Method for forming a silicon-based ceramic core with super gas-cooled blade double-wall by partitioned light curing

Abstract

The application discloses a super-air-cooled-vane double-layer-wall silicon-based ceramic core partitioned light-curing forming method, belongs to the field of light-cured additive manufacturing of ceramic cores for precision casting of aero-engines, and aims at solving the problem of excessive curing of an outer wall and an open area and insufficient curing of an inner wall, a suspended section and a light-shielded area caused by unified exposure parameters. The application establishes a pair of boundary units and a three-dimensional dose influence matrix, sets a shared dose upper limit and a remaining dose budget in the same inner cavity protection zone, switches from exposure amount compensation to geometric compensation when the supplementary lap dose exceeds the budget, generates a basic exposure mask and a supplementary exposure mask, and performs layer-by-layer partitioned exposure, so that the technical effects of considering wall thickness accuracy, interlayer lap integrity, and smoothness of micro-holes and narrow flow channels are achieved.

Description

Technical Field

[0001] This invention relates to the field of photopolymer additive manufacturing of ceramic cores for precision casting of aero-engines, and particularly to a method for partitioned photopolymerization molding of double-walled silicon-based ceramic cores for super-air-cooled blades. Background Technology

[0002] Super-air-cooled blades typically incorporate double-walled structures, micropores, narrow flow channels, and connecting structures to meet the forming requirements of complex cooling channels. Silicon-based ceramic cores can be used to form these internal structures, while photopolymerization additive manufacturing can form ceramic preforms layer by layer based on the three-dimensional model of the core, making it suitable for the integral manufacturing of thin-walled and complex internal cavity structures. Current photopolymerization processes typically set uniform exposure parameters for the same slice or entire core based on slurry calibration results, and compensate for curing broadening through contour offset.

[0003] However, ceramic particles in silicon-based ceramic slurries absorb, scatter, and block incident light, resulting in different propagation paths and transmission conditions between adjacent layers on different walls, suspended sections, and connecting roots. Using uniform exposure parameters can easily increase the curing width of the outer walls and open areas, leading to excessive wall thickness, reduced micropore size, or blocked flow channels. Simultaneously, it may result in insufficient curing depth on the inner walls, connecting free sections, and suspended areas, causing insufficient interlayer overlap, localized cracking, collapse, or misalignment of the inner and outer walls.

[0004] Existing single-sided contour compensation or single-area exposure enhancement typically treats a wall independently, without including the superposition of scattered doses from the physical boundaries on both sides of the micro-orifice or narrow channel to the same negative space in the shared constraint. Therefore, when supplementing the overlapping dose on one side, it may still intrude into the target clearance between the opposite walls, and it also lacks a control mechanism to switch from energy compensation to geometric compensation after the remaining cavity dose budget is exhausted.

[0005] Therefore, there is a need for a partitioned photopolymerization molding method for double-walled silicon-based ceramic cores of super-air-cooled blades that can overcome the shortcomings of the existing technology. Summary of the Invention

[0006] One objective of this invention is to propose a partitioned photopolymerization molding method for double-walled silicon-based ceramic cores of super-air-cooled blades. Addressing the problems of over-curing of the outer wall and open areas, insufficient curing of the inner wall and suspended areas, and the intrusion of relative wall surface scattering into micropores or narrow channels when using uniform exposure parameters in existing technologies, this invention proposes a technical solution that uses partitioned slice data to construct paired boundary units, establishes a three-dimensional dose influence matrix based on slurry optical calibration parameters, sets a shared dose upper limit and remaining dose budget within the same paired boundary units, and switches from exposure compensation to contour normal displacement, additional exposure area shrinkage, or pixel redistribution geometric compensation when the additional dose exceeds the budget. This invention can limit the cumulative dose in the inner cavity protection zone while meeting the requirements for solid core curing and interlayer overlap, and performs layer-by-layer partitioned exposure through a base exposure mask and a supplementary exposure mask, thereby balancing the size of the double-walled structure, interlayer integrity, and inner cavity unobstructedness.

[0007] This invention provides a method for partitioned photopolymerization molding of a double-walled silicon-based ceramic core for super-air-cooled blades, comprising: S1. Obtain the three-dimensional model of the core and the optical calibration parameters of the silicon-based ceramic slurry. Slice the three-dimensional model into layers and divide the solid forming area and the inner cavity protection area according to the geometric features of the slices to obtain the partitioned slice data. S2. Based on the partitioned slice data, determine the entity boundaries on both sides of the inner cavity protection zone, and generate paired boundary units that include the target clearance, the target wall thickness on both sides, and the structural changes of adjacent layers. S3. Based on the optical calibration parameters, establish a three-dimensional dose influence matrix for candidate exposure pixels on the solid core zone, interlayer overlap zone and inner cavity protection zone on both sides of the paired boundary unit. Set a shared inner cavity protection dose constraint for the same paired boundary unit and determine the remaining dose budget to obtain dose constraint data. S4. Solve the partitioned exposure parameters constrained by the remaining dose budget based on the dose constraint data, and switch to geometric compensation when the interlayer overlap dose compensation by exposure amount does not meet the shared cavity protection dose constraint to obtain paired boundary control parameters. S5. Based on the paired boundary control parameters, generate a basic exposure mask and a supplementary exposure mask, perform partitioned exposure layer by layer to obtain a core blank, and obtain a double-walled silicon-based ceramic core through cleaning, degreasing and sintering.

[0008] Optionally, S1 includes: projecting the entity connectivity components of the current slice onto the previous slice and the next slice respectively to obtain interlayer connectivity markers; The ratio of the absolute value of the difference in projected area between connected components of connected entities to the sum of their projected areas is used to obtain the rate of change in area. Based on the interlayer connectivity marker, the area change rate, and the preset area change threshold, the sliced ​​solid domain is divided into an outer wall bearing area, an inner wall bearing area, a micro-connection root area, and a connection free segment or suspended area. The sliced ​​negative space domain, which is enclosed by the solid boundary and whose target width is not greater than the preset inner cavity width threshold, is divided into a micropore protection zone or a narrow flow channel protection zone. The target width is the design clearance dimension of the corresponding negative space section in the three-dimensional model. The solid forming area includes each sliced ​​solid domain partition, and the inner cavity protection zone includes the micropore protection zone and the narrow flow channel protection zone. Furthermore, before step S1, wall thickness calibration samples and pore volume calibration samples are prepared using the same material as the silicon-based ceramic slurry. The wall thickness calibration samples and the pore volume calibration samples are sequentially subjected to cleaning, degreasing and sintering to obtain the sintering shrinkage rate in the wall thickness direction and the sintering shrinkage rate in the pore volume direction. The target wall thicknesses on both sides are corrected according to the sintering shrinkage rate in the wall thickness direction, and the target clearance is corrected according to the sintering shrinkage rate in the channel direction. Steps S1 to S5 are then performed with the corrected target wall thicknesses on both sides and the target clearance.

[0009] Optionally, S2 includes: selecting boundary sampling points at a preset arc length interval on the boundary of the micropore protection zone or the narrow channel protection zone, calculating the local normal pointing to the interior of the negative space from each boundary sampling point, and determining the current local normal by the sum of the local normal vectors of two adjacent boundary sampling points at the point where the boundary curvature is discontinuous; Along the current local normal and its reverse direction, select the first entity boundary intersection point adjacent to the same negative space connected component, and pair the two entity boundary intersection points as relative entity boundaries; When the sum of the local normal vectors is zero or there is no intersection point of the entity boundary in any direction, insert a new boundary sampling point between adjacent boundary sampling points and search again; The paired boundary units are generated based on the distance between the intersection points of the two entity boundaries, the distance from each intersection point to the corresponding outer boundary of the wall, the boundary curvature, and the positional change of the corresponding boundary intersection points in adjacent slices.

[0010] Optionally, S3 includes: determining the critical curing energy, depth penetration parameter, and lateral broadening parameter by performing single-layer exposure calibration on the silicon-based ceramic slurry with different illuminance and exposure time; The incident dose is the product of the illuminance of the candidate exposure pixel and the exposure time. The attenuation of the incident dose along the forming depth direction is determined according to the depth penetration parameter. The diffusion of the incident dose in the slice plane is determined according to the lateral widening parameter. Combined with the occlusion relationship of the propagation path of the three-dimensional model, the dose contribution of each candidate exposure pixel to each voxel in the solid core zone, interlayer overlap zone and inner cavity protection zone is calculated to form the three-dimensional dose influence matrix. Furthermore, the entity core band is formed by offsetting a first bandwidth from the relative entity boundary toward the interior of the entity, the interlayer overlap band is obtained by intercepting the overlapping area of ​​the projection of the current slice entity domain and the adjacent slice entity domain according to the second bandwidth, and the inner cavity protection band is obtained by intercepting the negative space between two relative entity boundaries according to the protection bandwidth. The first bandwidth is determined based on the target wall thickness on both sides and the exposure pixel size; the second bandwidth is determined based on the slice layer thickness and depth penetration parameters; and the protection bandwidth is determined based on the target clearance and lateral widening parameters. The upper limit of the shared dose is set as the product of the critical curing energy and the safety factor, wherein the safety factor is greater than zero and less than one; Based on the three-dimensional dose influence matrix, with the constraint that the cumulative dose of each voxel in the core zone on both sides reaches the critical curing energy and the cumulative dose of each voxel in the inner cavity protection zone does not exceed the upper limit of the shared dose, the initial basic exposure pixel set and the initial exposure dose of each pixel are solved. The upper limit of the cumulative dose of each voxel in the inner cavity protection zone is taken as the scalar dose of the inner cavity protection zone. The transmission exposure of the exposed adjacent layers is calculated using the corresponding exposure record. The transmission exposure of the adjacent layers to be exposed is calculated using the planned exposure parameters of the corresponding slice. The transmission exposure contribution of the side without adjacent layers is taken as zero. Subtracting the initial exposure dose and the scalar dose generated by the transmission exposure of adjacent layers within the inner cavity protection zone from the upper limit of the shared dose yields the remaining dose budget, which has the same dimensions as the upper limit of the shared dose.

[0011] Optionally, S4 includes: performing adjacent layer superposition exposure calibration in advance using a calibration sample with the same silicon-based ceramic slurry, slice layer thickness and exposure conditions as the silicon-based ceramic slurry, to obtain the adjacent layer superposition exposure calibration result; The cumulative dose boundary value that satisfies the preset interlayer bonding load threshold in the adjacent layer superposition exposure calibration results is determined as the lower limit of the overlap dose. The difference between the lower limit of the overlap dose and the current cumulative dose of the interlayer overlap is calculated to obtain the overlap dose gap; Based on the three-dimensional dose influence matrix, candidate exposure increments for filling the overlap dose gap on both sides are calculated respectively, and the upper limit of the additional cumulative dose generated by each voxel in the same inner cavity protection zone of each candidate exposure increment is taken as the scalar additional dose. Candidate exposure increments whose scalar additional dose does not exceed the remaining dose budget are determined as allowable exposure increments, and the exposure intensity or exposure time of the corresponding area is increased according to the allowable exposure increments; Furthermore, when the scalar additional dose used to fill the overlap dose gap on either side exceeds the remaining dose budget, switch to the compensation backoff branch and resolve in at least one of the following ways: within the allowable deviation of the target wall thickness on both sides, perform normal displacement of the corresponding solid contour into the solid; Reduce the coverage area of ​​the supplementary exposure mask within the neighborhood of the inner cavity protection zone boundary; Additional doses are redistributed in ascending order of the dose influence coefficient of candidate exposure pixels on the cavity protection zone; After each rollback action, if the coverage of the entity contour or supplementary exposure mask changes, the corresponding entries of the affected paired boundary units, entity core zone, interlayer overlap zone, inner cavity protection zone and the three-dimensional dose influence matrix are updated, and the overlap dose gap and the remaining dose budget are recalculated in any case. When the cumulative dose of the solid core strip reaches the critical curing energy, the cumulative dose of the interlayer overlap strip reaches the lower limit of the overlap dose, and the scalar dose of the inner cavity protection strip does not exceed the upper limit of the shared dose, the paired boundary control parameters containing the allowable exposure increment on both sides, the contour displacement, and the mask range are generated. When the contour displacement reaches the allowable deviation boundary of the target wall thickness, there are no candidate exposure pixels available for reallocation, or the number of re-solutions reaches the upper limit of the preset number of iterations, the current slice is marked as unformable, the issuance of the exposure mask for the current slice is stopped, and the process returns to step S1 to adjust the slice layer thickness or partition boundary.

[0012] Optionally, S5 includes: writing the initial basic exposure pixel set and the initial exposure dose of each pixel into the basic exposure mask, and writing the overlap dose compensation portion corresponding to the area located on the inner wall or forming a connecting free segment or suspended segment in the solid forming area, the overlap dose compensation portion corresponding to the area forming a micro-connection root in the solid forming area, and the geometric compensation portion of the inner cavity protection band boundary into the corresponding supplementary exposure mask. Assign grayscale value, illuminance, exposure time and contour displacement to each exposure mask, and perform exposure on the same slice in the order of basic exposure mask first and supplementary exposure mask later; Furthermore, after the current slice exposure is completed, the gray value, illuminance and exposure time of each exposure mask are recorded, the contour measurement data of the current cured layer is obtained, and the contour measurement data is registered with the target contour of the current slice to obtain the wall thickness deviation and clearance deviation. The transmission exposure scalar dose generated by the current slice for the next slice is corrected based on the wall thickness deviation, the clearance deviation, and the recorded exposure parameters, and the corresponding planned exposure estimate in the budget calculation of the next slice is replaced with the corrected transmission exposure scalar dose.

[0013] The beneficial effects of this invention are: 1. By constructing the physical boundaries on both sides of the micro-orifice or narrow channel into paired boundary units with shared control, and setting a common dose limit for the inner cavity protection zone located between the two sides, the superposition of scattered doses can be taken into account when solving the exposure parameters on both sides, reducing the net intrusion caused by separate reinforcement of the relative walls.

[0014] 2. By subtracting the base exposure and the transmission exposure of adjacent layers from the upper limit of the shared dose to obtain the remaining dose budget, and switching between increasing the exposure and geometric compensation based on whether the additional dose exceeds the budget, it is possible to limit the shrinkage of micropores and the blockage of narrow channels while supplementing the dose of the interlayer overlap of the inner wall, the root of the connection and the suspended section.

[0015] 3. By decomposing the solution results into a base exposure mask and supplementary exposure masks for different forming areas, and configuring exposure and contour displacement parameters for each mask, the partitioned dose constraint can be transformed into layer-by-layer executable forming control, improving the consistency of double-layer wall thickness, the integrity of interlayer connection, and the retention of the inner cavity after sintering. Attached Figure Description

[0016] 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 partitioned photopolymerization molding method for the double-walled silicon-based ceramic core of the super air-cooled blade of the present invention.

[0017] Figure 2 This is a flowchart of the adaptive switching between exposure compensation and geometric compensation based on the remaining dose budget constraint in S4 of the present invention. Detailed Implementation

[0018] 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.

[0019] refer to Figures 1-2 A method for partitioned photopolymerization molding of double-walled silicon-based ceramic cores for super-air-cooled blades, including: S1. Obtain the three-dimensional model of the core and the optical calibration parameters of the silicon-based ceramic slurry. Slice the three-dimensional model into layers and divide the solid forming area and the inner cavity protection area according to the geometric features of the slices to obtain the partitioned slice data. S2. Based on the partitioned slice data, determine the entity boundaries on both sides of the inner cavity protection zone, and generate paired boundary units that include the target clearance, the target wall thickness on both sides, and the structural changes of adjacent layers. S3. Based on the optical calibration parameters, establish a three-dimensional dose influence matrix for candidate exposure pixels on the solid core zone, interlayer overlap zone and inner cavity protection zone on both sides of the paired boundary unit. Set a shared inner cavity protection dose constraint for the same paired boundary unit and determine the remaining dose budget to obtain dose constraint data. S4. Solve the partitioned exposure parameters constrained by the remaining dose budget based on the dose constraint data, and switch to geometric compensation when the interlayer overlap dose compensation by exposure amount does not meet the shared cavity protection dose constraint to obtain paired boundary control parameters. S5. Based on the paired boundary control parameters, generate a basic exposure mask and a supplementary exposure mask, perform partitioned exposure layer by layer to obtain a core blank, and obtain a double-walled silicon-based ceramic core through cleaning, degreasing and sintering.

[0020] In this specific embodiment, S1 includes: In this specific embodiment, a surface projection curing device with a wavelength of 405 nm is used to process the same batch of silicon-based ceramic slurry. The solid phase of the slurry is fused silica powder with a median particle size of 3.2 μm and a solid phase volume fraction of 52%. The device pixel size is 50 μm and the nominal slice layer thickness is 50 μm. The controller establishes a molding task record with the core part number, slurry batch number, device light source number and calibration version number. Before step S1, five wall thickness calibration samples and five pore calibration samples are printed with this batch of slurry. The same cleaning, degreasing and sintering process as the core is completed in sequence. The length and pore measurement values ​​of the dimensions before degreasing and after sintering are written into the shrinkage calibration record. Controller according to and Calculate the sintering shrinkage rate in the wall thickness direction and the sintering shrinkage rate in the pore direction, and according to... and Generate a pre-compensated target size for slicing, where, This represents the sintering shrinkage rate in the wall thickness direction obtained from the same batch of wall thickness calibration samples. It is a dimensionless quantity, and the median value of five valid samples is taken. In this specific embodiment, it is calibrated to be 0.142. This indicates the gauge length of the wall thickness calibration sample before degreasing, in mm, and is derived from coordinate measuring machine (CMM) measurements. This indicates the dimensions after sintering at the same gauge length, in mm. This represents the sintering shrinkage rate in the pore direction obtained from the same batch of pore calibration samples. It is a dimensionless quantity and is calibrated to 0.136 in this specific embodiment. This indicates the pore diameter or groove width of the calibrated sample before degreasing, in mm. This indicates the sintered dimensions corresponding to the aperture or groove width, in mm. In the slice model, the first... The pre-compensation target wall thickness on the side, in mm. Indicates the index of the first or second side of a pair of boundaries, with a value of 1 or 2. The three-dimensional model gives the first Design wall thickness after side sintering, in mm. This represents the pre-compensation target clearance in the slice model, in mm. This indicates the design clearance after sintering, in mm. When the shrinkage rate of a single sample deviates from the median value of five samples by more than 0.015, the controller removes the abnormal sample and makes a replacement sample in the same batch. If there are fewer than four valid samples, the slice generation task is stopped. The shrinkage calibration record is bound to the molding task through the slurry batch number and the sintering furnace curve version number. Subsequently, S1 to S5 read the pre-compensated target wall thickness and the pre-compensated target clearance and no longer use the uncompensated design value. The slicing module reads the 3D model of the core after shrinkage pre-compensation, unifies the model to the right-hand Cartesian coordinate system, repairs isolated triangular faces smaller than one pixel and normal flips, calculates the intersection line of the plane and the triangular mesh along the equipment forming direction with a layer thickness of 50 μm, performs closed-loop merging on the intersection line with an endpoint distance of no more than 10 μm, and rasterizes each layer's closed contour into a pixel occupancy map. The pixel record includes the layer number, connected component number, entity or negative space marker, design wall side, contour point number, and pre-compensation dimension version number. For the current number Each entity connected component of the layer, the slicing module is respectively connected to the first... Layer and first When the connected components of a layer are projected along the forming direction, an inter-layer connectivity marker is written when the projection intersection contains at least four pixels and the proportion of the intersection area to the area of ​​the lower of the two connected components is not less than 0.10. When there are one-to-many or many-to-one projections in the same adjacent layer, the one with the largest intersection proportion is selected first. If the intersection proportions are the same, the one with the smallest projection centroid distance is selected as the corresponding component. For each adjacent side with a valid corresponding component, the following steps are taken: Calculate the rate of change of the area on one side, and according to... Obtain the rate of change of the area of ​​the current component, where, This indicates the current slice layer number, with values ​​ranging from 1 to the total number of layers. This indicates the adjacent index, and takes the value -1 or +1. This indicates the set of adjacent indices that actually exist in the current layer and have valid corresponding components. This represents the projected area of ​​the current entity's connected components, in mm. 2 And obtained from pixel occupancy statistics, Indicates the first The projected area of ​​the corresponding connected component of the selected layer, in mm. 2 , Indicates the first Rate of change of area of ​​adjacent side layers This represents the maximum value of the rate of change of area on each side; both are dimensionless quantities and range from 0 to 1; the first or last layer is represented only by... Calculate the side that exists in the middle. When the value is empty, write the flag indicating no inter-layer correspondence and do not proceed to the area threshold determination. This specific implementation stores a preset area change threshold in the partitioning parameter record and sets it to 0.18, and exhaustively classifies it according to a unique priority: the 3D model topology requires inter-layer continuity but... When a component is empty, crosses different wall labels, or cannot be uniquely selected according to the aforementioned rules, a re-slicing state is written and the component is prohibited from continuing to flow. Among the remaining components, those connected to adjacent layers, with an area change rate greater than 0.18 and a minimum connection width to the main wall not greater than 0.30 mm are classified as fine connection root regions. Those with only a single-sided interlayer connectivity marker and a continuous length greater than 0.50 mm after leaving the connection root are classified as connection free segments or suspended regions. All remaining entity pixels that do not meet the first two entity category conditions but have valid 3D model labels are classified as outer wall bearing regions or inner wall bearing regions according to inner and outer wall labels. This includes double-sided connectivity with an area change rate not greater than 0.18, double-sided connectivity with a connection width greater than 0.30 mm, single-sided connectivity with a continuous length not greater than 0.50 mm, and cases where the 3D model allows no interlayer connectivity at the end. Therefore, there are no unclassified entity pixels. The negative space processing module performs Euclidean distance transformation on each connected component of the negative space enclosed by the entity boundary, extracts twice the minimum boundary distance along the center line of the connected component as the target width of the current layer, and checks it against the pre-compensated target clearance of the corresponding negative space section of the 3D model. When the difference between the two is greater than one pixel, the pre-compensated target clearance of the 3D model is used as the standard and a geometric check mark is written. In this specific implementation, the preset inner cavity width threshold is set to 0.80 mm. The negative space with a closed section and a target width of no more than 0.30 mm is divided into a micropore protection zone. The negative space with an extension length of no less than three times the target width along the slice and a target width greater than 0.30 mm but not greater than 0.80 mm is divided into a narrow flow channel protection zone. The slicing module writes the outer wall bearing area, inner wall bearing area, micro-connection root area, and connection free segment or suspended area into the solid forming area layer. It writes the micro-hole protection area and narrow flow channel protection area into the inner cavity protection area layer. For each pixel, it saves the layer number, region category, connected component number, adjacent layer correspondence number and signed distance from the solid boundary. After forming the partition slicing data and the corresponding geometric version number, it is passed to step S2. If any closed negative space cannot establish a consistent connected component correspondence in two adjacent layers, the layer is marked as needing to be re-sliced ​​and the step is re-executed with a layer thickness of 25 μm. When steps S2, S3, S4, or S5 return a state of pairing failure, dose prior exceeding limit, unformable, no mapping across layers, continuous low confidence in measurement, or size action conflict, S1 uniformly reads the failure feedback object. This object includes the source step, failure layer number, negative space connectivity component number or failure paired boundary cell number, failure arc coordinates, failure reason, current local layer thickness, budget exceeding limit, core zone dose gap, side wall thickness deviation, net deviation, measurement confidence, cross-layer mapping status, cumulative contour displacement, minimum mask retention width, number of encrypted samplings, number of retests, number of resolutions, action history, and geometry version number. S1 uses local layer thickness selection, boundary sampling density, solid partition boundary pixel set, and negative space protection boundary pixel set as reslicing action variables, with the control objective of eliminating failure cells and minimizing the number of adjusted layers, and the constraints of the pre-compensation target wall thickness lower limit, the pre-compensation target net limit, and the three-dimensional model topology invariance. S1 retains the complete action history of S2, S3, S4, and S5 written to the unified feedback object and executes a unique retry order according to the source: S2 If pairing fails, first reduce the local layer thickness and then increase the boundary sampling density; S3 If the core band is unreachable or the dose prior exceeds the limit, first reduce the local layer thickness and then reconstruct the partition boundary; S4 If compensation fails, resolve according to at least one of the following methods: within the allowable deviation of the target wall thickness on both sides, perform normal displacement of the corresponding entity contour towards the interior of the entity, reduce the coverage of the supplementary exposure mask in the neighborhood of the inner cavity protection zone boundary, or reallocate the additional dose in ascending order according to the dose influence coefficient of the candidate exposure pixels on the inner cavity protection zone, and sequentially prioritize the contour displacement in the allowable direction, supplementary mask shrinkage, and pixel reallocation. Only when all the aforementioned actions reach the limit and are still not feasible, change the slice layer thickness or partition boundary; S5 If there is no mapping or size action conflict between cross-layer units, only perform 25 once on the failed unit cover layer and the next layer. If the local slice is re-sliced ​​at μm and the mapping is reconstructed from S1, and there is still no mapping or conflict after reconstruction, the task is frozen and transferred to manual review; if the continuous measurement low confidence state is still below 0.90 after S5 completes a re-permutation and retest, geometric retry is not triggered, the task is directly frozen and transferred to manual review; each allowed retry action uses the source step, failure layer number, failure cell number, failure reason, current local layer thickness and geometric version in the unified failure feedback object as status input, increases the corresponding retest number or re-solve number and updates the geometric version; In addition to the previously specified direct freezing of the S5 continuous measurement low-confidence state, the reslicing controller first changes the layer thickness of the failed unit's overlay layer and the two layers above and below it from 50 μm to 25 μm, keeping the remaining layer thicknesses unchanged, and re-executes the S1 to failure source step; if the regenerated data still causes the same S2, S3 or S4 source unit to fail, the sampling interval of the corresponding entity boundary and negative space protection boundary is densified to 25 μm according to the failure arc coordinates or budget overlimit voxel projection, but the pre-compensation target clearance or 3D model topology must not be changed, and then the execution proceeds from S1 to S4 sequentially; if the same failed object from S2, S3 or S4 is still infeasible after two consecutive rounds, the task is frozen, and the S5 source object is processed according to the previously specified one-time reconstruction or direct freezing rules, and all outputs include a unified failure feedback object and a manual review record of the entire version history; The resulting partitioned slicing data uses the task number of the super-air-cooled blade double-walled silicon-based ceramic core partitioned photopolymerization molding method as the primary key, and records the corrected target wall thickness on both sides, target clearance, solid molding area, inner cavity protection area, interlayer connectivity marker, area change rate, local layer thickness and feedback iteration number, and includes the slurry batch number, optical calibration version number and shrinkage calibration version number, as the only input for step S2 to determine the relative solid boundary.

[0021] In this specific embodiment, S2 includes: The boundary pairing module reads the partition slice data formed by S1, and takes the connected component of each micropore protection zone or narrow channel protection zone as the processing unit. It selects sampling points along its closed boundary at a preset arc length interval of 50 μm. The sampling point record includes the layer number, negative space connected component number, arc length sequence number, two-dimensional coordinates, adjacent entity region category and initial displacement from the corresponding sampling point of the previous layer. For the Each boundary sampling point, the module according to Calculate the local normal vector, and then, by moving one pixel along the obtained normal vector, check the sign of the signed distance to unify its direction to point towards the interior of negative space. Indicates the first Each sampling point points to a unit local normal within the negative space, which is a dimensionless two-dimensional vector. This represents a fixed operator that rotates a two-dimensional vector counterclockwise by 90°. This indicates the coordinates of the next boundary sampling point, in mm, and is derived from the current negative space contour. This represents the coordinates of the previous boundary sampling point, in mm. Indicates the index of sampling points within the same closed boundary, increasing in arc length; The boundary pairing module calculates the discrete boundary curvature by dividing the included angle between adjacent line segments by the average arc length of the two sampling points. The curvature is set to be no greater than 2.0 mm. -1Emit search lines along the current local normal and its reverse direction, when the absolute value of curvature is greater than 2.0 mm. -1 When curvature discontinuity is formed, according to Determine the current search normal, where, The normalized search normal used at the point of curvature discontinuity is represented by a dimensionless two-dimensional vector. This represents the unit local normal of the previous sampling point. This represents the unit local normal of the next sampling point. If the magnitude of the vector sum is less than 0.05, it is determined to be a zero vector and enters the encrypted sampling branch. The module moves pixel by pixel along the search normal in both directions. It selects the first entity boundary intersection encountered after leaving the current negative space connected component, and only accepts entity boundaries adjacent to the same negative space connected component. If no intersection is found after the search distance exceeds 1.5 times the pre-compensated target clearance, the midpoint of the arc length is inserted between the current sampling point and the adjacent sampling point and the normal is recalculated. If no valid intersection is found after three consecutive encryptions, the subsequent processing of the current layer is stopped. The source step S2, the failed layer number, the negative space connected component number, the first and last coordinates of the failed arc segment, the reason for failure (missing boundary pairing), the current local layer thickness, the number of encryption samplings, the action history, and the geometry version number are written into the unified failure feedback object agreed upon in S1. Other inapplicable fields are set to zero or empty set and S1 is returned. S1 first changes the local layer thickness from 50 μm to 25 μm and re-executes S1 to S2. If it still fails, the boundary sampling interval is changed from 50 μm to 25 μm and it is executed again. If it still fails the second time, the task is frozen and a manual review record is output. The intersection points of the two valid entity boundaries are denoted as follows: and The boundary pairing module follows The local clearance is obtained, and the corresponding outer boundary of the wall is searched from the two intersection points along the direction away from the negative space, according to... and The wall thicknesses of the two target sides are obtained, among which, Indicates the first The clearance of local targets in pairs, in mm. This represents the coordinates of the intersection point of the entity boundaries in the positive direction of the search, in mm. This represents the coordinates of the intersection point of the entity boundaries in the opposite direction of the search normal, in mm. This indicates the target wall thickness on the positive side, in mm. This represents the coordinates of the intersection point of the outer boundary of the wall on the positive side, in mm. This indicates the target wall thickness on the opposite side, in mm. This indicates the coordinates of the intersection point of the outer boundary of the wall on the opposite side, in mm. The module uses the nearest arc length and the same negative spatial connectivity component sign to find the corresponding intersection point in adjacent slices, according to... Record the structural changes of adjacent layers, among which, Indicates the first The maximum boundary position change of each paired position relative to the entire piece, in mm. and They represent the first Coordinates of the positive and negative intersection points of paired boundaries of the layer, in mm. and These represent the coordinates of the corresponding intersection points in the previous layer, in mm. This indicates the slice layer number, with zero structural variation between adjacent layers of the first layer. When the order of the intersection points on both sides of adjacent sampling points is consistent, the local clearance difference is no greater than 0.10 mm and the normal angle is no greater than 15°, the module merges them into the same pair of boundary units and saves the unit number, layer number, negative space connectivity component number, sequence of intersection points on both sides, target wall thickness on both sides, pre-compensated target clearance, boundary curvature, structural change of adjacent layers and entity region category. When any index exceeds the merging threshold, a new pair of boundary units is established to avoid crossing micro-hole corners or flow channel bifurcation. After completing one layer of pairing, the module verifies whether the boundary arcs of the same inner cavity protection zone are all covered by one and only one paired boundary unit. For overlapping units, the first unit is retained in ascending order of local clearance value. Uncovered arcs are resampled and encrypted to obtain a set of paired boundary units indexed by layer number and negative space connectivity component number. Then, based on the corresponding negative space connectivity component number of S1, the module projects the arcs of each unit in the current layer to the next layer along the forming direction, ensuring that only arcs in the same corresponding negative space whose projected arc length is not less than 0.50 of the lower value of the arc length of the two units and whose average distance between the two intersection points is not greater than 0.15. A cross-layer mapping is established for mm-level units; when a current unit corresponds to multiple next-level units, all targets are retained and weights are recorded in a normalized manner according to the projected overlap arc length; when multiple current units correspond to one next-level unit, all sources are retained; one-to-one, one-to-many, many-to-one, and no-correspondence states are written into the relation type respectively; the cross-layer unit mapping table stores the current layer number and unit number, the next layer number and target unit number set, relation type, overlap weight, negative spatial connectivity component corresponding number, and two-layer geometric version number, together with the paired boundary unit set, geometric version number, and partition category as the common input for step S3 to establish the zone and dose influence matrix and step S5 to place the feedback quantity.

[0022] In this specific embodiment, S3 includes: The dosage calibration module reads the slurry batch number from the S1 task record and prepares a single-layer stepped exposure sample with the same 50 μm layer thickness, 405 nm light source, and resin bath temperature of 25±1 ℃ as the actual molding. The illuminance is set to 8 mW / cm². 212 mW / cm 2 and 16 mW / cm 2 The exposure time ranged from 0.4 s to 3.0 s, with eight levels. Three samples were prepared for each level, and the curing depth and lateral widening relative to the design boundary were measured using a cross-sectional microscope. Measurements with a deviation of more than 10% from the median value of the same level were discarded to form optical calibration samples. For each target file, the module is configured according to... Calculate the incident dose and use Fit the critical curing energy and depth penetration parameters, where, Indicates the first The incident dose for each calibration range, in mJ / cm². 2 , Indicates the first The measured illuminance for each standard setting, in mW / cm². 2 And obtained by a measured optical power meter, This indicates the exposure time, measured in seconds (s). Indicates the calibration file index. This indicates the curing depth of the corresponding sample, in mm. This represents the penetration depth parameter of the slurry, in mm, and is obtained by least-squares fitting. This represents the critical curing energy of the slurry, expressed in mJ / cm². 2 Furthermore, this is a converted value of the fitting intercept. When the fitting determination coefficient is lower than 0.98, the current batch molding is stopped and recalibrated. The lateral broadening parameter is obtained by fitting the distance between the cured boundary of the linewidth stepped sample and the optical mask boundary. In this specific embodiment, the effective lateral dose kernel is set as a Gaussian kernel, and the fitting standard deviation is... Write it into the optical calibration record, where, This indicates the lateral width expansion parameter under the current slurry batch and equipment light source conditions, in mm, with an effective range of 0.03 mm to 0.20 mm. Calibration records are also saved. , , Temperature, layer thickness, pixel size, and light source number are all factors that, if any change exceeds the recording tolerance, will trigger recalibration. The strip construction module reads paired boundary units one by one from S2, offsets the relative entity boundaries towards their respective entities with a first bandwidth to form the core strips on both sides, and extracts the overlapping portion of the projections of the current slice entity domain and the adjacent slice entity domain according to a second bandwidth to form an interlayer overlap strip. Furthermore, it extracts an inner cavity protection strip along the boundary in the negative space between two relative entity boundaries. This specific implementation follows... , and Three bandwidths are determined, among which, This represents the first bandwidth, in mm. This indicates the exposure pixel size, in mm, and in this specific embodiment, it is 0.05 mm. This represents the lower of the target wall thicknesses on both sides of a pair of boundary elements, in mm. This indicates the second bandwidth, in mm. The thickness of the slice is indicated in mm, and in this specific embodiment it is 0.05 mm. This represents the depth penetration parameter, in mm. This indicates the protection bandwidth, in mm. This indicates the pre-compensation target clearance for this unit, in mm. This indicates the lateral width parameter, in mm. The dose matrix module establishes voxel centers within each zone with a 25 μm planar spacing and a 25 μm depth spacing, for voxels that may affect the corresponding voxels and whose planar distance does not exceed [a certain value]. Depth distance not exceeding Establish candidate relationships based on the exposure pixels, and according to Calculate the dose effect coefficient, where, Indicates candidate exposure pixels voxels The dimensionless dose influence coefficient ranges from 0 to 1. Indicates the banded voxel index. Indicates the candidate exposure pixel index. This represents the propagation coefficient obtained from the ray occlusion check of the 3D model. It is a dimensionless quantity and is taken as 1 when there is no occlusion, 0 when completely occluded by the entity, and the corresponding calibrated transmittance when passing through a solidified thin layer. This represents the distance between the slice planes of the pixel center and the voxel center, in mm. This represents the non-negative distance between the two along the forming direction, in mm. and All were taken from the same optical calibration version; The matrix module writes each non-zero coefficient into a sparse 3D dose influence matrix according to the layer number, paired boundary unit number, band category, voxel number, and pixel number, for a given pixel dose vector. according to Cumulative voxel dose, of which, Voxel representation The cumulative dose, in mJ / cm 2 , Voxels in the three-dimensional dose effect matrix The corresponding dimensionless row vector, This represents a column vector composed of the doses of each candidate exposure pixel, with each component in mJ / cm². 2Furthermore, it is determined by grayscale, measured illuminance, and exposure time, and matrix entries less than 0.001 are set to zero to control sparsity; The controller sets the shared dose limit to [value]. ,in, Indicates the first The upper limit of the shared dose of the lumen protection zone corresponding to each paired boundary unit, in mJ / cm². 2 , Indicates the paired boundary cell index of the current layer. The safety factor is a dimensionless quantity, and in this specific embodiment, it is calibrated to 0.78 based on the condition that the microporous calibration sample does not adhere. This represents the critical curing energy, expressed in mJ / cm². 2 The shared dose limit applies to both sides of the entity boundary of the same unit without being allocated separately; Before any slice exposure, the controller first establishes a full-layer planned transmission prior in ascending order of layer number: For adjacent layers that have not yet been exposed, the upper confidence limit of the transmission dose obtained in the same batch of double-layer transmission calibration under the combination of the equipment's 3.0 s exposure limit, the current optical calibration illuminance limit, and the most unfavorable shielding coefficient of the corresponding guard band is used as the sole conservative planned value; this value is directly calculated from the calibration samples without depending on the mask to be determined in the current or future layers, thereby eliminating circular dependence, and a planned version number is generated by the slurry batch, optical calibration version, transmission calibration version, layer number, and cross-layer unit mapping version; the conservative value is used in the first round of solution, and after the current layer exposure measurement is completed, only the measured correction value of S5 is used to replace the contribution of the exposed side, and the unexposed side always maintains the same planned value and does not perform mutual substitution; if there are fewer than 30 sets of corresponding valid calibration samples, the upper confidence limit cannot be obtained, or the planned version is inconsistent with the current geometric version, the flow of that layer is stopped and S1 is returned according to the dose prior missing; The initial exposure solver uses constrained non-negative least squares to obtain the initial base exposure pixel set and the initial exposure dose of each pixel, with all pixel doses being non-negative and not exceeding the dose at the device's maximum exposure duration of 3.0 s as the boundary. Each unit has a protective band voxel for its internal cavity. The sum of the measured upper transmission bound of the already exposed adjacent layer mapped to its position via cross-layer unit mapping and the conservatively planned upper transmission bound of the adjacent layer to be exposed is taken as a fixed term. And simultaneously satisfy the following during the solution process ,in, Indicates the adjacent layer at the 1st Unit voxel The transmitted dose fixed term formed at the location is expressed in mJ / cm. 2 , Indicates the banded voxel index. Represents paired boundary cell indices. This indicates the base exposure of the current layer at voxels. The cumulative dose generated at the site is expressed in mJ / cm². 2 , This indicates the upper limit of the shared dose for this unit; the solution also ensures that each voxel in the core band of the two sides satisfies And among the solutions that satisfy all constraints, the solution with the smallest total pixel dose is selected first; if there is a voxel in the core band that cannot reach the critical curing energy or the protection constraint is infeasible after adding the fixed term of the adjacent layer, the generation of the slice mask is stopped immediately, the core band dose gap or dose prior limit is written according to the unified failure feedback object and returned to S1, and the dead state is no longer handed over to S4. For the Each unit, the controller takes the upper limit of the cumulative dose generated by the initial base exposure that has passed the protection constraint for each voxel in the inner cavity protection zone as the unit. Take each voxel fixed item The upper boundary as The contribution of one side of an adjacent layer is zero; only when According to Determine the remaining dose budget, where, Indicates the initial base exposure at the 1st The scalar dose generated within the inner cavity protection zone is measured in mJ / cm². 2 , This represents the scalar dose generated by transmission exposure of adjacent layers, in mJ / cm². 2 , This represents the non-negative residual dose budget with the same dimensions as the upper limit of shared dose, expressed in mJ / cm². 2 If the sum of the above is greater than It must not continue to flow after being truncated to zero, but instead writes the dose prior limit exceeded state and returns to S1; The dose constraint data is stored by layer number and paired boundary unit number, including the voxel index of the solid core zone, interlayer overlap zone and inner cavity protection zone, the version of the three-dimensional dose influence matrix, the critical curing energy, the shared dose upper limit, the initial basic exposure pixel set of the current layer, the initial pixel dose, the measured transmission dose of the exposed side, the planned transmission dose of the side to be exposed, the plan version number, the cross-layer unit mapping version and the remaining dose budget. A budget depletion flag is set for any unit with a remaining dose budget of zero and the interlayer overlap zone has not yet met the dose lower limit. Only when the core zone curing constraint, the protection constraint containing the fixed items of adjacent layers and the remaining budget check all pass, is the complete dose constraint data bound to the current layer number, paired boundary unit number, geometric version number and matrix version number used as the input for solving the partition exposure parameters in step S4.

[0023] In this specific embodiment, S4 includes: Before solving for the zonal exposure parameters, the calibration module pre-performs adjacent layer superposition exposure calibration using calibration samples with the same silicon-based ceramic slurry, slice layer thickness, and exposure conditions as the stated sample, to obtain the adjacent layer superposition exposure calibration results, specifically at 10 mJ / cm 2 Up to 60 mJ / cm 2 Five double-layer stacked samples were prepared for each cumulative dose step. Before degreasing, three-point bending loading was performed on the interface between adjacent layers. The lowest cumulative dose at which the interface failure load reached 80% of the median failure load of the complete layer samples in the same batch was determined as the lower limit of the overlap dose. The sample batch, cumulative dose, failure load, loading speed, number of effective samples and calibration version were saved in the record. The solver uses the paired boundary cell numbers of the current layer, the minimum dose of the core bands on both sides, the minimum dose of the overlap band, the maximum dose of the guard band, the shared dose limit, the remaining dose budget, the initial base exposure pixel set and its dose of the current layer, the geometry version number, and the matrix version number output in step S3 to form the static solution state; S5 only feeds back the measured transmission dose, the wall thickness deviation on both sides, the clearance deviation, the measurement confidence, and the feedback geometry version number of the previous exposed layer after cross-layer mapping, and forms the current layer's running state; since the first layer has no previous exposed layer, the measured transmission contribution, the wall thickness deviation on both sides, and the clearance deviation are uniquely initialized to zero, and the measurement confidence is initialized to 1.0. 0, only retain the conservative planned transmittance of the side to be exposed given by S3; S4 entry checks the current layer number, paired boundary unit number, cross-layer mapping version, geometric version and matrix version of the static state and the running state item by item. If any is inconsistent, the solution stops and returns to S1; the action vector includes the pixel increment dose on both sides, the exposure time increment, the contour normal displacement in the allowed direction, the basic mask reduction set, the supplementary mask shrinkage width and the pixel redistribution order. The goal is to minimize the total new dose required to make up for the overlap dose gap, and the constraints are the core band solidification lower limit, the overlap band bonding lower limit, the protective band sharing upper limit, the target wall thickness allowable deviation and the equipment exposure upper limit; For each pair of boundary units The interlayer overlap zone, the solver takes the minimum value of the current voxel cumulative dose as the interlayer overlap zone. and in accordance with Calculate the overlap dose gap, where, Indicates the first The overlap dose gap per unit, in mJ / cm² 2 , This indicates the lower limit of the overlap dose obtained from the exposure calibration of adjacent layers, in mJ / cm². 2 , This represents the lower bound of the current cumulative dose within the interlayer overlap zone of this unit, expressed in mJ / cm². 2 When the gap is zero, no additional exposure will be added to the unit; When the overlap dose gap is greater than zero, the solver extracts sub-matrices from the three-dimensional dose influence matrix of the interlayer overlap zone and the same inner cavity protection zone on both sides of the unit, and jointly solves the candidate exposure increments on both sides, so that the dose generated by the candidate increments of each voxel in the overlap zone on each side is not less than the corresponding overlap dose gap, with the objective of minimizing the number of newly added pixels and the sum of the dose increments on both sides; the candidate vectors on both sides are... The merging increment vector of the assembly into the same unit, where, and They represent the first The incremental dose column vector of candidate exposure pixels on the first and second sides of each unit, with each component in mJ / cm². 2 , This represents the transpose of a vector. This represents the combined incremental dose column vector representing the shared budget used by both sides, with each component in mJ / cm². 2 ; Controller according to Calculate the scalar additional dose generated by the combined candidate increments from both sides on the same lumen protection zone, where, Indicates the first The scalar additional dose after merging the two sides of each paired boundary unit is expressed in mJ / cm². 2 , This indicates the third step in the three-dimensional dose influence matrix of step S3. The dimensionless submatrix formed by merging candidate pixels in the inner cavity protection band. This represents the merged incremental dose column vector defined in the previous section. The max operation indicates obtaining the upper bound of each voxel component of the protection band in the matrix product; after reallocating pixels on either side, reassembly is performed. And calculate ; When both sides combine scalar additional dose Not greater than the same remaining dose budget given in step S3 When the candidate exposure increments on both sides are jointly determined as the allowable exposure increment, the device prioritizes increasing the exposure duration of the corresponding pixel while maintaining the illuminance within the calibrated illuminance range. Once 3.0 s is reached, the corresponding grayscale illuminance is further increased. The controller writes the allowable exposure increments on both sides, the corresponding pixel set, the exposure intensity increment, and the exposure duration increment into the paired boundary control parameters. If the merging is successful... Greater than Even if the increment on either side does not exceed the limit, it will immediately enter the compensation backoff branch; The compensation rollback first selects the action direction based on the runtime dimensional deviation: only when the wall thickness deviation on either side is positive or the clearance deviation is negative, and moving inward into the solid will not cause the wall thickness on the other side to fall below the allowable lower limit or the clearance to exceed the allowable upper limit, is it allowed to move the corresponding solid outline along the internal normal of the solid by one pixel; when the wall thickness deviation is negative or the clearance deviation is positive, inward movement on that side is prohibited, and the allowable increment is preferentially redistributed to the corresponding wall core zone boundary pixels within the existing target solid domain to increase its effective curing width; when the action directions of different deviation requirements conflict, the measurement confidence is lower than 0.90, or all allowable actions exceed the shared budget, the unformable state is directly written; in this specific implementation, the cumulative normal displacement on one side is limited to 3% of the pre-compensated target wall thickness and not greater than 0.10 mm, and the local clearance and the wall thickness on both sides are recalculated based on the solid boundary after the movement. If the movement causes the predicted wall thickness on either side to fall below the design wall thickness allowable lower limit, the action is canceled; The second back-off action starts from the boundary between the supplementary exposure mask and the inner cavity protection zone and reduces the coverage area by one pixel width, while retaining the supplementary exposure area of ​​at least two pixels on both sides of the center line of the micro-connection root area. The third back-off action sorts the candidate pixels in ascending order according to the maximum dose influence coefficient of the inner cavity protection zone. Under the premise of still covering the overlap dose gap, the incremental dose is redistributed to the pixels ranked first. Under the same influence coefficient, the pixels with high dose contribution value to the interlayer overlap zone are selected first. Each time contour displacement, mask shrinkage, or pixel reassignment is performed, the controller updates the intersection sequence of affected paired boundary units, the solid core zone, the interlayer overlap zone, the cavity protection zone, and the 3D dose influence matrix entries. When the contour moves into the solid, the basic exposure pixel set is regenerated using the Boolean intersection of the new solid domain and the old basic exposure domain. The Boolean difference between the old basic exposure domain and the new solid domain is used as the basic mask reduction set, and the grayscale of pixels in the reduction set is forced to zero. For pixels located within the new solid domain that are adjacent to the boundary of the moved cavity protection zone by one pixel width and do not belong to the basic mask... If the retained boundary pixels of the deleted set are found to require a positive exposure increment to bring the core or overlap band to the lower limit after recalculation, then the positive exposure increment is defined as the geometric compensation part of the inner cavity protection band boundary and included in the calculation of the merged increment vector on both sides, the scalar additional dose, and the same remaining budget; the deleted area must not be written into the positive exposure supplement mask; the controller recalculates the base dose, overlap gap, merged additional dose on both sides, and remaining budget based on the new set, and the update record is saved in pairs with the geometric version number and the dose matrix version number, and it is prohibited to use the old matrix or the old base pixel set to evaluate the new contour; The rollback solution iterates in the order of allowable contour displacement, base mask reduction and supplementary mask shrinkage, and pixel redistribution, up to a maximum of eight times. Convergence is determined when the cumulative dose of each voxel in the solid core bands on both sides is not lower than the critical curing energy, the cumulative dose of each voxel in the interlayer overlap band is not lower than the overlap dose lower limit, the combined additional dose on both sides does not exceed the same remaining budget, and the scalar dose of the inner cavity protection band is not higher than the shared dose upper limit. This generates a positive exposure set containing the updated base exposure pixel set and dose, the base mask reduction set, the allowable positive exposure increments on both sides, and the geometric compensation portion of the inner cavity protection band boundary. The parameters include the set of light pixels and dose, contour displacement, positive exposure supplement mask range, geometric culling record mask for traceability only with zero dose for all pixels, positive exposure order, geometric version number, and dose matrix version number. Among these, the geometric compensation part of the inner cavity protection zone boundary has already included the combined additional dose and remaining budget on both sides. The geometric culling record mask only encodes pixels deleted by the base mask, does not participate in device projection, and is not included in the additional dose or remaining budget. The aforementioned pairwise boundary control parameters containing all fields are used as the sole input for generating the complete set of masks in step S5. When the cumulative contour displacement reaches the target wall thickness allowable deviation boundary, the mask has been reduced to the minimum retention width, there are no candidate exposure pixels available for reallocation, the size deviation action direction conflicts, or all constraints are still not met after eight re-solutions, the controller writes the current slice into the unformable state record, stops issuing any exposure mask for the slice to the device, and returns to step S1 according to the unified failure feedback object, carrying the source step S4, failure layer number, failure unit number, failure arc coordinates, failure reason, current local layer thickness, budget over-limit, wall thickness deviation on both sides, clearance deviation, measurement confidence, cumulative contour displacement, minimum mask retention width, number of re-solutions, action history and geometry version number. Prioritize adjusting the local layer thickness from 50 μm to 25 μm and then re-divide the solid forming area and the inner cavity protection area.

[0024] In this specific embodiment, S5 includes: The mask generation module reads the paired boundary control parameters output by S4 according to the layer number, and only uses the basic exposure pixel set and dose that are consistent with the final geometry version and have been resolved to generate the basic exposure mask, and sets the grayscale of the pixels in the basic mask deletion set to zero; the allowable overlap dose compensation pixels of the inner wall, the free or suspended segments of the connection, and the root region of the fine connection are written into the corresponding positive exposure supplement mask, and the boundary geometry compensation part of the cavity protection zone determined by S4 that is located in the new entity domain and has been included in the shared budget is written into the cavity boundary geometry compensation positive exposure mask; the Boolean difference set generated by the contour shrinkage is only written into the pixels. A geometric culling recording mask with a constant dose of zero must not be written into any positive exposure supplement mask, and the device's list of distributed recording masks must explicitly exclude such recording masks. Before generating a mask, the module must re-verify that the basic pixel set, all positive exposure supplement pixel sets, geometric version number, and dose matrix version number are consistent, and recalculate the core band lower limit, overlap band lower limit, side-merged additional dose, and guard band shared upper limit. If any item fails, the distribution will be stopped. Each mask pixel records the mask category, paired boundary unit number, gray value, calibrated illuminance, exposure time, contour displacement, whether it participates in projection, geometric version number, and dose matrix version number. Equipment controller based on Convert the permissible exposure dose into equipment instructions, where, Represents pixels Exposure dose, in mJ / cm 2 , This represents the full grayscale illuminance obtained from the current light source calibration, in mW / cm². 2 , Represents pixels The 8-bit grayscale value is a dimensionless integer ranging from 0 to 255. Represents pixels Exposure duration, in seconds. This indicates the pixel index in the current mask. If the calculation time exceeds 3.0 s, the grayscale will be increased within the calibrated illumination range and the calculation will be recalculated. Pixels that still exceed the limit will be returned to S4 for reallocation. For the same slice, the controller first projects the base exposure mask and waits for the resin tank to settle for 0.5 seconds. Then, it projects the positive exposure supplement mask in the order of inner wall or suspended section compensation, micro-connection root compensation, and inner cavity boundary geometry compensation. The inner cavity boundary geometry compensation mask only contains the retained boundary positive exposure pixels within the new entity domain that have been recalculated by S4 and included in the shared budget. The contour shrinkage deletion part is still implemented by the base mask reduction and zero grayscale. The geometry reduction record mask is only archived and does not enter the projection queue. Exposure is only allowed when the geometry version number and dose matrix version number in the control parameters are consistent with the current slice task. If any positive exposure supplement mask is empty, the mask is skipped but the sequence record is retained, thereby avoiding the supplement dose from changing the curing boundary of the base exposure in advance. After the current layer exposure is completed, the forming platform first descends to detach from the scraper and then moves to the sealed measurement position on the side of the resin tank. The measurement position is equipped with a fused silica transparent window parallel to the current layer and a dispersive confocal profile measurement head arranged along the normal of the window. The inner side of the window is separated from the main slurry by an isolation scraper. A transparent replacement liquid, prepared with the same batch of resin base liquid and free of ceramic powder, is used to locally replace the uncured slurry within a thickness of 0.30 mm above the surface of the cured layer at a pressure not exceeding 0.02 MPa, so that the measurement optical path does not pass through the 52% solid phase slurry. The controller performs coordinate compensation based on the refractive index, window thickness, and axial offset stored in the same transparent window and replacement liquid calibration block. The current cured layer profile point cloud is acquired from two opposite scanning directions at a sampling interval of 25 μm. The actual gray value of each mask, the illuminance fed back by the optical power meter, the start and end times, the platform position, and the anomaly code are recorded. The point cloud has an outlier radius of 0.10. After eliminating flying points according to the rules of mm and minimum neighbor number 6, rigid nearest point registration is performed with the current slice target contour according to the paired boundary unit number, and the measurement confidence is calculated based on the effective point ratio and coincidence error in bidirectional scanning. The deviation calculation is only performed if the confidence is not lower than 0.90. If it is lower than 0.90, the replacement is repeated and the measurement is repeated once. If it is still lower than 0.90, the next layer exposure is stopped. The source step S5, the failed layer number, the failed unit number, the failure reason (continuous measurement low confidence), the measurement confidence, the number of retests (1), and the current geometry version are written into a unified failure feedback object and returned to S1. S1 processes it according to the direct freeze and manual review rules. After the measurement is passed, the replacement liquid is recovered, the same batch of slurry is used to replenish the liquid, and the platform is returned to the material laying position to execute the next layer of material laying. After registration, the module re-acquires the two side entity boundaries and the outer boundary of the wall on the normal sampling line of each pair of boundary elements, according to... and Calculate the wall thickness deviation and the clearance deviation, where, Indicates the first Side wall thickness deviation, in mm. A positive value indicates that the measured wall thickness is greater than the pre-compensated target wall thickness. Indicates the first The measured wall thickness of the current cured layer, in mm. Indicates the first Side pre-compensation target wall thickness, in mm. This indicates the indices of the two side walls, and the values ​​are either 1 or 2. This indicates the current clearance deviation of the cured layer, expressed in mm. A positive value indicates that the measured clearance is greater than the pre-compensated target clearance. This indicates the measured clearance, in mm. This indicates the target clearance for pre-compensation, in mm. Transmission dose correction module per pair of boundary units Execute independently to recover the actual dose of each pixel from the actual grayscale, illuminance, and exposure time records of this layer. and in accordance with and The actual scalar doses of the base exposure and supplementary exposure for the protective band of this unit were obtained respectively, and then... Estimate the corrected transmission dose of the current layer for the next slice; where, Represents paired boundary cell indices. Indicates the current slice layer number. Indicates the first The set of voxel indices for the protective band within each unit cavity. This indicates the voxel index. and These represent the sets of actual exposed pixel indices in the final base exposure mask and all supplementary exposure masks of this layer, respectively. Indicates pixel index, Indicates the number of pixels in the current matrix version voxels The dimensionless dose influence coefficient, Represents pixels The actual dose is obtained by multiplying the measured illuminance by the exposure time, and the unit is mJ / cm². 2 , and They represent respectively to The base exposure scalar dose and supplementary exposure scalar dose, obtained by summing the values ​​of all voxels and taking the maximum value, are both in mJ / cm². 2 , Indicates the next slice The corrected transmission dose used in the unit budget calculation is in mJ / cm². 2 , This indicates the intercept of interlayer transmission calibration within the same batch, in mJ / cm. 2 , and represent the dimensionless regression coefficients of the base exposure dose and the supplemental exposure dose, respectively. , and This represents the dimensional deviation coefficient fitted by the interlaminar transmission calibration sample, with units of mJ·cm. -2 ·mm -1 , , and All measurements are derived from the current layer contour and are in mm. Indicates the same first The upper limit of the shared dose per unit, in mJ / cm² 2 The clip operation ensures that the correction value of this unit is not less than zero and does not exceed its shared dose limit; In the formula for front-end transmission dose This function represents the cutoff point, used to restrict the regression model output between the lower bound of zero dose and the upper bound of shared dose for the same pair of boundary units. It takes zero when the regression model output is less than zero, the upper bound of shared dose when it is greater than the upper bound of shared dose for that unit, and retains the regression model output in all other cases. Both input and output units are mJ / cm². 2 ; The calibration source of the above regression model is the double-layer transmission calibration sample of the same batch of slurry. The calibration record should include at least the basic exposure dose, supplementary exposure dose, curing wall thickness on both sides, curing clearance and the measured transmission dose of the next layer summarized by paired boundary units. No less than 30 sets of records are used for least squares fitting with intercept and 20% of the records are reserved for verification. When the root mean square error of the verification is no more than 5% of the upper limit of the corresponding shared dose, the coefficient version is enabled. If it exceeds 5%, the current molding task is immediately frozen, the next layer mask is prohibited from being issued and the transmission sample of the same batch of slurry is recalibrated. If the recalibration still fails, the manual review record is output. The budget of the next layer must not be updated with the unverified model or the undefined rollback output. When the current layer number When the number of layers equals the total number of layers in the molding task, S5 writes the final layer flag after completing the final layer contour measurement, dimensional deviation calculation, and exposure record archiving. Indicates the current layer number, which is an integer from 1 to the total number of layers; then skips the transmission dose placement to the next layer, cross-layer unit mapping check and next layer budget update. If there is no next layer after the last layer, it is a normal termination state. Do not write cross-layer unit no mapping or other failure feedback, and directly proceed to subsequent cleaning, degreasing and sintering processes. The controller will only assign the source unit of the current layer to the source layer according to the S2 cross-layer unit mapping table if the current layer is not the last layer, the transmission regression model has been validated, and the profile measurement confidence level is not lower than 0.90. The corrected transmission dose and size deviation are assigned to the next slice target cell; one-to-one relationships directly replace the planned transmission estimate of the target cell; for one-to-many relationships, for safety reasons, the upper bound of the corrected transmission dose from the same source and the same set of size deviations are copied to each target cell without being reduced by weight; many-to-one relationships add up the corrected transmission doses from all sources and truncate them at the upper limit of the shared dose of the target cell; the wall thickness deviation and net clearance deviation on both sides are selected from the source values ​​with the largest absolute values ​​and their signs are retained; when the source in a many-to-one relationship gives opposite action directions for the same size, a size action conflict is written; when there is no effective target mapping or the two layers of geometry are inconsistent, no mapping is written for the cross-layer cell; both stop the next layer mask delivery, write the source step S5 and the corresponding mapping status into the unified failure feedback object and return to S1; S1 is only allowed to be executed once. If local reslicing and mapping reconstruction at μm fails, the task is frozen and transferred to manual review. If mapping succeeds, the next layer's runtime state, composed of the target cell number, source cell number set, relation type, mapping version, corrected transmission dose, wall thickness deviation on both sides, clearance deviation, measurement confidence level, geometric version number, and matrix version number, is sent to S3 and S4. The remaining dose budget for the next layer is only recalculated by S3 using this corrected transmission dose and passed as the static solution state from S3 to S4. S5 does not generate or pass the remaining dose budget. S4 executes fixed decisions based on the deviation sign: when the wall thickness deviation is positive or the clearance deviation is negative, the contour is allowed to move inward or shrink the exposure domain only if the other dimension does not exceed the limit; when the wall thickness deviation is negative or the clearance deviation is positive, inward movement is prohibited, and the remaining budget in the static state after recalculation in S3 is used preferentially to redistribute the dose to the corresponding wall core zone boundary pixels within the original target entity domain; if the absolute value of the wall thickness deviation on either side exceeds 0.10 mm or the absolute value of the clearance deviation exceeds 0.08 mm... This decision must be made when the exposure is in mm, and the next layer of exposure must not be increased unconditionally. After layer-by-layer exposure, the core blank is rinsed in the same batch of diluted cleaning solution with low pressure for 120 seconds. Then, clean gas at a pressure not exceeding 0.15 MPa is used to alternately blow the micropores and narrow channels until no visible suspended matter is visible in the outflow liquid. The blank is dried at 40 ℃ for 4 h, and then the temperature is increased to 200 ℃ at 0.5 ℃ / min and held for 2 h, increased to 450 ℃ at 0.3 ℃ / min and held for 3 h, and increased to 600 ℃ at 0.5 ℃ / min and held for 2 h to complete degreasing. When the rate of quality decline in the furnace exceeds the upper limit allowed by the calibration curve of the same batch, the heating is stopped until the quality change stabilizes. The degreased preform was heated to 1200℃ at 2℃ / min in air atmosphere and held for 2 hours before being cooled in the furnace. The sintering regime used the same furnace curve version as the S1 shrinkage calibration sample. After cooling, the wall thickness of the double-layer wall, the diameter of the micropores and the clearance of the narrow flow channel were re-measured using a go / no-go gauge and an industrial CT. The sintered dimensions were calculated back according to the design coordinate system used in S1. When the measurement results met the tolerance of the three-dimensional model, a double-layer wall silicon-based ceramic core forming file was generated, which included the forming task number, slurry batch, mask version, exposure record, cleaning and degreasing sintering curve and final dimension record. The shrinkage compensation data used in this step comes from the wall thickness calibration sample and the pore calibration sample prepared with the same material as the silicon-based ceramic slurry before step S1. Both samples are subjected to the same cleaning, degreasing and sintering process as the core blank to obtain the sintering shrinkage rate in the wall thickness direction and the sintering shrinkage rate in the pore direction. The target wall thickness and target clearance on both sides are corrected respectively. The final dimensions are only accepted after the furnace curve version is checked to be consistent with the shrinkage calibration version in S5. The final core and its molding file are associated with all geometry, dosage, mask, feedback and heat treatment records from S1 to S5 using the task number of the super air-cooled blade double-wall silicon-based ceramic core partition photopolymerization molding method. The double-wall thickness, micropore size and narrow flow channel clearance are all measured after sintering as acceptance values.

[0025] 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.

[0026] This invention uses a three-dimensional core model, the structural relationship between adjacent slices, and the optical calibration parameters of the slurry as input. It organizes the solid boundaries on both sides of the double-walled inner cavity into paired boundary units and utilizes a three-dimensional dose influence matrix to uniformly calculate the cumulative dose received by the solid core zone, interlayer overlap zone, and inner cavity protective zone. This ensures that the compensation process for solid curing is constrained by the same negative space dose budget. Therefore, the adjustment of exposure parameters is no longer limited to a single wall surface but simultaneously considers the combined influence of the relative wall surfaces on the target clearance.

[0027] This invention incorporates a trigger switching mechanism for energy compensation to geometric compensation: when the additional dose generated by supplementing the interlayer overlap dose remains within the remaining dose budget, compensation is achieved through exposure adjustment; when the additional dose exceeds the remaining dose budget, further exposure increases are stopped, and contour displacement into the solid interior, shrinkage of the additional exposure area, or pixel redistribution of the additional dose are employed. This structural improvement maintains the curing requirements of the solid region and interlayer overlap region while protecting the incurable areas within the cavity.

Claims

1. A method for partitioned photopolymerization molding of a double-walled silicon-based ceramic core for super-air-cooled blades, characterized in that, include: S1. Obtain the three-dimensional model of the core and the optical calibration parameters of the silicon-based ceramic slurry. Slice the three-dimensional model into layers and divide the solid forming area and the inner cavity protection area according to the geometric features of the slices to obtain the partitioned slice data. S2. Based on the partitioned slice data, determine the entity boundaries on both sides of the inner cavity protection zone, and generate paired boundary units that include the target clearance, the target wall thickness on both sides, and the structural changes of adjacent layers. S3. Based on the optical calibration parameters, establish a three-dimensional dose influence matrix for candidate exposure pixels on the solid core zone, interlayer overlap zone and inner cavity protection zone on both sides of the paired boundary unit. Set a shared inner cavity protection dose constraint for the same paired boundary unit and determine the remaining dose budget to obtain dose constraint data. S4. Solve the partitioned exposure parameters constrained by the remaining dose budget based on the dose constraint data, and switch to geometric compensation when the interlayer overlap dose compensation by exposure amount does not meet the shared cavity protection dose constraint to obtain paired boundary control parameters. S5. Generate a basic exposure mask and a supplementary exposure mask based on paired boundary control parameters, perform partitioned exposure layer by layer to obtain a core blank, and obtain a double-walled silicon-based ceramic core through cleaning, degreasing and sintering.

2. The method for partitioned photopolymerization molding of double-walled silicon-based ceramic core for super-air-cooled blades according to claim 1, characterized in that, S1 includes: Project the connected components of the current slice onto the previous slice and the next slice respectively to obtain the inter-layer connectivity marker; The ratio of the absolute value of the difference in projected area between connected components of connected entities to the sum of their projected areas is used to obtain the rate of change in area. Based on the interlayer connectivity marker, the area change rate, and the preset area change threshold, the sliced ​​solid domain is divided into an outer wall bearing area, an inner wall bearing area, a micro-connection root area, and a connection free segment or suspended area. The sliced ​​negative space domain, which is enclosed by the solid boundary and whose target width is not greater than the preset inner cavity width threshold, is divided into a micropore protection zone or a narrow flow channel protection zone. The target width is the design clearance dimension of the corresponding negative space section in the three-dimensional model. The solid forming area includes each sliced ​​solid domain partition, and the inner cavity protection zone includes the micropore protection zone and the narrow flow channel protection zone.

3. The method for partitioned photopolymerization molding of double-walled silicon-based ceramic core for super-air-cooled blades according to claim 2, characterized in that, S2 include: Boundary sampling points are selected at preset arc length intervals on the boundary of the micropore protection zone or the narrow channel protection zone. The local normal pointing to the interior of the negative space is calculated for each boundary sampling point. The current local normal is determined by the sum of the local normal vectors of two adjacent boundary sampling points at the point where the boundary curvature is discontinuous. Along the current local normal and its reverse direction, select the first entity boundary intersection point adjacent to the same negative space connected component, and pair the two entity boundary intersection points as relative entity boundaries; When the sum of the local normal vectors is zero or there is no intersection point of the entity boundary in any direction, insert a new boundary sampling point between adjacent boundary sampling points and search again; The paired boundary units are generated based on the distance between the intersection points of the two entity boundaries, the distance from each intersection point to the corresponding outer boundary of the wall, the boundary curvature, and the positional change of the corresponding boundary intersection points in adjacent slices.

4. The method for partitioned photopolymerization molding of double-walled silicon-based ceramic core for super-air-cooled blades according to claim 1, characterized in that, S3 includes: By performing single-layer exposure calibration on the silicon-based ceramic slurry with different illuminance and exposure time, the critical curing energy, depth penetration parameter and lateral broadening parameter were determined. The incident dose is the product of the illuminance of the candidate exposure pixel and the exposure time. The attenuation of the incident dose along the forming depth direction is determined according to the depth penetration parameter. The diffusion of the incident dose in the slice plane is determined according to the lateral widening parameter. Combined with the occlusion relationship of the propagation path of the three-dimensional model, the dose contribution of each candidate exposure pixel to each voxel in the solid core zone, interlayer overlap zone and inner cavity protection zone is calculated to form the three-dimensional dose influence matrix.

5. The method for partitioned photopolymerization molding of double-walled silicon-based ceramic core for super-air-cooled blades according to claim 4, characterized in that, The core zone of the entity is formed by offsetting a first bandwidth from the relative entity boundary toward the interior of the entity. The interlayer overlap zone is obtained by intercepting the overlapping area of ​​the projection of the current slice entity domain and the adjacent slice entity domain according to the second bandwidth. The inner cavity protection zone is obtained by intercepting the negative space between two relative entity boundaries according to the protection bandwidth. The first bandwidth is determined based on the target wall thickness on both sides and the exposure pixel size; the second bandwidth is determined based on the slice layer thickness and depth penetration parameters; and the protection bandwidth is determined based on the target clearance and lateral widening parameters. The upper limit of the shared dose is set as the product of the critical curing energy and the safety factor, wherein the safety factor is greater than zero and less than one; Based on the three-dimensional dose influence matrix, with the constraint that the cumulative dose of each voxel in the core zone on both sides reaches the critical curing energy and the cumulative dose of each voxel in the inner cavity protection zone does not exceed the upper limit of the shared dose, the initial basic exposure pixel set and the initial exposure dose of each pixel are solved. The upper limit of the cumulative dose of each voxel in the inner cavity protection zone is taken as the scalar dose of the inner cavity protection zone. The transmission exposure of the exposed adjacent layers is calculated using the corresponding exposure record. The transmission exposure of the adjacent layers to be exposed is calculated using the planned exposure parameters of the corresponding slice. The transmission exposure contribution of the side without adjacent layers is taken as zero. Subtracting the initial exposure dose and the scalar dose generated by the transmission exposure of adjacent layers within the inner cavity protection zone from the upper limit of the shared dose yields the remaining dose budget, which has the same dimensions as the upper limit of the shared dose.

6. The method for partitioned photopolymerization molding of double-walled silicon-based ceramic core for super-air-cooled blades according to claim 5, characterized in that, S4 include: Adjacent layer superposition exposure calibration was performed using a calibration sample with the same silicon-based ceramic slurry, slice layer thickness, and exposure conditions as the silicon-based ceramic slurry, to obtain the adjacent layer superposition exposure calibration results; The cumulative dose boundary value that satisfies the preset interlayer bonding load threshold in the adjacent layer superposition exposure calibration results is determined as the lower limit of the overlap dose. The difference between the lower limit of the overlap dose and the current cumulative dose of the interlayer overlap is calculated to obtain the overlap dose gap; Based on the three-dimensional dose influence matrix, candidate exposure increments for filling the overlap dose gap on both sides are calculated respectively, and the upper limit of the additional cumulative dose generated by each voxel in the same inner cavity protection zone of each candidate exposure increment is taken as the scalar additional dose. Candidate exposure increments whose scalar additional dose does not exceed the remaining dose budget are determined as allowable exposure increments, and the exposure intensity or exposure time of the corresponding area is increased according to the allowable exposure increments.

7. The method for partitioned photopolymerization molding of double-walled silicon-based ceramic core for super-air-cooled blades according to claim 6, characterized in that, When the scalar additional dose used to fill the overlap dose gap on either side exceeds the remaining dose budget, switch to the compensation backoff branch and resolve in at least one of the following ways: normal displacement of the corresponding solid profile into the solid within the allowable deviation of the target wall thickness on both sides; Reduce the coverage area of ​​the supplementary exposure mask within the neighborhood of the inner cavity protection zone boundary; Additional doses are redistributed in ascending order of the dose influence coefficient of candidate exposure pixels on the cavity protection zone; After each rollback action, if the coverage of the entity contour or supplementary exposure mask changes, the corresponding entries of the affected paired boundary units, entity core zone, interlayer overlap zone, inner cavity protection zone and the three-dimensional dose influence matrix are updated, and the overlap dose gap and the remaining dose budget are recalculated in any case. When the cumulative dose of the solid core strip reaches the critical curing energy, the cumulative dose of the interlayer overlap strip reaches the lower limit of the overlap dose, and the scalar dose of the inner cavity protection strip does not exceed the upper limit of the shared dose, the paired boundary control parameters containing the allowable exposure increment on both sides, the contour displacement, and the mask range are generated. When the contour displacement reaches the allowable deviation boundary of the target wall thickness, there are no candidate exposure pixels available for reallocation, or the number of re-solutions reaches the upper limit of the preset number of iterations, the current slice is marked as unformable, the issuance of the exposure mask for the current slice is stopped, and the process returns to step S1 to adjust the slice layer thickness or partition boundary.

8. The method for partitioned photopolymerization molding of double-walled silicon-based ceramic core for super-air-cooled blades according to claim 7, characterized in that, S5 include: Write the initial basic exposure pixel set and the initial exposure dose of each pixel into the basic exposure mask, and write the overlap dose compensation part corresponding to the area located on the inner wall or forming the connecting free segment or suspended segment in the solid forming area, the overlap dose compensation part corresponding to the area forming the micro-connection root in the solid forming area, and the geometric compensation part of the inner cavity protection zone boundary into the corresponding supplementary exposure mask. Assign grayscale values, illumination, exposure time, and contour displacement to each exposure mask, and perform exposure on the same slice in the order of basic exposure mask first, followed by supplementary exposure mask.

9. The method for partitioned photopolymerization molding of double-walled silicon-based ceramic core for super-air-cooled blades according to claim 8, characterized in that, After the current slice exposure is completed, the gray value, illuminance and exposure time of each exposure mask are recorded, the contour measurement data of the current cured layer is obtained, and the contour measurement data is registered with the target contour of the current slice to obtain the wall thickness deviation and clearance deviation. The transmission exposure scalar dose generated by the current slice for the next slice is corrected based on the wall thickness deviation, the clearance deviation, and the recorded exposure parameters, and the corresponding planned exposure estimate in the budget calculation of the next slice is replaced with the corrected transmission exposure scalar dose.

10. The method for partitioned photopolymerization molding of double-walled silicon-based ceramic core for super-air-cooled blades according to claim 9, characterized in that, Before step S1, wall thickness calibration samples and pore volume calibration samples are prepared using the same material as the silicon-based ceramic slurry. The wall thickness calibration samples and the pore volume calibration samples are sequentially cleaned, degreased and sintered to obtain the sintering shrinkage rate in the wall thickness direction and the sintering shrinkage rate in the pore volume direction. The target wall thicknesses on both sides are corrected according to the sintering shrinkage rate in the wall thickness direction, and the target clearance is corrected according to the sintering shrinkage rate in the channel direction. Steps S1 to S5 are then performed with the corrected target wall thicknesses on both sides and the target clearance.

Images