A method for detecting and analyzing heat treatment performance of a high-temperature alloy forge piece

Abstract

The application discloses a kind of high-temperature alloy forgings heat treatment performance detection analysis methods, it is related to alloy forgings heat treatment performance detection field, comprising: the present application obtains the three-dimensional geometric point cloud of forgings surface, continuous sound velocity field distribution data and residual stress data in unified space coordinate system by adopting optical scanning, ultrasonic scanning and micro area X-ray diffraction;Recognize abnormal area by obtaining sound velocity field distribution data, plan abnormal area encryption, and the non-uniform test point array of normal area sparse, to generate the continuous distribution nephogram of sound velocity and residual stress;Temperature history curve is generated by real-time acquisition temperature data with time stamp to divide heat treatment stage, and space information of surface geometric defect area and material performance abnormal area is combined to carry out space-time matching, recognize surface geometric defect area and material performance abnormal area cause, realize non-destructive performance detection, and can accurately trace the process root of defect.

Description

Technical Field

[0001] This invention belongs to the field of heat treatment performance testing technology for alloy forgings, and relates to a method for testing and analyzing the heat treatment performance of high-temperature alloy forgings. Background Technology

[0002] High-temperature alloy forgings are widely used in hot-end components of critical equipment such as aero-engines and gas turbines, and their performance and reliability directly depend on the quality of the heat treatment process. The heat treatment process, by controlling temperature and time, induces the evolution of the microstructure within the forging and forms a residual stress distribution, thereby endowing the material with the required comprehensive properties. Therefore, comprehensive testing and analysis of the performance of heat-treated forgings is a core step in ensuring product quality.

[0003] Currently, there are several technical bottlenecks in the testing technology for the heat treatment performance of high-temperature alloy forgings: First, traditional testing methods are mostly destructive, forcing forgings to be scrapped directly after testing. This not only increases production costs but also makes it impossible to conduct full inspection of key components, posing a risk of missing quality defects. Second, existing technologies mostly use sampling measurements of uniformly and discretely distributed test points. The data obtained cannot reflect the continuous distribution characteristics of mechanical property parameters and the details of performance fluctuations. This point-to-surface approach makes it difficult to comprehensively evaluate the heat-treated diffusion layer, decarburized layer, and internal performance gradient, which can easily lead to a one-sided assessment of the overall performance status of the forging and omission of local performance abnormalities.

[0004] Finally, traditional testing only performs final performance sampling after heat treatment. The online monitoring of the heat treatment process and the final performance testing after heat treatment are disconnected from each other, and the data of each dimension are isolated from each other. Once the performance is unqualified, it is difficult to trace the specific process stage and abnormal parameters that caused the defect. This not only leads to low traceability efficiency, but also makes it easy for batches of defective products to flow out, and it is difficult to control the risk of batch data dispersion. Summary of the Invention

[0005] In view of this, in order to solve the problems mentioned in the background art, the present invention provides a method for testing and analyzing the heat treatment performance of high-temperature alloy forgings.

[0006] The objective of this invention can be achieved through the following technical solution: a method for detecting and analyzing the heat treatment performance of high-temperature alloy forgings, comprising: S1, real-time acquisition of the temperature inside the heat treatment furnace to form a temperature history curve with timestamps.

[0007] S2. After heat treatment, use an optical scanner to acquire point cloud data of the forging. Based on a unified spatial coordinate system, perform ultrasonic scanning and micro-area X-ray diffraction tests on the surface of the forging to obtain sound velocity field distribution data and surface residual stress data.

[0008] S3. Compare the point cloud data of the forging with the original CAD forging model to identify the surface geometric defect areas.

[0009] S4. Correlate the sound velocity field distribution data with the surface residual stress data to locate areas of abnormal material properties.

[0010] S5. Divide the heat treatment stages according to the temperature history curve and the recorded timestamps.

[0011] S6. The spatial information of the surface geometric defect area and the material performance abnormal area is spatiotemporally matched with the temperature change of the corresponding spatial location in each stage of heat treatment to determine the heat treatment stage that leads to the surface geometric defect area and the material performance abnormality.

[0012] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention uses optical scanning, ultrasonic scanning and micro-area X-ray diffraction to obtain the three-dimensional geometric point cloud, continuous sound velocity field distribution data and residual stress data of the forging surface in a unified spatial coordinate system. It adopts non-destructive testing methods to realize non-destructive testing and evaluation of the heat treatment performance of forgings, and solves the problems of high cost and inability to fully inspect caused by destructive sampling inspection.

[0013] (2) This invention identifies abnormal regions by acquiring sound velocity field distribution data, plans to densify abnormal regions and sparse non-uniform test point arrays in normal regions, thereby generating a continuous distribution cloud map of sound velocity and residual stress, overcoming the limitations of traditional discrete sampling that uses points to represent surfaces, and reducing the risk of missing detection of local performance abnormal regions.

[0014] (3) This invention generates a temperature history curve by collecting temperature data with timestamps in real time to divide the heat treatment stage, and combines the spatial information of surface geometric defect area and material performance abnormal area to perform spatiotemporal matching, identify the causes of surface geometric defect area and material performance abnormal area, solve the situation of the separation between traditional detection results and process, thereby avoiding the outflow of batch defective products and effectively controlling the risk of batch data dispersion. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a diagram illustrating the implementation steps of the method of the present invention.

[0017] Figure 2 This is a flowchart for determining abnormal areas of heat-treated materials in forgings according to the present invention.

[0018] Figure 3This is a flowchart illustrating the determination of the dominant causes of heat treatment defects and abnormal areas in forgings according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 As shown, the present invention provides a method for detecting and analyzing the heat treatment performance of high-temperature alloy forgings, including: S1, real-time acquisition of the temperature inside the heat treatment furnace to form a temperature history curve with timestamps.

[0021] According to one embodiment of the present invention, the real-time acquisition of the temperature inside the heat treatment furnace to form a time-stamped temperature history curve is achieved by setting a high-temperature resistant infrared observation window on the wall of the heat treatment furnace and fixing an infrared thermal imager outside the furnace.

[0022] The infrared observation window is made of quartz glass. It is installed in a position that avoids strong airflow, can cover the main monitoring area, can work stably for a long time in high-temperature environments, and allows infrared radiation of a specific band to pass through effectively.

[0023] Because heat treatment furnaces often contain protective atmospheres or trace amounts of combustion products, the mid-wave infrared band, located in the gap between the main gas absorption bands, has higher transmittance and is less affected by changes in the furnace atmosphere. Therefore, mid-wave infrared imagers, typically in the 3–5 μm range, are preferred to minimize the impact of atmosphere absorption on temperature measurement accuracy. Furthermore, the temperature measurement range of the infrared imager must fully cover and be wider than the actual temperature range of the entire process. For example, if the highest process temperature is 1200℃, a model with a measurement upper limit of no less than 1500℃ should be selected to ensure effective measurement even during instantaneous temperature fluctuations or localized overheating, avoiding signal saturation.

[0024] Throughout the heat treatment process, an infrared thermal imager is used to acquire thermal images inside the furnace, and a unified timestamp is recorded synchronously for each frame of the image.

[0025] Each acquired thermal image frame undergoes radiation signal processing to convert it into two-dimensional temperature field data.

[0026] Radiation signal processing includes, but is not limited to, preprocessing such as non-uniformity correction, bad pixel replacement, and background noise filtering to improve image quality.

[0027] To eliminate the interference of thermal radiation from the furnace inner wall on the surface temperature measurement of forgings, background radiation separation is implemented. Specifically, firstly, the furnace inner wall is regarded as an approximately uniform radiation source, and its radiation energy will be superimposed on the radiation signal of the forging. By measuring the temperature of the furnace wall itself and combining it with its emissivity, the radiation contribution of the furnace wall in the receiving band of the thermal imager is calculated. Then, the radiation contribution of the furnace wall is subtracted from the total radiation signal of the forging measurement area to obtain the net radiation signal of the forging surface.

[0028] Furthermore, to convert the obtained net radiation signal into a temperature value, the emissivity of the forging material surface needs to be dynamically corrected. The specific implementation method is as follows: First, empirical formulas are obtained in advance through experiments to show the evolution of the emissivity of the forging surface with temperature and holding time under the same heat treatment atmosphere for high-temperature alloys; then, during temperature measurement, the estimated emissivity value at the current moment is calculated based on the preliminary temperature value obtained from the inversion and the cumulative heating time, according to the empirical formulas; then, the updated emissivity value is substituted into the radiation calculation formula for iterative calculation, and finally the corrected true temperature is output.

[0029] Based on the two-dimensional temperature field data sequence and its corresponding timestamps, a time-stamped temperature history curve is generated to characterize the furnace. With time as the horizontal axis and representative temperature values ​​as the vertical axis, a continuous time-stamped temperature history curve is plotted, recording the trajectory of temperature change over time throughout the entire heat treatment process.

[0030] The timestamped temperature history curve contains two-dimensional spatial distribution data of the temperature field inside the furnace and on the surface of the forging at every time point during the entire heat treatment process. Based on this, the temperature time series during the entire heat treatment process can be extracted from any spatial coordinate.

[0031] S2. After heat treatment, use an optical scanner to acquire point cloud data of the forging. Under a unified spatial coordinate system, perform ultrasonic scanning and micro-area X-ray diffraction tests on the surface of the forging to obtain sound velocity field distribution data and surface residual stress data.

[0032] The steps to obtain point cloud data of forgings are as follows: set up reference marker points in the non-critical detection areas on the surface of the forgings after heat treatment.

[0033] Non-critical testing areas refer to regions on the surface of forgings that do not directly affect performance evaluation, or where controllable data obstruction and minor interference are permissible. Specifically, these areas are defined as those that do not directly participate in the main mechanical load-bearing capacity of the component, are not located in high-temperature combustion gas channels, or do not have strict requirements for cooling airflow organization. For example, for guide vane forgings, the inner cavities of the side plates of the blade that are not part of the cooling airflow channels, and the planes of the upper and lower mounting plates that do not mate with other components, can generally be considered non-critical testing areas. Selecting such areas for setting up reference points can minimize physical obstruction and interference to performance testing signals.

[0034] The optical scanner is controlled to adjust its scanning orientation around the forging in both horizontal and vertical directions to perform multi-view scanning.

[0035] It should be noted that before planning the scan path, it is necessary to ensure that at least three non-collinear common reference markers can be observed from any adjacent scan viewpoint.

[0036] If a complex forging is scanned directly in a single scan, data loss may occur due to line-of-sight obstruction. Furthermore, multi-view scanning without a stitching reference can lead to misalignment, overlap, or gaps in the final point cloud model, severely impacting the accuracy of subsequent geometric defect identification. Therefore, the overlapping area between two adjacent scanning views should be greater than 30% of the field of view of a single scan, and this overlapping area should contain at least three complete reference marker points to acquire multiple sets of local point cloud data covering the entire measured surface of the forging.

[0037] Based on the common reference markers in the scanning data from different perspectives, the multi-view scanning data are stitched together to obtain point cloud data of the forging surface.

[0038] The specific implementation process of the above data stitching is as follows: using the automatic identification function of the marker points in the scanner's software, the three-dimensional center coordinates of all reference marker points in each local point cloud are extracted; based on at least three non-collinear common reference marker point pairs, the least squares method is used to solve the optimal spatial transformation matrix between two adjacent viewpoint point clouds, thereby achieving point cloud alignment.

[0039] Furthermore, the above process is performed sequentially for all viewpoints, and the coordinates of the marked points of all viewpoints are adjusted as a whole to generate the final point cloud model of the forging.

[0040] The above-mentioned sound velocity field distribution data and surface residual stress data with precise spatial location are obtained. The specific implementation steps are as follows: Based on a unified spatial coordinate system, the three-dimensional surface area of ​​the forging is orthogonally projected onto a two-dimensional plane to generate a two-dimensional projection area.

[0041] Specifically, the least squares method is used to perform plane fitting on the point cloud data of the three-dimensional surface area of ​​the forging to obtain its normal vector. This normal vector is used as the projection direction, i.e., the Z-axis, to establish a local coordinate system, and the coordinates of the three-dimensional points are projected onto this plane, i.e., the XY plane. For non-developable surfaces, piecewise projection can be used.

[0042] The establishment of the unified spatial coordinate system is achieved through the following steps: First, before optical scanning, at least four non-coplanar reference marks are set on the surface of the forging; second, a metrologically calibrated contact coordinate measuring machine is used to accurately measure the coordinates of each mark in the machine tool coordinate system; then, the optical scanner, ultrasonic probe, and X-ray diffraction probe are controlled to sequentially align with the same mark to take readings, and the transformation matrix between the sensor coordinate system and the machine tool coordinate system is calculated respectively; finally, the data collected by each sensor is unified to the same machine tool coordinate system through the transformation matrix chain.

[0043] Based on the preset detection step size, a regular two-dimensional grid dot matrix is ​​generated in the two-dimensional projection area. The planar coordinates of each two-dimensional grid point are then projected back onto the corresponding three-dimensional surface of the point cloud data to obtain its three-dimensional spatial coordinates and generate the actual motion path point sequence of the ultrasonic probe.

[0044] The preset detection step size (Δx, Δy) is determined based on the nominal size of the smallest defect to be inspected by the ultrasonic probe. To ensure detection resolution, the step size is preferably no greater than half of the feature size of the smallest defect.

[0045] In the specific implementation of the above scheme, the generated regular two-dimensional grid dot matrix uses the circumscribed rectangle of the projection area as the boundary and a preset detection step size (Δx, Δy) as the interval to generate the coordinates (x, y) of all grid nodes covering the circumscribed rectangle. i y i ).

[0046] The corresponding 3D surface projected onto the point cloud data as described above passes through each 2D grid point (x i y i Find the nearest neighbor of the point cloud data in the XY plane projection coordinates, calculate the Euclidean distance from the nearest neighbor to the XY plane, and thus obtain the height on the original surface.

[0047] Based on the acquired set of three-dimensional spatial coordinate points, a sequence of motion paths for the ultrasonic probe is planned to ensure collision-free probe movement and coverage of all detection points. Specifically, a zigzag path can be used to sort all three-dimensional detection points, forming an ordered sequence of three-dimensional spatial coordinates.

[0048] The ultrasonic waves are controlled to scan the forging according to the path point sequence, and the longitudinal wave velocity value and three-dimensional coordinates of each point are measured and recorded to form a sound velocity field distribution dataset.

[0049] By analyzing the sound velocity field distribution dataset, abnormal regions of sound velocity values ​​that deviate from the normal range are identified.

[0050] The normal range can be determined by statistically analyzing the historical sound velocity data of qualified forgings in the same batch, taking the average sound velocity ± 3 times the standard deviation; for the initial production of new products without historical data, the sound velocity value measured after ideal heat treatment of the same material standard test bar can be used as a reference benchmark.

[0051] Spatial clustering analysis is performed on all data points identified as anomalies, and anomalies that are spatially close are grouped into a continuous region, and the boundaries of the region are marked.

[0052] A first test array is planned in the region of abnormal sound speed, and a second test array is planned in the region of normal sound speed, together forming a non-uniform test array.

[0053] In the specific implementation of the present invention described above, the specific steps for constructing a non-uniform test grid are as follows: determine the geometric center point and the smallest bounding rectangle of the sound velocity anomaly region, connect the midpoints of opposite sides of the smallest bounding rectangle to form a longitudinal central axis and a transverse central axis, and divide the rectangular region into four sub-rectangular regions.

[0054] The determination of the minimum bounding rectangle for the sound velocity anomaly region is achieved through the following calculation steps: First, extract the projected coordinates of all anomaly points in the sound velocity anomaly region onto the XY plane of a unified spatial coordinate system to form a two-dimensional point set; then, calculate the covariance matrix of the two-dimensional point set and perform eigenvalue decomposition, determine the direction of the eigenvector corresponding to the largest eigenvalue as the principal direction of the two-dimensional point set, and use this principal direction as the initial angle to obtain an initial oriented boundary rectangle, and calculate the convex hull of the two-dimensional point set; then, starting from the initial angle, rotate the rectangle step by step within a 90° range, and calculate the area of ​​the minimum axis-aligned rectangle enclosing the convex hull at each angle; finally, select the rectangle with the smallest area as the final minimum bounding rectangle.

[0055] An example of generating the aforementioned central axis is as follows: Let the four vertices of the smallest bounding rectangle be P1, P2, P3, and P4. Then, the vertical central axis is the line connecting the midpoint of side P1P2 and the midpoint of side P3P4; the horizontal central axis is the line connecting the midpoint of side P2P3 and the midpoint of side P4P1. The four sub-rectangular regions are formed by the intersection of the two central axes.

[0056] The first test point array was deployed at the geometric center, the midpoints of the four sides of the smallest bounding rectangle, and the center points of the four sub-rectangular regions. This enabled high-density sampling of the sound velocity anomaly region, thus allowing for the characterization of its internal stress gradient changes and distribution details. It can capture the stress distribution trend and potential gradient of the region with a minimal number of test points.

[0057] When the area of ​​the abnormal region exceeds 10 times the area of ​​the ultrasonic scanning unit, the above layout rules can be recursively applied to the four sub-rectangular regions.

[0058] Furthermore, firstly, the first test point array is laid out within the smallest bounding rectangle of the abnormal area using the method described above. Secondly, the area of ​​each of the four sub-rectangular regions is calculated. If the area of ​​a certain sub-rectangular region is still greater than 10 times the area of ​​the ultrasonic scanning unit, then that sub-rectangular region is considered a new independent region, and the same first test point array is recursively applied until the shortest side length of the sub-rectangular region is less than the effective beam diameter of the ultrasonic probe. Then, further fine-mapping of points in that sub-region is terminated. The area of ​​the ultrasonic scanning unit refers to the area of ​​a rectangle with a preset detection step length (Δx, Δy) as its side length. Each time a new point is recursively laid out, spatial coordinate deduplication is required to ensure that points at the same physical location are not repeatedly laid out due to overlapping recursion boundaries.

[0059] In the region of normal sound speed, a uniform grid is generated according to a preset sparse grid spacing, and a second test point array is set at the geometric center point of each grid.

[0060] The preset sparse grid spacing is set according to the detection efficiency requirements, typically 2 to 5 times the ultrasonic detection step size. The generated grid covers a continuous region with normal sound velocity, and points falling within the smallest bounding rectangle of any abnormal region must be removed from the uniform grid across the entire area.

[0061] The first and second test point lattices are projected onto the corresponding three-dimensional surface of the forging point cloud data to obtain its three-dimensional spatial coordinates.

[0062] For the two-dimensional plane coordinates of each deployment point, the nearest neighbor point is found in the forging point cloud data and local plane or surface fitting is performed. The three-dimensional coordinates of the nearest neighbor point on the actual surface of the forging and the surface normal vector of the nearest neighbor point are calculated by interpolation, which is used for accurate positioning of the X-ray diffractometer.

[0063] All test points ultimately obtain three-dimensional coordinates using the same set of point cloud data and the same projection algorithm, ensuring that they share the same unified spatial coordinate system as the sound velocity field data.

[0064] The micro-area X-ray diffraction system drives the probe to be positioned sequentially at each test point, and measures and records the surface residual stress value and its three-dimensional coordinates at each point, forming residual stress data corresponding to the spatial position of the sound velocity field.

[0065] S3. Compare the point cloud data of the forging with the original CAD forging model to identify the surface geometric defect areas.

[0066] The steps for identifying surface geometric defect regions are as follows: import the point cloud data of the forging surface and the original CAD forging model into the same data processing environment, and perform best-fit spatial alignment.

[0067] The optimal fitting spatial alignment is achieved through an iterative nearest-point algorithm. The specific steps include: first, preliminary localization based on three or more common feature points; then, in each iteration, finding the nearest point in the theoretical model for each point in the measured point cloud to form a point pair; then, based on all point pairs, solving for the optimal rigid body transformation using the least squares method; next, applying the rigid body transformation to update the position of the measured point cloud; repeating the above process until the average distance change between two iterations is less than a set threshold, such as 0.001 mm. This eliminates the systematic deviation caused by the inconsistency between the forging's orientation and the coordinate system during optical scanning.

[0068] After spatial alignment is completed, the minimum normal distance from each data point in the point cloud data of the forging surface to the theoretical surface of the original CAD forging model is calculated. The normal distance values ​​of all data points constitute the three-dimensional geometric deviation dataset of the forging surface. This three-dimensional geometric deviation dataset intuitively represents the deviation of the actual shape of the forging from the theoretical shape and its spatial distribution.

[0069] The normal distance is calculated as follows: For the measured point P m Find its closest point P in the theoretical model. n If the unit normal vector of the triangular facet is given, then the normal deviation is... This value is a signed distance; a positive value indicates excess material, and a negative value indicates missing material. This represents the unit normal vector.

[0070] Traverse the 3D geometric deviation dataset, filter out data points whose absolute deviation value exceeds the preset tolerance threshold, and mark these points as initial outliers.

[0071] The preset tolerance threshold is determined based on product technical requirements and process quality data: First, a baseline value is set according to the tolerance of the forging design drawings; then, statistical analysis of the geometric deviation data of historical qualified batches of forgings is performed to calculate its mean and standard deviation, and the mean plus or minus a certain multiple of the standard deviation is used as the statistical control limit; subsequently, the smaller value between the baseline value and the statistical control limit is selected as the tolerance threshold. For example, for a certain forging, the threshold can be set to ±0.15mm.

[0072] If the spatial distance between multiple initial anomalies is less than a preset proximity threshold, these points are merged into the same set, and the merging operation is performed iteratively until all initial anomalies that meet the proximity condition are merged to form a surface geometric defect region. This process aggregates spatially discrete initial anomalies into anomaly point clusters that are continuously distributed in three-dimensional space, with each anomaly point cluster corresponding to a potential surface geometric defect region.

[0073] The preset proximity threshold is determined based on the average point spacing of the measured point cloud data of the forging and the size characteristics of the smallest defect to be identified: First, the average nearest neighbor distance of the point cloud data in three-dimensional space is calculated as a basic reference value; then, combined with the minimum defect feature size to be detected according to the process requirements, the proximity threshold is set to an integer multiple of the average point spacing, and not greater than half of the minimum defect feature size, to ensure that discrete anomalies belonging to the same physical defect in space can be correctly merged into the same set.

[0074] The merging process is as follows: First, two key parameters are set: neighborhood radius and minimum number of points. Starting from any unprocessed initial outlier, neighboring points within its neighborhood radius are searched. If the number of neighboring points reaches the minimum number, a new region's core is formed, and this region is recursively expanded until no more new neighboring points can be included. This process is repeated until all spatially densely connected outliers are aggregated into independent surface geometric defect regions. The neighborhood radius can be set to 3-5 times the average spacing of the point cloud, and the minimum number of points can be set to 3-5 to filter noise points.

[0075] S4. Correlate the sound velocity field distribution data with the surface residual stress data to locate areas of abnormal material properties.

[0076] See Figure 2 As shown, the steps for locating abnormal material performance areas are as follows: spatial alignment and interpolation processing are performed on the sound velocity field distribution data and surface residual stress data to generate sound velocity distribution cloud maps and residual stress distribution cloud maps.

[0077] The aforementioned spatial alignment is based on a unified spatial coordinate system and reference markers established by optical scanning. The three-dimensional coordinates of each data point acquired by the ultrasonic scanning system and the X-ray diffraction system are uniformly transformed to this coordinate system. For spatial position deviations, nearest neighbor matching is used to find the corresponding stress data point within a radius of ±0.2 mm for each sound velocity data point, establishing a related point pair to ensure that subsequent analyses are based on the same spatial location.

[0078] The interpolation process employs Kriging interpolation, which, based on spatial data correlation, estimates the values ​​of unknown points by constructing a semi-variogram model, thereby estimating the spatial distribution. For the center point of each grid pixel, all original data points within a radius of 3 mm are searched; based on the values ​​of these original data points and their distances to the center point, weights are calculated using a spherical model; a weighted sum is then used to obtain the estimated sound velocity and stress values ​​for that pixel. This process iterates through all pixels, generating sound velocity distribution contour maps and residual stress distribution contour maps.

[0079] Calculate the gradient value of each point in the sound velocity distribution cloud map and the residual stress distribution cloud map respectively. Mark the continuous pixel or grid cell region in the distribution cloud map where the gradient value exceeds the corresponding threshold as the sound velocity change region and the stress change region.

[0080] The gradient value is obtained by calculating the gradient magnitude. For each grid pixel, a two-dimensional convolution is performed using the Sobel operator to obtain the gradient components in the X and Y directions, respectively. and Then the gradient value of that pixel is .

[0081] The corresponding threshold can sort the gradient values ​​of all pixels by size, and determine the pixels with the top 10% gradient values ​​as high gradient pixels.

[0082] Extract the outer contour lines of each abrupt change region to obtain the sound velocity characteristic boundary and the residual stress characteristic boundary, respectively.

[0083] Specifically, foreground pixels with gradient values ​​exceeding the corresponding threshold are marked as 1, while other background pixels are marked as 0, resulting in a binary image. Then, for each connected region composed of foreground pixels, starting from a seed pixel on the boundary, the process iteratively traces the region using 8-neighborhood until returning to the starting point, thereby extracting the outer pixel chain surrounding each connected abrupt change region. Finally, the coordinate sequence of the extracted original outer pixel chain is smoothed using B-spline processing to filter out jagged noise caused by pixel discretization, generating a vector feature boundary line. The spatial coordinates of all nodes of the vector feature boundary line are recorded to obtain the sound velocity feature boundary or residual stress feature boundary.

[0084] The sound velocity characteristic boundary and the residual stress characteristic boundary are superimposed in the same coordinate system, and the spatial distance between them is calculated. If the average spatial distance between continuous characteristic boundary segments is less than the distance threshold, and the area enclosed by the two characteristic boundary segments overlaps, then the area is determined to be a region with abnormal material properties.

[0085] The spatial distance is calculated by taking the Euclidean distance from each node on a sound velocity characteristic boundary to the nearest stress characteristic boundary node, and then averaging the distances of all nodes.

[0086] If the area where the sound velocity characteristic boundary intersects with the stress characteristic boundary is greater than zero, then it is determined that there is an area overlap.

[0087] The distance threshold is typically set to 3 to 5 times the pixel size of the sound velocity or stress distribution cloud map grid to allow for minor positioning errors.

[0088] Grain anomalies or microcrack initiation during heat treatment of forgings can simultaneously affect their elastic wave propagation characteristics and macroscopic internal stress state. An average spatial distance less than a distance threshold ensures that the two compared boundaries are spatially proximate, eliminating accidental anomaly matching from spatially distant locations. Furthermore, it requires that not only are the boundaries of sound velocity anomalies and residual stress anomalies close, but their respective defined anomalous regions must also spatially overlap, avoiding misjudgments caused by proximity of boundaries but separation of the anomalous core areas. Therefore, performance anomaly regions caused by inherent material variations should theoretically exhibit both acoustic and mechanical property anomalies spatially, and these two types of anomaly regions should have a strong spatial correlation.

[0089] S5. Divide the heat treatment stages according to the temperature history curve and the recorded timestamps.

[0090] Specifically, the first derivative of the temperature history curve is calculated to obtain the temperature change rate sequence.

[0091] The time point corresponding to the sign reversal in the temperature change rate sequence is determined as the candidate boundary point of the heat treatment stage, and the time period defined by each two adjacent candidate boundary points is processed in turn.

[0092] If the temperature change rate sequence is consistently positive and the end temperature of a time period is higher than the start temperature, then that time period is a warming phase.

[0093] It should be noted that when the temperature change rate sequence is consistently positive, it indicates that the furnace temperature remains in a thermodynamic state of increasing over time throughout the heat treatment process, which aligns with the process objectives and thermodynamic change characteristics of the heating phase. Furthermore, the fact that the end temperature of the time interval is higher than the starting temperature verifies the cumulative effect of temperature changes within that time interval, thus eliminating false temperature change rate feedback caused by instantaneous signal disturbances.

[0094] If the temperature change rate sequence is zero, and the absolute value of the difference between the start and end temperatures of a time period is less than the steady-state temperature difference tolerance, then that time period is the heat preservation stage.

[0095] A zero value in the temperature change rate sequence reflects a dynamic balance between heat input and heat loss within the furnace, satisfying the thermodynamic characteristics of the heat preservation stage and serving as the basis for determining the dynamic trend of the heat preservation status. However, considering the minor equipment fluctuations and environmental disturbances present during actual heat treatment, the temperature change rate sequence alone cannot completely eliminate misjudgments caused by instantaneous disturbances. Therefore, by ensuring that the absolute value of the temperature difference between the start and end of a time interval is less than the stable temperature difference tolerance, it is confirmed that the temperature has not fluctuated beyond the allowable range of the process throughout the entire time interval, ensuring the sustainability of the determined stable state.

[0096] The stable temperature difference tolerance is determined in the following way: based on a large amount of historical temperature data of qualified forgings during the heat preservation stage, the absolute value distribution of the difference between the initial temperature and the final temperature is statistically calculated, and the 95th percentile of the statistical distribution is taken as the basic tolerance value; finally, the stable temperature difference tolerance is set to a quantized value that is not less than twice the system temperature measurement error.

[0097] If the temperature change rate sequence is consistently negative, and the end temperature of a time period is lower than the start temperature, then that time period is the cooling phase.

[0098] The consistently negative values ​​of the temperature change rate sequence reflect a dynamic trend of decreasing temperature over time, consistent with the gradual reduction of forging temperature during the cooling phase. Furthermore, the fact that the end temperature of a time period is lower than the starting temperature confirms a downward trend in temperature throughout the entire time interval from the perspective of the cumulative effect of temperature change, thus eliminating false temperature change rate feedback caused by instantaneous signal disturbances.

[0099] S6. Spatiotemporally match the spatial information of the surface geometric defect area and the material performance abnormal area with the temperature change of the corresponding spatial location in each stage of heat treatment to determine the heat treatment stage that caused the surface geometric defect area and the material performance abnormal area.

[0100] See Figure 3 As shown, the heat treatment steps for determining the areas of surface geometric defects and abnormal material properties are as follows: Based on the spatial coordinates of each area of ​​surface geometric defects and abnormal material properties, extract its temperature time series from the temperature history curve.

[0101] Align the temperature time series with the standard process specification for heat treatment of the forging on the time axis.

[0102] Specifically, firstly, the first time point in the temperature time series to reach the preheating set temperature is selected as the first alignment reference point, and the preheating set temperature is usually set to 200-400℃. Then, the end time point of the heating stage specified in the standard process procedure is selected as the second alignment reference point. Next, with the first and second alignment reference points coinciding as constraints, the temperature time series is linearly scaled and aligned with the entire time axis of the standard process procedure. Finally, on the aligned time axis, the start and end time points of the heat preservation stage in the standard process procedure are marked as comparison reference intervals.

[0103] The time interval during which the actual temperature data of each heat treatment stage after identification and comparison continuously deviates from the target temperature conditions in the standard process specification.

[0104] Calculate the proportion of the total duration of the continuous deviation time interval to the total duration of the process stage, and obtain the deviation time percentage.

[0105] If the deviation time percentage of multiple heat treatment stages all exceeds the judgment threshold, the stage with the highest deviation time percentage will be judged as the dominant inducing stage, and the cause leading to the corresponding defect or abnormal area will be determined.

[0106] It should be noted that the determination threshold setting process is as follows: First, the temperature deviation time ratio data of each heat treatment stage of multiple batches of qualified forgings is statistically analyzed, and the mean and standard deviation are calculated. The mean plus 2 to 3 times the standard deviation is used as the determination threshold.

[0107] During the heat treatment of high-temperature alloy forgings, there may be multiple stages where the actual temperature data deviates from the standard process specifications. From the perspective of the process mechanism, the percentage of deviation time directly indicates the intensity of the continuous effect of the temperature anomaly at the corresponding stage. The higher the percentage, the more persistent and critical the interference of the temperature deviation at that stage on the surface geometry and internal material properties of the forging.

[0108] The stage with the highest deviation time is determined as the dominant cause stage. The causes leading to the corresponding defects or abnormal areas include: if the dominant cause stage is the insulation stage and the average insulation temperature deviation is negative, then the specific failure mode is determined to be insufficient insulation; otherwise, the specific failure mode is determined to be overheating.

[0109] A negative deviation value indicates a continuous lack of thermodynamic supply, resulting in insufficient microstructure transformation and incomplete stress release, which is consistent with the consequences of insufficient insulation failure mode. A positive deviation value indicates that the temperature is continuously too high, which can easily lead to microstructure problems such as abnormal grain growth and grain boundary damage, which is consistent with the causes of overheating failure mode.

[0110] If the dominant cause is the cooling stage, and the maximum cooling rate exceeds the upper limit allowed by the process specification, then the specific failure mode is determined to be localized excessive cooling.

[0111] The maximum cooling rate reflects the peak intensity of the temperature drop of the forging per unit time. When the dominant inducing stage is the cooling stage and the maximum cooling rate exceeds the upper limit allowed by the process specification, it indicates that the cooling rate of the forging exceeds the range that the material and process can withstand. This will lead to the formation of a temperature gradient inside the forging, causing local thermal expansion and contraction imbalance, and in turn inducing residual stress concentration. This is completely consistent with the formation mechanism and process consequences of the localized excessively rapid cooling failure mode.

[0112] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0113] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0114] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0115] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0116] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for testing and analyzing the heat treatment properties of high-temperature alloy forgings, characterized in that: include: S1. Real-time acquisition of temperature inside the heat treatment furnace to generate a time-stamped temperature history curve; S2. After heat treatment, use an optical scanner to acquire point cloud data of the forging. Under a unified spatial coordinate system, perform ultrasonic scanning and micro-area X-ray diffraction tests on the surface of the forging to obtain sound velocity field distribution data and surface residual stress data. S3. Compare the point cloud data of the forging with the original CAD forging model to identify surface geometric defect areas; S4. Correlate the sound velocity field distribution data with the surface residual stress data to locate areas of abnormal material properties; S5. Divide the heat treatment stages according to the temperature history curve and the recorded timestamps; S6. Spatiotemporally match the spatial information of the surface geometric defect area and the material performance abnormal area with the temperature change of the corresponding spatial location in each stage of heat treatment to determine the heat treatment stage that caused the surface geometric defect area and the material performance abnormal area. The identified surface geometric defect regions include: The point cloud data of the forging surface and the original CAD forging model are imported into the same data processing environment and the best-fit spatial alignment is performed. After spatial alignment is completed, the minimum normal distance from each data point in the point cloud data of the forging surface to the theoretical surface of the original CAD forging model is calculated. The normal distance values ​​of all data points constitute the three-dimensional geometric deviation dataset of the forging surface. The three-dimensional geometric deviation dataset is traversed, and data points whose absolute deviation values ​​exceed the preset tolerance threshold are filtered out and marked as initial outliers. The preset tolerance threshold is determined by product technical requirements and process quality data: first, a benchmark value is set based on the tolerance of the forging design drawings; then, the mean and standard deviation of the geometric deviation data of historical qualified batches of forgings are statistically analyzed, and the mean plus or minus a certain number of times the standard deviation is used as the statistical control limit; then, the smaller value between the benchmark value and the statistical control limit is selected as the tolerance threshold. If the spatial distance between multiple initial anomaly points is less than the preset proximity threshold, these points are merged into the same set, and the merging operation is performed iteratively until all initial anomaly points that meet the proximity condition are merged to form a surface geometric defect region. The preset proximity threshold is determined based on the average point spacing of the measured point cloud data of the forging and the size characteristics of the smallest defect to be identified: First, the average nearest neighbor distance of the point cloud data in three-dimensional space is calculated as the basic reference value; then, combined with the minimum defect feature size to be detected according to the process requirements, the proximity threshold is set to an integer multiple of the average point spacing, and not greater than half of the minimum defect feature size, to ensure that discrete anomalies belonging to the same physical defect in space can be correctly merged into the same set.

2. The method for testing and analyzing the heat treatment performance of high-temperature alloy forgings according to claim 1, characterized in that: The implementation of S1 includes: A high-temperature resistant infrared observation window is installed on the wall of the heat treatment furnace, and an infrared thermal imager is fixedly installed outside the furnace. Throughout the heat treatment process, an infrared thermal imager is controlled to acquire thermal images inside the furnace, and a unified timestamp is recorded synchronously for each frame of the image. Each frame of thermal image acquired is subjected to radiation signal processing to convert it into two-dimensional temperature field data. Based on the two-dimensional temperature field data sequence and its corresponding timestamps, a time-stamped temperature history curve is generated to characterize the furnace interior.

3. The method for testing and analyzing the heat treatment performance of high-temperature alloy forgings according to claim 1, characterized in that: The steps for obtaining point cloud data of the forging are as follows: Reference markers are set up in non-critical inspection areas on the surface of the heat-treated forgings. The optical scanner is controlled to adjust its scanning orientation around the forging in both horizontal and vertical directions to perform multi-view scanning. Based on the common reference markers in the scanning data from different perspectives, the multi-view scanning data are stitched together to obtain point cloud data of the forging surface.

4. The method for testing and analyzing the heat treatment performance of high-temperature alloy forgings according to claim 1, characterized in that: The steps for obtaining the sound velocity field distribution data and surface residual stress data are as follows: Based on a unified spatial coordinate system, the three-dimensional surface region of the forging is orthogonally projected onto a two-dimensional plane to generate a two-dimensional projection region; Based on the preset detection step size, a regular two-dimensional grid dot matrix is ​​generated in the two-dimensional projection area. The planar coordinates of each two-dimensional grid point are back-projected onto the corresponding three-dimensional surface of the point cloud data to obtain its three-dimensional spatial coordinates and generate the actual motion path point sequence of the ultrasonic probe. The ultrasonic waves are controlled to scan the forging according to the path point sequence, and the longitudinal wave velocity value and three-dimensional coordinates of each point are measured and recorded to form a sound velocity field distribution dataset. Analyze the sound velocity field distribution dataset to identify abnormal regions where the sound velocity values ​​deviate from the normal range; A first test array is planned in the region of abnormal sound speed, and a second test array is planned in the region of normal sound speed, together forming a non-uniform test array. The micro-area X-ray diffraction system drives the probe to be positioned sequentially at each test point, and measures and records the surface residual stress value and its three-dimensional coordinates at each point, forming residual stress data corresponding to the spatial position of the sound velocity field.

5. The method for testing and analyzing the heat treatment performance of high-temperature alloy forgings according to claim 4, characterized in that: The non-uniform test lattice includes: Determine the geometric center point and minimum bounding rectangle of the sound velocity anomaly region. Connect the midpoints of opposite sides of the minimum bounding rectangle to form the longitudinal and transverse central axes, and divide the rectangular region into four sub-rectangular regions. The first test point array is set up at the geometric center point, the midpoint of the four sides of the smallest bounding rectangle, and the center point of the four sub-rectangular regions. In the region of normal sound speed, a uniform grid is generated according to a preset sparse grid spacing, and a second test point array is set at the geometric center point of each grid. The first and second test point lattices are projected onto the corresponding three-dimensional surface of the forging point cloud data to obtain its three-dimensional spatial coordinates.

6. The method for testing and analyzing the heat treatment performance of high-temperature alloy forgings according to claim 1, characterized in that: The areas with abnormal performance of the positioning material include: Spatial alignment and interpolation are performed on the sound velocity field distribution data and the surface residual stress data to generate sound velocity distribution cloud maps and residual stress distribution cloud maps. Calculate the gradient value of each point in the sound velocity distribution cloud map and the residual stress distribution cloud map respectively. Mark the continuous pixel or grid cell region in the distribution cloud map where the gradient value exceeds the corresponding threshold as the sound velocity change region and the stress change region. Extract the outer contour lines of each abrupt change region to obtain the sound velocity characteristic boundary and the residual stress characteristic boundary, respectively. The sound velocity characteristic boundary and the residual stress characteristic boundary are superimposed in the same coordinate system, and the spatial distance between them is calculated. If the average spatial distance between continuous characteristic boundary segments is less than the distance threshold, and the area enclosed by the two characteristic boundary segments overlaps, then the area is determined to be a region with abnormal material properties.

7. The method for testing and analyzing the heat treatment performance of high-temperature alloy forgings according to claim 1, characterized in that: The division of heat treatment stages includes: Calculate the first derivative of the temperature history curve to obtain the temperature change rate sequence; The time point corresponding to the sign reversal in the temperature change rate sequence is determined as the candidate boundary point of the heat treatment stage, and the time period defined by each two adjacent candidate boundary points is processed in turn. If the temperature change rate sequence is consistently positive and the end temperature of a time period is higher than the start temperature, then that time period is the warming phase. If the temperature change rate sequence is zero, and the absolute value of the difference between the start and end temperatures of a time period is less than the steady temperature difference tolerance, then that time period is the heat preservation stage. If the temperature change rate sequence is consistently negative, and the end temperature of a time period is lower than the start temperature, then that time period is the cooling phase.

8. The method for testing and analyzing the heat treatment performance of high-temperature alloy forgings according to claim 1, characterized in that: The heat treatment stages that determine regions of surface geometric defects and abnormal material properties include: Based on the spatial coordinates of each surface geometric defect region and material property abnormality region, the temperature time series is extracted from the temperature history curve. Align the temperature time series with the standard process specification for heat treatment of the forging on the time axis. The time interval during which the actual temperature data of each heat treatment stage after identification and comparison continuously deviates from the target temperature conditions in the standard process specification. Calculate the proportion of the total duration of the continuous deviation time interval to the total duration of the process stage to obtain the deviation time percentage; If the deviation time percentage of multiple heat treatment stages exceeds the judgment threshold, the stage with the highest deviation time percentage will be judged as the dominant inducing stage, and the cause leading to the corresponding defect or abnormal area will be determined. The judgment threshold setting process is as follows: First, statistical data on the proportion of temperature deviation time in each heat treatment stage of multiple batches of qualified forgings are collected, and the mean and standard deviation are calculated. The mean plus 2 to 3 times the standard deviation is used as the judgment threshold.

9. The method for testing and analyzing the heat treatment performance of high-temperature alloy forgings according to claim 8, characterized in that: The stage with the highest deviation time percentage is determined as the dominant inducing factor stage, and the causes leading to the corresponding defect or abnormal area include: If the dominant cause stage is the insulation stage and the average insulation temperature deviation is negative, then the specific failure mode is determined to be insufficient insulation; otherwise, the specific failure mode is determined to be overheating. If the dominant cause is the cooling stage, and the maximum cooling rate exceeds the upper limit allowed by the process specification, then the specific failure mode is determined to be localized excessive cooling.

Images