lifted and the calibration result containing the failure flag, residual geometric deviation data and damage status is output. When each level of retesting passes, the verification results are recorded and the calibration results containing the pass flag are output.

[0012] The beneficial effects of this invention are: 1. Establishing a transfer frame positioning connection while the blank is still supported by the printing platform, and completing the platform separation and printing support removal under the same process reference, can reduce early warping and torsion caused by sudden changes in the load-bearing boundary.

[0013] 2. Based on the measured deviation and local structural risk, paired segmented conformal units are arranged, and the relative normal displacement is used for correction and the sum of the contact loads on both sides is limited. The correction amount and the upper limit of local damage can be controlled separately. The safety of thin walls, connecting ribs and air film holes can be taken into account by supporting and sharing and correcting the torsional couple.

[0014] 3. The constraint boundary of normal holding and tangential limited slippage runs through cleaning, drying and secondary light curing on both sides, so that the green blank releases in-plane shrinkage while maintaining the shape correction position, and reduces the risk of drying obstruction, springback and dark crack by stabilization condition judgment, graded release and damage priority stop. Attached Figure Description

[0015] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of the conformal constraint correction method for silicon-based ceramic cores of super-air-cooled blades for aero-engines according to the present invention.

[0016] Figure 2 This is a control flow chart for the S4 cleaning, drying and secondary photocuring stages of the present invention. Detailed Implementation

[0017] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0018] refer to Figures 1-2 A conformal constraint correction method for silicon-based ceramic cores of supercooled blades for aero-engines, including: S1. With the blank supported by the printing platform, the process positioning part of the blank is positioned and connected to the transfer frame, and the positioning connection is maintained to complete the separation of the platform and the removal of the printing support, so as to obtain a blank to be calibrated that maintains the process reference. S2. Measure the three-dimensional geometry of the blank to be corrected and register it with the nominal model to obtain geometric deviation data. Based on the local structural risk, divide the blank to be corrected into a correction zone, a support zone, and a clearance zone. S3. Based on the correction area, support area and avoidance area, corresponding segmented conformal units are arranged on both sides of the blank to be corrected. Relative normal displacement is applied according to the geometric deviation data, and the sum of the contact loads of the pair of segmented conformal units is limited to within the upper limit of the contact load to obtain a constrained blank. S4. Maintain the alignment position of the process reference and the constrained blank, and perform cleaning, drying and secondary photocuring on both sides in sequence. The segmented conformal unit maintains its position along the local normal of the contact interface and allows the non-process reference area to slide within a limited stroke along the tangent of the contact interface. S5. After the constrained preform reaches the conditions of quality stability, shrinkage stability and secondary photocuring dosage, the segmented conformal unit is released in stages, and the geometric deviation data and damage state are re-measured to obtain the calibration result.

[0019] In this specific embodiment, S1 includes: This specific implementation takes silicon-based ceramic core blanks from the same batch that have been cured by 405 nm photopolymerization and have not yet undergone final sintering as the object. After the blanks are printed, they are still supported horizontally by the printing platform. The transfer frame adopts a rectangular closed-loop frame and is equipped with three adjustable positioning seats. The three adjustable positioning seats respectively cooperate with the positioning plane of the core head end, the side positioning surface of the tenon, and the anti-rotation surface of the preset process block to form a process positioning relationship that restricts three translational degrees of freedom and three rotational degrees of freedom. The locking part adopts a pressure plate with a force limiting knob. A silicone rubber isolation pad with a hardness of Shore A 35 is set between the pressure plate and the process positioning part. The setting value of the force limiting knob is determined by the indentation test of the process positioning part of the same batch and recorded in the transfer process card. The process reference parameter table is established according to the batch number of the green blank, the category of the process positioning part, and the number of the transfer frame. The fields include the flatness of the positioning surface, the reference spherical coordinates, the locking load of the pressure plate, the allowable positioning displacement, the allowable registration residual, and the load transfer threshold. The flatness of the positioning surface is verified by a coordinate measuring machine. The locking load of the pressure plate is determined by taking 80% of the load when the core head, tenon and preset process block samples of the same batch are pressed step by step without producing indentations. The allowable positioning displacement and allowable registration residual are determined by adding 0.01 mm to the maximum value of the results of ten repeated clamping. The load transfer threshold is determined by the minimum stable ratio before the platform completely detaches in the transfer test of three complete green blanks of the same batch. If any batch is inconsistent, the parameter table is re-established instead of using the old record. The process reference parameter table uses the batch number plus a two-digit revision number as the version identifier. Every five pieces of the same batch of green billets are transferred or any of the positioning pads, locking parts, or weighing sensors are replaced, triggering a verification. During the verification, the positioning surface is checked with standard gauge blocks, the load channel is checked with standard weights, and the empty frame clamping is repeated once. If the verification result exceeds the original threshold, the revision number is increased and recalibrated. The controller matches the parameter table according to the positioning type and compares the registration threshold according to the reference residual. When the registration residual exceeds the limit, a re-clamping command is output. The old version is only retained for traceability and is not used for new green billets. The assembly and adjustment personnel measured the center coordinates of the three positioning reference spheres in the printing platform coordinate system and registered them with the corresponding reference points in the nominal model to establish a rigid body transformation relationship from the printing platform coordinate system to the transfer frame coordinate system. ,symbol This represents the three-dimensional coordinate vector of the same spatial point in the transport frame coordinate system, with units of mm and symbol [symbol missing]. This represents the three-dimensional coordinate vector of the point in the printing platform coordinate system, in mm, with the symbol... The orthogonal rotation matrix from the printing platform coordinate system to the transfer frame coordinate system is a dimensionless quantity obtained by least-squares rigid body registration of three reference spheres. The symbol is... This represents the translation vector from the origin of the transfer frame coordinate system to the origin of the printing platform coordinate system, expressed in the transfer frame coordinate system, with units of mm. That is, let... When the coordinates of the origin of the printing platform coordinate system are zero vectors in the transfer frame coordinate system, and the registration residual does not exceed 0.03 mm, the transformation matrix, blank batch number, printing platform number and transfer frame number are written into the process reference record. After positioning is completed, the locking load of the three pressure plates is increased in increments of 0.5 N. After each level of locking, the displacement of the core end and tenon is read by a dial indicator. When the displacement of any measuring point relative to the pre-locking state exceeds 0.02 mm, the load of the corresponding pressure plate is stopped and the load is returned to the previous level. When all positioning surfaces are in contact and the displacement does not exceed 0.02 mm, the force limiting knob is locked. Then, the three reference balls are re-measured. If the difference between the re-measured coordinates after the aforementioned rigid body transformation and the coordinates before locking does not exceed 0.03 mm, the process reference is determined to be established. If it exceeds this value, the printing platform support is maintained and the positioning seat is readjusted, and the platform separation process is not entered. The controller reads the allowable positioning displacement threshold and allowable registration residual threshold from the current valid version of the process reference parameter table according to the process positioning category, and executes the dual-condition matching rule. The current valid version is triggered for verification and update every five pieces of the same batch of blanks are transferred or any positioning component is replaced. Only when all three positioning surfaces are in contact, the displacement of all positioning measuring points is not greater than the allowable positioning displacement threshold, and the registration residual of the three reference balls is not greater than the allowable registration residual threshold, will the reference valid flag be output. If any condition is not met, the corresponding positioning seat number and the amount of deviation will be output. The operator adjusts the positioning seat and re-executes the complete matching while the printing platform continues to support the blank. The platform separation adopts a path extending from the area adjacent to the transfer frame to the unsupported span area, dividing the printing base plate into a first separation zone adjacent to the transfer frame, a second separation zone located in the middle section of the blade, and a third separation zone including the unsupported span of the leading and trailing edges. Only one connecting segment with a length of no more than 20 mm is removed at a time. After removing a connecting segment, it is held for 30 seconds and the weighing sensors and printing platform support sensors set at the four corners of the transfer frame are read. The sensor range is 0 to 50 N and the resolution is 0.01 N. The weighing values ​​at the four corners are added to obtain the load value of the transfer frame, and the support values ​​of the platform are added to obtain the remaining load value of the platform. Controller according to Calculate the first The load transfer ratio after the connection segment is released, symbol Indicates the first The load transfer ratio of the transfer frame after the connection section is released is a dimensionless quantity ranging from 0 to 1, with the sign... This represents the sum of the weights at the four corners of the transfer frame after the connecting section is removed, in N (unit: N). The symbol is... This represents the sum of the remaining support values ​​of the printing platform after the connection section is removed, in N, with the symbol... The index of the connecting segment is indicated by the sequential numbering of the first, second, and third separation zones. It is a positive integer. When the denominator is less than 0.10 N, it is determined that the sensor is not properly loaded and separation is suspended. The last platform connecting segment is allowed to be released only when the load transfer ratio reaches the load transfer threshold read from the current effective process reference parameter table and the change between two adjacent readings does not exceed 0.02. The load transfer threshold calibrated in this batch of tests is 0.95. After batch change or parameter table revision, the effective value in the new version record is adopted so that the load of the separated green billet is borne by the transfer frame. The removal of printing supports continues in the same direction. First, the short supports in the area adjacent to the transfer frame are removed, then the supports in the middle section of the blade are removed, and finally the supports below the leading and trailing edges and the unsupported span are removed. For each set of supports removed, the load difference at the four corners of the transfer frame and the three reference spherical coordinates are checked. When the load difference at any diagonal exceeds 10% of the self-weight of the complete blank in the same batch, the removal is paused and the load is balanced by adjusting the leveling screws of the transfer frame. When the reference spherical coordinates drift more than 0.03 mm, the supports that have not yet been removed are retained and the positioning seat is relocked. After all supports are removed and the reference is checked and approved, a blank record to be calibrated is generated, which includes rigid body transformation relationship, locking load at each level, separation sequence, load transfer curve and reference remeasurement value. The blank to be calibrated, which maintains the process reference, is transferred to step S2 along with the transfer frame.

[0020] In this specific embodiment, S2 includes: The three-dimensional measurement module reads the blank record to be calibrated from step S1, and reproduces the coordinate system of the transfer frame using three reference spheres on the transfer frame. At the measurement station with a temperature of 23±2 ℃ and a relative humidity of 45% to 55%, a blue light stripe projection scanner is used to collect the outer surface, front and rear edges, air film vents and visible double-wall boundaries of the blank. The point spacing is set to 0.05 mm. The point cloud of a single scan simultaneously saves the point coordinates, reflection intensity, scanning direction and acquisition time. When the splicing residual of the overlapping area between two scans exceeds 0.03 mm, the area is re-acquired. The measurement module first deletes points whose reflection intensity is lower than the lower limit of the effective range of the equipment. Then, it deletes outliers with fewer than eight neighboring points in a 0.20 mm radius neighborhood. Based on the rigid body transformation relationship saved in step S1, the point cloud is uniformly transformed to the transport frame coordinate system. Then, the corresponding points of the core end positioning plane, the tenon side positioning surface and the preset process block anti-rotation surface are selected for iterative nearest point registration under the reference constraint. The reference point weight is set to five times the weight of the ordinary surface point. The registration iteration terminates when the root mean square residual changes no more than 0.002 mm for three consecutive times or reaches fifty times. When the root mean square residual exceeds 0.05 mm, the measurement is marked as invalid and the reference sphere and scan coverage are rechecked. Point cloud data cleaning is performed frame by frame according to the acquisition frame number. First, the reflection intensity is corrected by the device dark field record. Then, points that exceed the nominal model outer envelope by 3 mm are deleted. After that, the median of the distance between points in the same region is calculated for the overlapping scan frames. Points with a distance greater than three times the absolute deviation of the median are marked as abnormal points and deleted. If there are fewer than twenty valid points in any 5 mm × 5 mm grid after cleaning, the grid is added to the rescan list. The cleaned point cloud version is generated only after the rescan is completed and the stitching residual with the original point cloud does not exceed 0.03 mm. For each surface sampling point after registration, the measurement module calculates the surface normal deviation along the unit normal of the corresponding point on the nominal model. ,symbol Indicates the first The signed normal deviation of each surface sampling point is expressed in mm. A positive value indicates that the measured surface is located on the outside normal side of the nominal model. The sign is... Indicates the nominal model number The outward normal unit column vectors of corresponding points are dimensionless quantities with the sign of... This represents the transpose of the outward normal unit column vector, which is a dimensionless row vector. The superscript indicates that the transpose is the transpose of the outward normal unit column vector. The transpose operation is represented by the symbol. Indicates the first The coordinate vectors of each measured sampling point in the coordinate system of the transport frame, in mm, with the symbol... Indicates the nominal model number The coordinate vectors of corresponding points in the same coordinate system, in mm, with symbol [symbol missing]. The index of the sampling point on the surface is a positive integer. The calculation result and the dimensional tolerance at that position on the nominal model are written together into the geometric deviation record. A cross-section is established every 5 mm along the blade length. The torsion of the leading and trailing edges is determined by the signed angle between the chord direction of the measured cross-section and the chord direction of the nominal cross-section, in degrees. The offset of the film gas vent axis is determined by the shortest distance between the center and normal axis of the fitted vent point set and the nominal vent axis, in mm. The displacement of the intersection point of the measured fitted axes relative to the intersection point of the nominal axis is calculated on the vent inlet and outlet reference planes specified in the nominal model. This displacement is projected onto the positive local thickness direction from the pressure surface to the suction surface and the positive local chord direction orthogonal to it, from the leading edge to the trailing edge, to obtain the inlet and... Each outlet has two signed axis offset components; the geometric deviation record is saved according to the hole number, including two reference surface identifiers, measured and nominal intersection coordinates, two positive direction unit vectors, four signed components, corresponding tolerances and validity marks, and is transferred to step S3 along with the partitioned dataset; the double wall spacing deviation is determined by the measured spacing of the corresponding inner surface points of the pressure wall and suction wall in the same cross section along the local thickness direction minus the nominal spacing, in mm. If the number of fitting points is less than twelve or the hole opening is blocked by the support residue for more than one-quarter of the perimeter, the corresponding data will be marked as to be re-measured and will not participate in the correction area determination; Risk assessment uses samples from the same powder batch, slurry ratio, printing parameters, and initial curing state as the blank to be calibrated. The samples are grouped according to local wall thickness, unsupported span, distance from connecting ribs, and distance from the edge of the air film hole. Under the condition that the compliant surface contact pad material, contact area, loading direction, and loading rate are consistent with step S3, 0.5N is used as the first level of loading, and each level is held for 60 s. Indentation, cracks, thin-wall indentation, or connecting rib damage are detected by microscopic imaging and displacement curves. When any damage occurs for the first time, the damage initiation load and corresponding structural parameters are recorded. Take another blank sample with a preset geometric deviation from the same batch, and apply paired normal loads step by step without damage. The sum of the loads on both sides when the geometric deviation first enters the nominal model size tolerance is used as the target correction load. Perform monotonically piecewise linear fitting on the damage initiation load curves of the four types of structural parameters, and read the structural parameters where the curve is equal to the target correction load to obtain the wall thickness limit, the unsupported span limit, the adjacent distance limit of the connecting rib, and the edge distance limit of the air film hole. The calibration record is saved with the material batch number and the contact pad batch number as the joint search key. If either search key does not match, the old batch limit shall not be used. When the target correction load is between the minimum and maximum values ​​of a certain damage initiation load curve, curve interpolation is used to obtain the corresponding structural limit value. When the target correction load is lower than the minimum damage initiation load of the entire curve and there is no equivalent intersection, the boundary of the most unfavorable structural parameter within the test coverage of the curve is used as the conservative limit value and the cells exceeding the test coverage are classified into the avoidance zone. When the target correction load is higher than the maximum damage initiation load of the entire curve and there is no equivalent intersection, all cells controlled by the structural parameter are classified into the avoidance zone and the target load exceeds the safety calibration range is output. If any of the four types of curves enters the next branch and results in no correction zone, the calibration of this batch is stopped and the curve is not extrapolated. The risk calibration record also saves the sample number, contact pad hardness, effective contact area, loading rate, loading direction, load levels, holding time, damage type, damage image and curve version. The controller retrieves records according to the rule of first matching the material batch with the contact pad batch, and then matching the contact area with the loading direction. If there are two records of the same level, the one with the lower corresponding damage initiation load value is used. If no complete record is found, the relevant area is classified into the avoidance zone and the missing field is output. Extrapolation values ​​are not used to generate risk limits. The risk zoning module divides the registered point cloud into cells using a nominal surface grid of 5 mm × 5 mm. For each cell, it reads the local wall thickness, maximum unsupported span, distance to the nearest connecting rib, and distance to the edge of the nearest air film hole. If any of the following conditions are met: local wall thickness less than the wall thickness limit, maximum unsupported span greater than the unsupported span limit, distance between connecting ribs less than the adjacent distance limit of connecting ribs, or distance to the edge of the air film hole less than the edge distance limit of the air film hole, the cell is marked as a avoidance zone. The condition judgment uses the same unit value in the same batch of calibration records. For cells not marked as avoidance zones, the risk zoning module reads the surface normal deviation, the torsion of the leading and trailing edges of the section, the offset of the axis of the adjacent air film hole, and the double wall spacing deviation. If any of these effective deviations exceeds the corresponding dimensional tolerance given by the nominal model and the location has a contact channel on both sides of the core that does not overlap with the avoidance zone, it is marked as a correction zone. The remaining cells with contact channels that can distribute the correction load are marked as support zones. Cells that cannot be reliably measured or cannot be approached from both sides are incorporated into the avoidance zone. After partitioning, the region number, three-dimensional center coordinates, local normal, four types of geometric deviations, four types of structural parameters, corresponding limits, tolerances, validity indicators, and region categories of each cell are written into the partitioned dataset. Adjacent cells of the same type are connected and merged to form the boundaries of the correction zone, support zone, and avoidance zone. The partition boundaries are superimposed on the nominal model for the operator to check. Checking only allows for returning the measurement for remeasurement or changing the questionable cells to avoidance zones. It does not allow avoiding zones to be changed to correction zones. The partitioned dataset and geometric deviation data that have passed the check are transferred to step S3.

[0021] In this specific embodiment, S3 includes: The calibration controller reads the partition dataset and geometric deviation data output in step S2, and generates a block conformal unit arrangement list according to the surface boundary of each connected calibration zone. The list fields include unit number, left and right side identification, target contact center, local normal, effective contact area, initial gap, displacement stop value, load channel number and adjacent support unit number. The calibration zone sets up unit pairs with corresponding centers on both sides of the core. The support zone sets up support units with a coverage area of ​​at least 1.5 times the effective contact area of ​​adjacent calibration zone units. The boundary of the avoidance zone extends outward by 1 mm as a safety interval. No load-bearing block or compliant surface contact pad may enter this safety interval. Each segmented conformal unit uses a resin-printed and precision-machined locally conformal bearing block. The contact surface of the bearing block is generated with a nominal profile offset of 0.8 mm, and a 0.8 mm thick silicone rubber compliant surface contact pad with a Shore A hardness of 20 to 30 is bonded to it. Roller guides arranged along the local normal of step S2 are installed on the back of the bearing block. A helical microhead constitutes the propulsion mechanism, a mechanical stop ring constitutes the displacement stop, and a series miniature load sensor with a range of 0 to 20 N and a resolution of 0.01 N constitutes the load limiting component. After assembly, the lift and descent are calibrated at six load points of 0 N, 4 N, 8 N, 12 N, 16 N, and 20 N using a standard force gauge. Units with linear error or hysteresis error exceeding 0.5% of full scale are not calibrated. The controller only allows the upper limit of the contact load and the allowable support load to be no greater than 18 N and within the effective calibration range of the sensor. If the limit calculated according to the damage initiation load curve is greater than 18 N, the unit is not calibrated. If N is not found, replace it with a load sensor with a larger range while maintaining the adjacent safety support. After completing the same lift and descent calibration with a standard load covering 0 N to at least 1.1 times the required limit, update the load channel number. If the replacement and coverage calibration cannot be completed, write the location into the unreachable list and do not proceed. The unit assembly module uses the target contact center and local normal as search keys to select candidate carrier blocks from the carrier block library. The carrier blocks must have a curvature radius difference of no more than 10% of the nominal local curvature radius, an effective contact boundary that does not exceed the safety interval of the avoidance zone, and an angle of no more than 2° between the propulsion axis and the local normal. If multiple candidates meet the conditions at the same time, the one with the lower curvature radius difference value is selected first. If they are still the same, the one with the higher effective contact area value is selected. If there are no candidates, the local conformal carrier block is remade and the contact surface is re-measured. The carrier blocks in the adjacent area are not used as substitutes. The contact load limit was determined from the damage records of the same batch of specimens. For local wall thickness, unsupported span, distance between connecting ribs, and distance to the edge of the air film pore, the monotonic piecewise linear damage initiation load curve saved in step S2 was used. First, the reference damage initiation load was read at the most favorable test boundary of each curve, and then... The local damage initiation load is obtained by linear interpolation within the curve at the actual structural parameters at each location; the corresponding reduction factor is the ratio of the local damage initiation load to the reference damage initiation load of the same curve, and results greater than 1 are truncated to 1. The minimum value among the reference damage initiation loads of the four curves is taken as the reference damage initiation load. The four reduction factors are multiplied and combined to ensure that any one of the hazard factors only reduces the load limit without increasing it. This combination rule serves as a conservative weak chain rule when multiple weak factors exist simultaneously, and is saved along with the curve version, reference boundary, and interpolation interval; the controller calculation... ,symbol Indicates the first The upper limit of the contact load of the first correction unit pair or the first The allowable load-bearing capacity of each supporting unit, in N, symbol Indicates the local wall thickness reduction factor, symbol Indicates the reduction factor for unsupported spans, symbol Indicates the reduction factor for adjacent distances of connecting bars, symbol This represents the reduction factor for the distance from the edge of the air film pore. All four reduction factors are dimensionless quantities ranging from 0 to 1, with the symbol... This represents the minimum value among the four types of reference damage initiation loads under the same material batch, contact pad material, contact area, and loading direction, in N, with the symbol [symbol missing]. The index represents the block conformal element or correction element pair index, which is a positive integer. When any reference damage initiation load is missing, equal to zero, the actual structural parameters exceed the effective range of the curve, or the curve version is mismatched, the reduction factor is not calculated, and the position is remarked as the avoidance zone. During the calibration zone probe, the controller causes the left and right segmented conformal units to advance alternately in 0.01 mm increments. Each advance is held for 5 seconds and the loads on the left and right are collected. When the load of a certain unit first reaches the preset probe load of that unit, the current position is recorded as the initial contact position of that unit. The preset probe load is 5% of the upper limit of the corresponding contact load and not less than ten times the sensor resolution. The sum of the preset probe loads on the left and right is always less than the upper limit of the pair of contact loads. During the probe, if the sum of the loads on both sides of any sampling period reaches the upper limit of the contact load, the advancement stops and the unit retracts by 0.02 mm. The area where stable contact cannot be established after retraction is converted into an avoidance zone. For the support area, the segmented conformal unit is advanced with a probe load less than the corresponding allowable support load, with an advancement step size of 0.01 mm. The probe load is 5% of the allowable support load. The position where the probe load is first reached is taken as the initial contact position of the support. Subsequently, the support load is maintained at 10% of the allowable support load to eliminate one-sided suspension. The contact area, initial contact position, probe load, holding load and allowable support load of the support unit are written into the support status table. The support status table is indexed by adjacent correction units for subsequent load sharing. For the correction zone, the relative normal displacement between the two elements is used as the local correction amount, and the controller adjusts the first... Each correction unit according to Determine the unsigned local closure quantity, and follow... Implement security constraints, symbols This represents the sum of the local normal advances of the left and right units along their respective directions towards the blank, in mm and not less than zero, symbol [symbol missing]. This represents the normal advance of the left element relative to its initial contact position on the left, in mm, with the symbol... This represents the normal advance of the right-side element relative to its initial contact position on the right, in mm, with the symbol... Indicates the first The sum of the contact loads of each correction unit pair, in N, symbol This indicates the contact load of the left-hand element, in N, with the symbol... The right-side unit contact load is indicated by N. Other symbols follow the definitions in this step. The left and right advance amounts are both positive when pointing towards the blank and are adjusted differentially in 0.01 mm increments. During profile correction, the controller, according to the left and right side markings in the arrangement list, records the side containing each out-of-tolerance profile sampling point as the target side and the opposite unit as the opposite side, and reads the signed profile normal deviation and its upper and lower tolerance limits recorded in step S2. When the target point is on the left profile, the target side is the left unit and the opposite side is the right unit; when the target point is on the right profile, the two are swapped. When the deviation is higher than the upper tolerance limit, the target side unit advances 0.01 mm into the blank while the opposite side maintains stable contact with a load not lower than the probe load. When the deviation is lower than the lower tolerance limit, the opposite side unit advances 0.01 mm into the blank while the target side maintains stable contact. After each step is maintained for 60 seconds, the signed deviation is remeasured according to the same normal and registration rules in step S2. If the absolute value of the deviation does not decrease by at least 0.005 mm after the first trial advance, the advancing unit is first retreated by 0.01 mm. If the distance is mm to the position before the test advance, then the advance polarity mark of the unit pair is reversed once and the advance mapping of the target side and the opposite side is swapped. If the absolute value of the deviation is not reduced by at least 0.005 mm after two consecutive steps after reversal, the path is stopped and the unit pair number, the deviation before and after reversal and the load are written into the unreachable list. No further attempts to correct it by adding load are made. Torque control is handled separately. Establish a signed torsional displacement, sign Indicates the first The difference in left and right advance of each correction element pair under the positive direction of rotation of the chord of the uniform section is expressed in mm. A positive value indicates that the left element advances more than the right element, and a negative value indicates that the right element advances more than the left element. The zero point is the state when both left and right elements are in their respective initial contact positions. The left and right advance and element pair index in the formula follow the previous definition. Unsigned local closure quantities continue to be used for profile correction and load constraints. Signed torsional displacement is only used to generate torsional couples. After each 0.05 mm local correction is completed, hold for 60 seconds and quickly retest the corresponding noodle strip. When the retest deviation enters the target range determined by the nominal model size tolerance, lock the displacement stop. When the retest deviation has not yet entered the target range but the sum of the contact loads reaches the upper limit of the contact load, stop the advancement of the unit and select the support unit that is closest to the boundary of the correction zone and still has load margin from the support status table. Increase the support load in increments of 0.2 N, but do not exceed its allowable support load. After increasing, try again to advance the correction by no more than 0.02 mm. When the nearest supporting unit reaches the allowable supporting load but still cannot bring the deviation into the target range, the controller will transfer the load to another adjacent supporting unit that has not reached the allowable supporting load, according to the spatial distance from near to far. If the distance is the same, the supporting unit with the higher effective contact area value will be selected. After all adjacent supporting units have reached the allowable supporting load, the correction path will be stopped. The established safe support will be maintained and the residual geometric deviation that has not reached the target range, the sum of the contact load at the time of stopping, and the reason for the limitation will be recorded. The correction will not be continued by increasing the upper limit of the load by increasing the correction unit. To address the torsion at the leading and trailing edges, at least two pairs of segmented conformal elements are staggered at 20 mm to 30 mm intervals along the core length direction, ensuring that the signed torsional displacement of adjacent element pairs satisfies the following conditions. ,symbol Indicates the relationship with the first The unit with the adjacent first Signed torsional displacement of each unit pair, in mm, symbol Using the definition mentioned above in this step, the negative sign indicates that the left and right advance differences of adjacent sections are in opposite directions, thus forming a torque couple. After each displacement, the torsion is remeasured in the direction of the chord of the section and the displacement stop is locked when the angular tolerance is entered. To address the offset of the air film orifice axis, the controller reads the directional offset of the measured axis point relative to the nominal axis point on the orifice inlet reference plane and the orifice outlet reference plane, respectively. It then decomposes this offset into signed components along the thickness direction pointing from the local pressure surface to the suction surface and along the local chord direction orthogonal to it. When any component exceeds the corresponding tolerance, the controller selects a pair of candidate correction units outside the orifice edge safety interval whose combined propulsion direction is opposite to that component and has the smallest included angle. The positive component is propulsed by the positive side unit in 0.01 mm increments, and the negative component is propulsed by the negative side unit in 0.01 mm increments. The other side maintains stable contact at a level not lower than the probe load. After each step is maintained for 60 seconds, the orifice is rescanned and the axis is fitted. The propulsion side is updated according to the new offset component. When the inlet and outlet have the same sign, adjustments are made in the same direction to correct the translational offset; when the inlet and outlet have different signs, adjustments are made in opposite directions to correct the axis tilt. Neither side may exceed the air film orifice edge safety interval or increase the upper limit of the contact load. The axis correction of the air film hole is passed when both the inlet and outlet directional offsets are within their respective tolerances. If there are no candidate unit pairs with opposite composite propulsion directions and an included angle of no more than 30°, the absolute value of the out-of-tolerance component in three consecutive re-measurements has not decreased, or any candidate reaches the displacement stop or the upper limit of the contact load, the axis correction of the hole is stopped, the safe load before the last re-measurement is restored, the adjacent support unit is maintained, and the hole number, inlet and outlet residual offsets, used unit numbers and reasons for unreachability are written into the unreachable list. Correction is not continued by crossing the avoidance zone, changing the hole edge safety interval or increasing the load limit. For the deviation of the double wall spacing, the controller reads the signed value of the measured spacing minus the nominal spacing along the local thickness direction specified in step S2; when the positive deviation exceeds the upper tolerance, the correction units on both sides of the cross section advance symmetrically from their respective initial contact positions, with each side taking a step of 0.01 mm and the difference in load increment on both sides not exceeding 0.05 N, holding each step for 60 s and re-measuring the spacing of the same pair of inner surface points. Once the tolerance is reached, the displacement stop is locked; when the negative deviation is lower than the lower tolerance, no reverse tension is applied. The supporting load of the cross section and one adjacent cross section before and after it is first synchronously reduced to the preset probe load in increments of 0.2 N, and the load is held for 60 s at each level and re-measuring to eliminate the effect of reversible compression. If the deviation decreases with unloading, the unloading continues step by step until the tolerance is reached. The correction unit maintains probe contact without increasing the advance. The positive deviation branch of the double wall spacing stops advancing when any correction unit reaches the displacement stop or the upper limit of the contact load, and the positive deviation does not decrease after three consecutive re-measurements. The negative deviation branch stops unloading when the relevant support load has been reduced to the probe load but is still below the lower tolerance, the absolute value of the negative deviation does not decrease after three consecutive re-measurements, or continued unloading will cause suspension. After stopping, all are restored to the safe support state that meets the load limit and has no suspension in the most recent test. The section number, deviation symbol, residual spacing deviation, re-measurement history and the reason for the limitation are written into the unreachable list. Double wall spacing that has not entered the tolerance is not marked as corrected. Each correction path is executed from the blade root to the blade tip according to the correction zone number. When there are surface normal deviation and torsional deviation in the same section, the two sides are first executed in the same direction differential correction and the target surface position is locked. Then the adjacent unit is executed to the opposite torsional correction. Before each advance, the controller reads the remaining displacement margin and load margin of the current candidate path. In the candidate paths where both margins are positive, they are first sorted by load margin from large to small. If the load margins are the same, they are sorted by displacement margin from large to small. If both margins are the same, they are executed in ascending order according to the correction zone number. The execution result is immediately updated to the layout list to prevent subsequent paths from repeatedly occupying the same avoidance zone boundary. After all correction paths are processed, the calibration controller continuously records for 120 seconds at a sampling frequency of 10 Hz. Only when the margin of the upper limit of the corresponding contact load minus the sum of the contact loads of each pair of calibration units is not less than 0.02 N, the margin of the allowable support load minus the actual load of each support unit is not less than 0.02 N, the load drift of any channel is not greater than 0.05 N, the four types of geometric deviations specified in step S2 are all within the corresponding tolerances and the unreachable list is empty, the end steady-state verification of S3 is deemed to be passed, and the unit layout list, initial contact position, local correction amount, load limit, support sharing relationship, residual deviation and stop locking status are written into the constrained blank status file, and the transfer frame together with the constrained blank is transferred to step S4 as a whole. During the S3 end steady-state verification, if the load margin of any correction unit or support unit is less than 0.02 N, immediately freeze all advances, trigger the channel to retract the most recent 0.01 mm advance or reduce the support load by one level by 0.2 N and maintain it for 60 s. If the load drift of any channel exceeds 0.05 N and the load margin is still not less than 0.02 N, lock all stops, stop the load, and maintain it for 10 min. After completing the retraction, load reduction, or maintenance, first use the same coordinate system, scanning parameters, and registration rules of step S2 to remeasure the surface normal deviation, leading and trailing edge torsion, the offset of the signed axis between the inlet and outlet of the film air hole, and the double wall spacing deviation, and update the residual deviation record. If the remeasured deviation exceeds the tolerance, recalibrate according to the surface, torsion correction, film air hole, or double wall spacing mapping corresponding to this step. If the calibration is unattainable, write it into the unattainable list. Only when the remeasured deviation re-enters the tolerance and the unattainable list is empty will the timer start from zero and execute a complete 120 seconds. S-Verification: Each adjustment, reason for maintenance, load before and after adjustment, retested geometric deviation, and verification curve are written into the steady-state retry record. The same constrained state can be retried a maximum of two times. If after the second retry there is still a load margin of less than 0.02 N, load drift exceeds the limit, geometric deviation exceeds the tolerance, stable contact cannot be maintained after retreat, or the unreachable list is not empty, step S4 is not entered. The process positioning part and the most recent safe support without exceeding the limit are maintained. The output is a failure calibration result consisting of residual geometric deviation, unreachable list, trigger channel, number of retry and S3 steady-state failure flag.

[0022] In this specific embodiment, S4 includes: The post-processing controller reads the constrained blank status file from step S3 and installs the entire assembly onto the processing table with a liquid collection tank without releasing the positioning connection between the process positioning part and the transfer frame or changing the setting value of the displacement stop. The local conformal bearing block and the compliant surface contact pad are prefabricated with cleaning drainage holes, drying airflow holes and secondary curing light-receiving areas that communicate with the surface of the blank. When generating the layout list in step S3, the channels and light-receiving areas avoid the main bearing surface of the correction area and all avoidance areas. In this specific embodiment, a 0.10 mm thick PTFE tangential micro-slip layer is provided between the compliant surface contact pad and the locally conformal support block. The micro-slip layer is held in place by a displacement stop in the local normal direction. An elongated hole limiting structure is provided in the contact tangential plane. No micro-slip layer is provided at the connection between the process positioning part and the transfer frame, and the tangential relative displacement is kept to zero by the positioning surface. In other non-process reference areas, the compliant surface contact pad is allowed to slide tangentially relative to the locally conformal support block as the preform shrinks. The micro-slip layer is provided with through-holes corresponding to each cleaning drain hole and drying airflow hole, and a light-transmitting window is provided corresponding to the secondary curing light-receiving area. The coverage dimensions of the elongated holes and windows on both sides of the preset tangential axis are not less than the channel or the projection size of the light-receiving area plus twice the single-sided tangential travel limit plus 0.20 mm assembly allowance, so that the compliant surface contact pad does not block the medium channel or the 405 nm optical path at the positive and negative maximum sliding positions. After assembly, 0.02 mm of pressure is applied at the zero position, positive to limit position, and negative to limit position. MPa flow rate check and same batch of calibration light sheet illuminance check. If the flow rate or illuminance is lower than 90% of the zero position at any position, it is prohibited to enter the post-processing. The distance between the limit pin and the two ends of the elongated hole and the inspection results are written into the micro-slip status table. For the first In a non-process reference area, the displacement acquisition device follows... To monitor tangential microslip, the fixed end of the displacement acquisition device is mounted on a locally conformal bearing block, and the measuring end is connected to a thin marking plate on the back of the compliant contact pad. The readings at both ends when assembled and without tangential load are taken as zero points. Indicates the first The real-time relative displacement of each compliant surface contact pad with respect to its local conformal bearing block along a preset tangential axis is expressed in mm. The positive and negative signs indicate the two directions along the preset tangential axis, respectively. This indicates the unilateral tangential travel limit of the kinematic pair, determined by the difference between the effective length of the elongated hole and the diameter of the limiting pin, in mm, and stored in the micro-slip state table by unit number, with the symbol... The index of the non-process reference block conformal unit is a positive integer. In this specific embodiment, the single-sided tangential travel limit is set to 0.30 mm. The subsequent limit, return to center and displacement change rate are all judged based on the same kinematic pair and the same zero point. When the travel limit is reached, the current post-processing action is paused and the normal position is maintained, and the tangential limit dedicated recovery branch is entered. When using a limited-stroke planar micro-movement pair to replace the tangential micro-slip layer, a pair of parallel, precision-ground stainless steel sliding plates are installed between the mounting base of the local normal guide and the processing table base. A 0.10 mm thick PTFE friction-reducing sheet is sandwiched between the two sliding plates. A guide key is installed along a preset tangential axis, and adjustable stop screws are installed in both positive and negative directions, ensuring that the mounting base can only move along this tangential axis, with the effective stroke on one side also set to 0.30 mm. The fixed end of the displacement acquisition device is installed on the processing table base, and the measuring end is connected to the mounting base of the local normal guide. The zero point is defined as the position with no tangential load and equal stop margins on both sides. During assembly, the normal movement is verified to be no more than 0.01 mm at the zero position and at both limit positions. mm; The unit number, real-time displacement, single-sided stroke limit, zero point, positive and negative directions, end stop margins and normal axial movement verification results of the micro-motion pair are all written into the same micro-slip state table. Its limit determination, displacement change rate calculation and recovery conditions adopt the same rules as the previous paragraph; when any stop is reached, the current post-processing action is immediately paused, the fluid supply and light source output are set to zero and the normal position is maintained. The adjacent support unit is symmetrically unloaded according to the tangential limit-reach dedicated recovery branch. The interrupted state is only restored after the stroke limit is returned to within 80% and the displacement change rate meets the recovery conditions. If the three-level unloading still fails to return to the center or the recovery conditions are not met after accumulating 10 minutes, the unrecoverable tangential limit-reach flag and the non-passing state are output. The dedicated recovery branch for tangential to limit immediately sets the fluid supply and light source output to zero. First, it subtracts the displacement in the same direction 60 seconds ago from the current interface displacement to obtain the signed displacement difference. Then, it takes the absolute value of the entire signed displacement difference and divides it by 1 minute to obtain the interface displacement change rate, which is in mm / min. At the same time, it symmetrically reduces the load of the adjacent support units on both sides of the unit in increments of 0.1 N, but not less than 10% of their respective allowable support loads. The unit of the adjacent support loads is N. After each load reduction, it is maintained for 60 seconds and a maximum of three levels are executed to reduce the contact interface friction and release the elastic deformation of the compliant pad. When the interface displacement returns to within 80% of the stroke limit for 60 consecutive seconds and the interface displacement change rate is not greater than 0.005 mm / min, the interruption state is restored. If the three-level load reduction still fails to return to the center or the recovery conditions are not met after 10 minutes from the trigger, the process reference and normal safety constraints are maintained, cleaning, drying and irradiation are stopped and a failure abnormality flag is output. During the cleaning phase, the entire component is kept in the direction of gravity as it was printed. Cleaning fluid, compatible with the photocurable resin system and verified by the same batch of blank samples not to corrode the silicone rubber contact pads, is input through the cleaning drain hole located on the upper side along the direction of gravity, with an inlet pressure set at 0.02 MPa. The cleaning fluid is discharged from the lower cleaning drain hole and collection tank. The turbidity of the discharged fluid is collected every 2 minutes. Cleaning ends when the turbidity change does not exceed 5% of the calibration blank value in two consecutive tests and the cumulative cleaning time reaches 10 minutes. If the inlet pressure abnormally rises to 0.03 MPa or there is no continuous drainage from the lower side for more than 30 seconds, the fluid supply is immediately stopped while maintaining the process baseline and normal constraint. Reverse pulse cleaning is used with a pressure not exceeding 0.01 MPa, each pulse lasting 5 seconds with a 20-second interval, for a maximum of three pulses. After each pulse, forward fluid supply at 0.02 MPa is resumed for 30 seconds. Recovery is considered successful when the inlet pressure is below 0.025 MPa and continuous drainage from the lower side reaches 30 seconds. Cleaning continues from the interruption point. If recovery is still not achieved after three pulses or the cumulative troubleshooting reaches 10 seconds since the abnormality was triggered, the cleaning is considered complete. The liquid supply is shut off at min time, the abnormal channel, pressure and drainage history are saved, the unrecoverable cleaning abnormality mark and non-pass status are output, drying and secondary light curing are not allowed, and forced cleaning by increasing pressure is not allowed. Post-processing parameters are recorded by material batch, cleaning fluid batch, and light source number, including cleaning pressure threshold, drainage interruption threshold, drying gas dew point, temperature range, switching cycle, irradiance correction coefficient, and target dose source. The state machine records the current state, state start time, interruption time, remaining duration, abnormal flag, and cumulative irradiance, with an execution cycle of 100 ms. The controller starts in the cleaning state, with the inlet action being to start the 0.02 MPa liquid supply, and the outlet condition being that the cleaning turbidity and cumulative time simultaneously meet the requirements. The inlet action for the alternating drying state is to shut down the cleaning liquid supply and start the drying gas, and the outlet condition is that the quality stability judgment in the subsequent step S5 has valid continuous data. The inlet action for the two-sided photocuring state is to shut down the drying gas and start the current irradiation side, and the outlet condition is that the cumulative irradiance of all measuring points reaches the target dose. If the outlet condition of the previous state is not met, entering the next state is prohibited. If any load exceeds the limit or the tangential stroke reaches the limit, the state timer is frozen and the interruption position is saved. After the conditions are restored, it continues from the interruption position without repeating the already accumulated irradiance. During the drying stage, drying gas at a temperature of 25±2 ℃ and a dew point not exceeding -20 ℃ is alternately introduced into both sides of the core through the drying gas flow holes. Each side is ventilated for 60 seconds before switching sides. The inlet pressure is set to 0.03 MPa, and the exhaust side remains connected to the liquid collection tank. Before the first weighing, the external liquid supply, drainage, and drying pipelines not attached to the frame are disconnected at the calibration quick-connect joint of the transfer frame. The drain outlet is kept facing downwards and left to stand for 30 seconds without dripping for 10 consecutive seconds. The unified weighing boundary includes the blank, transfer frame, fixed pipe section attached to the frame, and detachable connectors. The transfer frame, fixed pipe section attached to the frame, and detachable connectors are weighed as a whole under no-load before assembling the blank to obtain a unified tare weight. The tare weight value is recorded according to the material batch, processing system version, and tare weight version. The weighing module is used every 10... Min records the total mass within the weighing boundary at the same disconnection position and drainage posture, and deducts the uniform tare weight of the same version to obtain the net mass of the green blank for use in steps S4 and S5. If there is still dripping, inconsistent quick connection position, or mismatched tare weight version before weighing, the current reading is marked as invalid and the continuous interval count is reset. The historical values ​​of different weighing boundaries are not compared with each other. The infrared temperature sensor synchronously monitors the core surface temperature. When the surface temperature exceeds the range of 23 ℃ to 30 ℃, the gas flow rate is reduced and the current side ventilation time is extended to avoid new warping caused by temperature difference. In the secondary photocuring stage, a surface light source with a center wavelength of 405 nm is used. After secondary curing, the light area is irradiated from both sides of the core in a time-division manner. After continuous irradiation on one side for 30 seconds, the light source is switched to the other side. After every two single-side cycles, the incident angle of the light source is changed once around the length of the transport frame so that each measuring point is covered by at least two incident directions. The irradiance sensor is placed at the edge of the light area and records the local irradiance at 1 Hz. The distance correction coefficient between the measuring point and the sensor is obtained by irradiance mapping calibration of the same batch of flat plate samples. Post-processing controller according to Cumulative irradiance at each measuring point, symbol Indicates the first The cumulative radiation dose at each measuring point during the time-sharing irradiation process on both sides is expressed in mJ / cm². 2 ,symbol Indicates the first The measuring point at the ... Irradiance after distance correction for each sampling interval, in mW / cm² 2 ,symbol Indicates the first The duration of each sampling interval, in seconds. The irradiance within parentheses is multiplied by the duration to obtain the irradiance for that sampling interval, and the whole is summed as a single term. The mW is multiplied by s to obtain mJ. (Symbol: mW) Indicates the first The total number of sampling intervals accumulated at each measuring point, which is a positive integer, with sign... Represents the measurement point index, which is a positive integer, with the sign... This represents the index of the irradiation sampling interval, which is a positive integer. The cumulative irradiation amount is written into the curing dose record with the blank batch number, measurement point number, and irradiation side as the key. Throughout the entire process of cleaning, drying, and secondary photocuring, the sum of the contact loads of each pair of calibration units and the loads of each support unit as defined in step S3 are continuously monitored at 10 Hz. When any calibration unit reaches its corresponding upper limit of contact load or any support unit reaches its allowable support load, the fluid or light irradiation action is immediately stopped, the timer is frozen, and the process reference and the original normal stop value are maintained. When a calibration unit reaches its limit, the load of the adjacent support unit is increased to share the local effect without causing other support units to reach their allowable support loads. When a support unit reaches its limit, the load of the trigger support unit is reduced in increments of 0.1 N, and the reduction is shared according to the remaining load margin of the adjacent candidate units. Each increment is maintained for 30 s without changing the tangential stroke. It is prohibited to eliminate the over-limit by loosening the process reference, expanding the tangential stroke, or increasing the medium pressure. The load over-limit interruption record saves the trigger time, post-processing status, trigger type, over-limit channel, contact load of each correction unit, load of each support unit, tangential displacement, completed duration, and cumulative irradiance. The controller constructs a candidate set consisting of adjacent support unit numbers, remaining load margin, and spatial distance. After excluding support units that have reached their limits, it selects units in descending order of remaining load margin. If the margins are the same, it selects the one with the lower spatial distance value and outputs the selected unit number and target support load. The safety parameter set sets the fluid supply and light source output to zero and updates the recovery timer every 100 ms. When the load of all previously triggered channels is below 90% of its respective load limit for 30 consecutive seconds and there is no tangential limit indicator, the freeze is lifted. The cleaning or drying after recovery continues for the remaining duration, and photocuring continues to accumulate from the saved cumulative irradiance. The freeze occurs when the candidate set is empty, any channel after load sharing still cannot be reduced to below 90% of its load limit, or the load is frozen for the second time in the same state, or the freeze lasts for 10 seconds. If the recovery conditions are still not met, the post-processing of this batch is stopped, the process reference and normal safety constraints are maintained, and the output distinguishes the load over-limit stop flag and non-pass status of the correction unit for over-limit and the support unit over-limit. After completing the predetermined cleaning, alternating drying, and two-sided time curing, the post-processing controller verifies all load channels and micro-slip states with a 100 ms execution cycle, maintains the local normal position of all segmented conformal units, saves the micro-slip displacement history, cleaning turbidity, drying temperature and quality history, cumulative irradiance at each measuring point, contact load history, and abnormal indicators, and transmits the constrained green blank state, which is still under normal constraint and allows limited tangential slip in the non-process reference area, to step S5 for stability condition determination and graded release.

[0023] In this specific embodiment, S5 includes: The stability determination module first reads all the abnormal flags saved in step S4, and uses the irreversible tangential limit flag, the load over-limit termination flag (distinguishing between the over-limit of the correction unit and the over-limit of the support unit), and the irreversible cleaning abnormal flag to form a priority access control. When any priority access control flag exists, the process reference and normal safety constraints are maintained, stability condition determination and graded release are prohibited, and a non-pass calibration result containing the abnormality category, abnormal channel or passage, current residual geometric deviation, and safety handling status is directly generated. Only when all recoverable interruptions have met the recovery conditions and the flags have been cleared, and no priority access control flag exists, does the stability determination module read the mass history, micro-slip displacement history, and curing dosage record saved in step S4, and reads the mass change threshold, shrinkage change threshold, and preset time interval according to the material batch number and processing system version. Under the condition that all segmented conformal units continue to maintain their local normal positions, the net mass of the green blank is obtained according to the external pipeline disconnection position, drainage posture, unified weighing boundary, and tare version specified in step S4. The preset time interval obtained from the calibration of the stable section of this batch is 10. The distance between the feature points is measured using six pairs of optical feature points set near the core tip, the middle section of the blade, and the leading and trailing edges. Each measurement saves the timestamp, ambient temperature and humidity, net mass value, weighing boundary mark, tare version, feature point number, distance value, and measurement validity mark. Stability determination module calculation and ,symbol Indicates the first The change in the mass of the green blank within a preset time interval, in grams, symbol [symbol missing]. Indicates the first The measured mass of the green blank is in grams, with the symbol [symbol missing]. This indicates the mass of the green blank measured previously, in grams, with the symbol... Indicates the first The maximum value of the spacing change of all feature point pairs within a preset time interval, in mm, symbol [symbol missing]. Indicates the first For feature points in the th The interval between measurements, in mm, symbol [symbol missing]. This represents the distance between the same feature point and the previous measurement, in mm, with the symbol... The index of the feature point pair is a positive integer, and the sign is... This indicates the measurement index numbered according to a preset time interval. It is a positive integer, and the preset time interval for this batch is 10 minutes. The mass change threshold, shrinkage change threshold, and preset time interval are obtained from the stable segment of the same batch of green blank samples. At least five samples from the same batch are treated with the same cleaning, drying, and secondary light curing regime as in this specific embodiment. When the mass change and the change in the distance between the maximum feature points are within three times the repeatability standard deviation of their respective measurement systems for three consecutive intervals, the upper quantile of the mass change in the stable segment is taken as the mass change threshold, the upper quantile of the change in the distance between the maximum feature points in the stable segment is taken as the shrinkage change threshold, and the shortest sampling interval that can distinguish the change in the stable segment is determined as the preset time interval. The threshold records are saved according to the material batch and the version of the processing regime. When the mass change of the blank to be calibrated does not exceed the mass change threshold for two consecutive preset time intervals, it is determined that the quality stability condition has been met. When the change of the maximum feature point spacing for two consecutive preset time intervals does not exceed the shrinkage change threshold, it is determined that the shrinkage stability condition has been met. If any mass reading is marked as invalid due to residual liquid, weighing vibration, change of weighing boundary, or mismatch of tare weight version, the consecutive intervals are recounted. If any feature point is obstructed, the constraint is maintained and the measurement line of sight is cleared, and other feature points are not substituted for the obstructed point. In the initial stabilization gate, if only the quality stability condition is not met, the alternating drying regime of step S4 is continued and weighing is performed according to the preset time interval. If only the shrinkage stability condition is not met, all local normal positions are maintained and no load is added, and the spacing of six pairs of feature points is measured according to the preset time interval. If neither condition is met, alternating drying is continued first, and the feature point spacing is measured synchronously at the same re-judgment point. All stability conditions are re-judged after each preset time interval. This initial gate processing takes precedence over the graded release and is not counted in the number of abnormal recovery times after unloading. A total of 60 processing is performed from the first gate failure. If the quality stability condition or shrinkage stability condition is not met, stop drying and irradiation, maintain the process reference and normal safety constraints, and output the failure calibration result consisting of unmet conditions, net mass history, spacing history, residual geometric deviation and initial stability timeout mark. The curing dose condition uses the minimum irradiance corresponding to the flexural modulus and residual double bond conversion rate of the same batch of material samples reaching the process target value as the target dose. The stability judgment module reads the cumulative irradiance of each measuring point on both sides of step S4 one by one. When all measuring points are not less than the target dose, it is determined that the secondary photocuring dose condition has been met. If any measuring point is insufficient, only the side of the measuring point and the corresponding light-receiving area are irradiated additionally, and the contact load and tangential stroke monitoring continue to be performed during the additional irradiation. When any measuring point starts to be irradiated additionally, the already obtained cumulative irradiance is retained, but the connection between the quality stability condition and the shrinkage stability condition is immediately changed. After the count of qualified intervals is reset to zero, the first valid weighing value and the first valid six-pair feature point spacing value are used as the new common baseline after the additional irradiation ends. The mass change and the maximum feature point spacing change are reacquired from zero for two consecutive complete preset time intervals. Only when the two types of reacquired continuous data meet their respective thresholds and the cumulative irradiation of all measuring points reaches the target dose will the graded release be initiated. If additional irradiation is initiated again during the recounting period, the count of the two types of qualified intervals will be reset to zero again and the common baseline will be rebuilt after the irradiation ends. The stability judgment result before any additional irradiation will not be used. The graded unloading process first addresses the correction zone unit pairs, then addresses the support zone unit pairs. Each unit pair in the correction zone is reduced at three levels: 70%, 40%, and 0% of the current contact load. After all correction zone levels have been retested and passed, the support zone is reduced at the same three levels. Each level is unloaded symmetrically in increments of 0.2 N and held for 2 minutes. During unloading, the connection between the process positioning part and the transfer frame is not changed. After each level is held, the normal deviation of the surface, the torsion of the leading and trailing edges, the offset of the air film hole axis, and the deviation of the double wall spacing are retested using the same coordinate system, point cloud processing parameters, and registration rules as in step S2. Indentations, cracks, thin-wall depressions, and damage to connecting ribs are also inspected using microscopic imaging. Damage re-inspection uses 10x microscopic imaging to cover the entire compliant surface contact area, connecting rib projection area, and air film hole edge area. The current level image is registered with the image before constraint in step S3 according to the reference point. A continuous dark line width of 0.03 mm is judged as a crack. A local surface depression depth exceeding the dimensional tolerance at that position is judged as a thin-wall depression. A permanent indentation depth in the contact area exceeding 0.02 mm is judged as an indentation. A discontinuity in the connecting rib outline or a relative nominal position offset exceeding the corresponding tolerance is judged as connecting rib damage. If there is a suspected dark line, the current level load is maintained and a re-image is taken using another illumination direction. The process is not to be released in the state of uncertainty. For each level of retesting, the evaluation module follows... Calculate the dimensionless rebound ratio, sign Indicates the first The percentage of the maximum rebound threshold occupancy among all retest items after level unloading is a dimensionless quantity not less than zero, with the sign... This represents the set of retest items participating in this level of rebound gate control, including linear displacement items such as surface normal deviation, air film hole axis offset, and double wall spacing deviation, as well as angular items such as leading and trailing edge torsion. The symbols are... Indicates the first The retest item in the first The unloading process has a signed deviation; the linear displacement is in mm and the angle is in degrees. This indicates that the same retest item has a sign deviation within the same unit before unloading at that level. The nominal model is the first The absolute value of the strictly positive-zero dimensional tolerance specified for each re-measurement item, the unit for linear displacement item is mm and the unit for angular item is degrees, symbol... This represents a dimensionless proportionality constant that uses a certain proportion of the corresponding dimensional tolerance as the springback threshold. In this specific implementation, we take... This means that 20% of the corresponding dimensional tolerance is used as the springback threshold, symbol... Represents the index of the repeated test item, which is a positive integer, with the sign... The unloading level index, numbered sequentially according to the three levels of the calibration zone and the three levels of the support zone, is a positive integer, with the numerator and denominator of each fraction having the same unit; the process parameter table is established according to material batch, nominal model version, and retesting system version, and the fields include... Number of each retest item , unit, source document and effective date The minimum tolerance ratio, determined by at least five samples from the same batch undergoing non-destructive unloading tests, enables the rebound gate to distinguish between stable and unstable levels. Each time the material batch, nominal model version, or retesting procedure is changed, the version number is recalibrated and increased. The controller only reads parameters from currently valid records where all three version keys match perfectly; older versions are only used for traceability. No calculation is performed if the value is zero, missing, or the corresponding tolerance record is invalid. Maintain the current level constraint and output the retest item number and tolerance data anomaly flag. After supplementing the valid tolerance that has been verified and is strictly greater than zero, retest. If supplementation is not possible, terminate and output the failure calibration result. When the dimensionless springback ratio is not greater than 1, all retest deviations do not exceed the corresponding dimensional tolerances, and no damage is detected, the level is deemed passed and the next level is entered. When any damage is detected, the dimensionless springback ratio is no longer judged, the release is immediately stopped, and the calibration result containing the failure flag, current residual geometric deviation, damage type, damage location, and unloading level is output. When no damage is detected but the dimensionless springback ratio is greater than 1, the load before unloading at this level is restored and the abnormal recovery branch is entered. After the abnormal recovery branch is restored to the load, the three conditions are re-evaluated according to the set of preset time intervals and thresholds. If the quality stability condition is not met, the alternating drying is extended and weighing continues. If the shrinkage stability condition is not met, the condition is maintained under the recovery load for 20 minutes and the distance between six pairs of characteristic points is measured. If the secondary photocuring dose condition is not met, only the corresponding side is irradiated for the insufficient measurement points. If there are more than two unmet conditions, the order of drying, constraint maintenance, and additional irradiation is followed and the condition is re-evaluated after each item is completed. When all three conditions are met but the dimensionless rebound ratio is still greater than 1, a residual stress relaxation maintenance cycle of 20 minutes calibrated by the same batch of samples is performed under the recovery load. The load is not increased or the constraint stroke is not expanded. After the maintenance is completed, the release and retest of this level are performed again. After each abnormal recovery branch is processed, the recovery reason, recovery load, processing action, extension time and retest data are saved. In this specific implementation, the condition for stopping is that the same level can be restored a maximum of twice or the cumulative extension processing time reaches 60 min. After the stopping condition is reached, if the dimensionless springback ratio is still greater than 1, the safety constraint after recovery is maintained and the failure result is output. When no damage is detected, the dimensionless springback ratio is not greater than 1, but the retest deviation exceeds the corresponding dimensional tolerance, the unloading is not continued and the failure result containing residual deviation is directly output. The calibration result data structure establishes a master record based on the unique number of the blank. The master record is associated with the load before and after each level of load release, four types of geometric deviations, springback change, damage image index, damage status, number of recovery times, cumulative processing time and judgment branch. At the end of each level, the evaluation module checks whether the measurement timestamp and unit number are complete. If any reference value before release, remeasured value after release, or damage image is missing, the current constraint is maintained and supplemented. Only after the data is complete is the pass or fail flag for that level written. After all levels of retesting in the calibration and support areas have passed, the positioning connection between the process positioning unit and the transfer frame is released. A final inspection and retest is performed on the core that has been removed from the transfer frame. The geometric deviation data and damage status are retested to obtain the calibration results. The final inspection geometric deviation data, damage status, load and springback changes at each level, stability condition judgment record, cumulative irradiation, and final pass mark are written into the calibration result file. If any final inspection deviation exceeds the tolerance or damage occurs, the final mark is changed to fail, thus obtaining a calibration result with a complete measurement chain and processing history.

[0024] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

[0025] This invention takes the blank to be corrected, which maintains the process reference, as the processing object. The deviation data obtained by three-dimensional geometric registration and the local structural risk partition are used for differential displacement control and block conformal unit arrangement, respectively. This allows the correction zone to bear the deviation correction, the support zone to distribute the load, and the avoidance zone to protect the connecting ribs and the structure adjacent to the air film hole. In this way, the continuous support of the complex surface is coordinated with the local safety constraints.

[0026] This invention uses the sum of paired contact loads as a safety control quantity, and employs boundary-through cleaning, drying, and secondary photocuring on both sides with normal position holding and tangential limited slip. When damage is detected, the release is stopped first, and recovery, re-judgment, and retry are performed when there is no damage rebound, so that post-processing shrinkage release, shape maintenance, and result verification form a closed control chain.

Claims

1. A conformal constraint correction method for silicon-based ceramic cores of super-air-cooled blades for aero-engines, characterized in that, include: S1. With the blank supported by the printing platform, the process positioning part of the blank is positioned and connected to the transfer frame, and the platform is separated and the printing support is removed while maintaining the positioning connection, so as to obtain a blank to be calibrated that maintains the process reference. S2. Measure the three-dimensional geometry of the blank to be corrected and register it with the nominal model to obtain geometric deviation data. Based on the local structural risk, divide the blank to be corrected into the correction zone, support zone and avoidance zone. The same batch of raw blank samples were divided into test groups according to different local wall thicknesses, unsupported spans, distances to connecting ribs, and distances to the edge of the air film hole. With the other structural parameters and contact conditions fixed, the load was applied step by step to obtain the relationship between the damage initiation load and the corresponding structural parameters of each test group. The structural parameters when the damage initiation load in the relationship is equal to the target correction load were determined as the wall thickness limit, the unsupported span limit, the adjacent distance limit of the connecting rib, and the distance limit of the air film hole edge, respectively. The target correction load was determined by the contact load when the geometric deviation of the same batch of raw blank samples with preset geometric deviations entered the nominal model size tolerance. The locations where the local wall thickness is less than the wall thickness limit, the unsupported span is greater than the unsupported span limit, the distance to the connecting rib is less than the adjacent distance limit of the connecting rib, or the distance to the edge of the air film hole is less than the edge distance limit of the air film hole are determined as avoidance zones. The locations where the geometric deviation exceeds the dimensional tolerance and can be contacted by both sides of the core are determined as correction zones. The remaining contactable locations used to disperse the contact load are determined as support zones. S3. Based on the correction zone, support zone and avoidance zone, corresponding segmented conformal units are arranged on both sides of the blank to be corrected. The two segmented conformal units located in the correction zone are set opposite to each other. The relative normal displacement is applied according to the geometric deviation data, and the sum of the contact loads of the two segmented conformal units set opposite to each other is limited to the upper limit of the contact load to obtain a constrained blank. S4. Maintain the alignment position of the process reference and the constrained green blank, and perform cleaning, drying and secondary photocuring on both sides in sequence. The segmented conformal units maintain their position along the local normal of the contact interface, and allow the non-process reference area to slide within a limited stroke along the tangent of the contact interface. S5. After the constrained preform reaches the conditions of quality stability, shrinkage stability and secondary photocuring dosage, the segmented conformal units are released in stages, and the geometric deviation data and damage status are re-measured to obtain the calibration results.

2. The conformal constraint correction method for silicon-based ceramic cores of aero-engine super-cooled blades according to claim 1, characterized in that, In step S1, the process positioning part is at least one of the core head, tenon, and preset process block; The transfer frame establishes a process reference through a positioning surface and locking components that cooperate with the process positioning part; The platform separation and printing support removal are carried out in the order of extending from the area adjacent to the transfer frame to the unsupported span area, and the transfer frame bears the load of the separated blank.

3. The conformal constraint correction method for silicon-based ceramic cores of aero-engine super-cooled blades according to claim 1, characterized in that, In step S2, the geometric deviation data includes the surface normal deviation, the torsion of the leading and trailing edges, the offset of the air film hole axis, and the deviation of the double wall spacing.

4. The conformal constraint correction method for silicon-based ceramic cores of aero-engine super-cooled blades according to claim 1, characterized in that, In step S3, each of the segmented conformal units includes a local conformal bearing block, a compliant surface contact pad disposed on the contact side of the local conformal bearing block, a local normal guide that limits the movement direction of the local conformal bearing block, a displacement stop that limits the amount of propulsion, and a load limiter that detects and limits the contact load. Two segmented conformal units located in the correction zone are arranged opposite to each other. The contact area of ​​the segmented conformal unit located in the support zone is larger than the contact area of ​​the adjacent segmented conformal unit in the correction zone. The segmented conformal unit located in the avoidance zone is spaced apart from the surface of the blank. For the correction zone, two relatively arranged segmented conformal units are advanced respectively. When the contact load of each segmented conformal unit first reaches the corresponding preset unit probe load, the initial contact position of the segmented conformal unit is determined. The sum of the probe loads of the two preset units is less than the upper limit of the contact load of the pair of segmented conformal units, and the sum of the contact loads of the two units during the probe process does not exceed the upper limit of the contact load. For the support area, the segmented conformal unit is advanced with a probe load less than the corresponding allowable support load, and the position where the contact load first reaches the probe load is determined as the initial contact position.

5. The conformal constraint correction method for silicon-based ceramic cores of aero-engine super-cooled blades according to claim 4, characterized in that, The determination of the upper limit of the contact load includes: Using samples of the same batch of materials, loading was performed with the same contact pad material, contact area and loading direction as the corresponding segmented conformal unit to obtain the damage initiation load when indentation, crack, thin-wall indentation or connecting rib damage first appeared. The damage initiation load is reduced based on the local wall thickness at the corresponding location, the unsupported span, and the distance from the connecting ribs and air film holes. For the correction zone, the reduced load is determined as the upper limit of the contact load of the pair of segmented conformal elements. For the support zone, the reduced load is determined as the allowable support load of the corresponding support zone segmented conformal elements.

6. The conformal constraint correction method for silicon-based ceramic cores of aero-engine super-cooled blades according to claim 5, characterized in that, For each pair of segmented conformal elements, the relative normal displacement between the two elements is used as the local correction amount, and the sum of the contact loads of the two elements is used as the local safety constraint amount. When the geometric deviation does not reach the corresponding target range and the local safety constraint reaches the upper limit of the contact load, the advancement of the pair of segmented conformal units is stopped, and the support load of the adjacent support segmented conformal units is increased within the range not exceeding the allowable support load. When an adjacent support zone block conformal unit reaches the allowable support load, the load is shared by another adjacent support zone block conformal unit that has not reached the allowable support load, or the corresponding correction path is stopped and the geometric deviation of the unreached target section is recorded. To counteract the torsion of the leading and trailing edges, at least two pairs of segmented conformal elements are staggered along the length of the core, and relative normal displacements in opposite directions are applied to adjacent pairs of segmented conformal elements to form a torsional couple. The target range is determined by the dimensional tolerances of the nominal model.

7. The conformal constraint correction method for silicon-based ceramic cores of aero-engine super-cooled blades according to claim 4, characterized in that, In step S4, a tangential micro-slip layer is provided between the compliant surface contact pad and the local conformal bearing block, or a limit planar micro-movement pair is provided between the local normal guide and the base on which the local normal guide is installed; The displacement stop defines a local normal position, and the tangential micro-slip layer or the limiting plane micro-movement pair defines a tangential stroke limit, so that the tangential relative displacement between the process positioning part and the positioning connection of the transfer frame is zero, and the non-process reference area outside the positioning connection slides within the tangential stroke limit as the blank shrinks.

8. The conformal constraint correction method for silicon-based ceramic cores of aero-engine super-cooled blades according to claim 7, characterized in that, The flexible surface contact pad and the local conformal support block are respectively provided with cleaning and drainage holes, drying airflow holes and secondary curing light-receiving areas that communicate with the surface of the blank. The cleaning fluid is input from the cleaning drain hole located on the upper side along the direction of gravity and discharged from the cleaning drain hole located on the lower side. Drying gas is alternately introduced into both sides of the core through the drying airflow holes; The secondary curing light irradiates the core from both sides in a time-division manner through the secondary curing process. During the cleaning, drying and secondary curing processes, the sum of the contact loads of each pair of segmented conformal units is continuously monitored, and the sum of the contact loads is ensured not to exceed the upper limit of the contact load of the pair of segmented conformal units.

9. The conformal constraint correction method for silicon-based ceramic cores of aero-engine super-cooled blades according to claim 8, characterized in that, In step S5, the quality of the green blank and the spacing between feature points are obtained according to a preset time interval; When the change in mass within two consecutive preset time intervals is not greater than the mass change threshold, it is determined that the mass stability condition has been met. When the change in the distance between feature points within two consecutive preset time intervals is not greater than the shrinkage change threshold, it is determined that the shrinkage stability condition has been met. The condition for secondary photocuring dose is determined to be met if the cumulative irradiation at each measuring point on both sides of the green blank is not less than the irradiation required for the corresponding material sample in the same batch to reach the target curing degree. The mass change threshold, shrinkage change threshold, and preset time interval are determined by the stable period of time between the mass and the distance between characteristic points of the same batch of green blank samples.

10. The conformal constraint correction method for silicon-based ceramic cores of aero-engine super-cooled blades according to claim 9, characterized in that, The graded removal of the segmented conformal unit includes first reducing the contact load of the correction segmented conformal unit, then reducing the support load of the support segmented conformal unit, and re-measuring the geometric deviation data and damage status after each level of load reduction. For each level of retest, if the change in the retest deviation relative to the deviation before release does not exceed the springback threshold determined by the nominal model size tolerance, the retest deviation does not exceed the corresponding size tolerance, and no indentation, crack, thin-walled depression or damage to the connecting rib is detected, the level of retest is determined to be passed, and the positioning connection between the process positioning part and the transfer frame is released after each level of retest is passed. When the damage is detected, regardless of whether the change exceeds the rebound threshold, the removal process is stopped and a correction result containing a failure flag, residual geometric deviation data, and damage status is output. For each level of retest, if no damage is detected and the change exceeds the rebound threshold, enter the abnormal recovery branch, restore the load before unloading of this level and extend the drying or secondary photocuring time, re-determine the quality stability condition, shrinkage stability condition and secondary photocuring dose condition, and if the condition is met, re-execute the release and retest of this level; if the condition is not met, maintain the constraint and continue drying or secondary photocuring. When no damage is detected, the change does not exceed the rebound threshold, but the retest deviation exceeds the corresponding dimensional tolerance, the release is stopped and a calibration result containing a failure flag, residual geometric deviation data, and damage status is output. When the number of abnormal recovery attempts or the cumulative processing time reaches the stopping condition determined by the same batch of green blank samples, the stop is lifted and the calibration result containing the failure flag, residual geometric deviation data and damage status is output. When each level of retesting passes, the verification results are recorded and the calibration results containing the pass flag are output.

Images