Adaptability test optimization method and system for element replacement

Abstract

The invention discloses an element replacement adaptability test optimization method and system, and relates to the technical field of element test optimization, and the method comprises the steps: extracting scarce element performance data and scarce element types in target element replacement scene information, and carrying out first-order multi-dimensional screening to obtain a to-be-verified rich element combination set; performing virtual representation reverse fine tuning to obtain a fine tuning rich element combination set to be verified; obtaining a to-be-verified fine-tuning rich element combination sample set and a sputtering experiment result set; and performing test evaluation according to a preset performance test type, and determining a target fine-tuning rich element combination. The technical problems that in the prior art, the test efficiency of replacing scarce elements with rich elements is low, the research and development loss is high, the combination performance of the rich elements is difficult to guarantee, and differentiation is insufficient are solved, replacement test optimization of the scarce elements with the rich elements is achieved, the research and development efficiency is improved, and the test cost is reduced. The research and development loss is reduced; and the combination performance and differentiation potential of rich elements are guaranteed.

Description

Technical Field

[0001] This invention relates to the field of element testing optimization technology, and specifically to an adaptive testing optimization method and system for element replacement. Background Technology

[0002] In the field of target manufacturing for high-end coating materials, the performance of the target directly determines the quality of the coated product. However, traditional target production relies on scarce elements, which suffer from problems such as limited resources, high costs, and unstable supply. Replacing scarce elements with rich elements has become a key direction for industry development. However, existing testing methods rely on experience to select rich elements, which easily overlooks suitable combinations. They lack dual control over the performance and differentiation of rich element combinations. At the same time, the disconnect between virtual prediction and experimentation leads to low testing efficiency and high R&D costs, seriously hindering the industrialization of the technology.

[0003] Existing technologies suffer from low testing efficiency, high R&D costs, and difficulties in guaranteeing the performance and differentiation of rich element combinations. Summary of the Invention

[0004] This application provides an adaptive testing optimization method and system for element substitution, which addresses the technical problems in the prior art of low testing efficiency, high R&D losses, and difficulty in guaranteeing the performance and lack of differentiation of rich element combinations.

[0005] In view of the above problems, this application provides an adaptive testing optimization method and system for element substitution.

[0006] The first aspect of this application provides an adaptive testing optimization method for element substitution, the method comprising: The performance data and types of scarce elements in the target element replacement scene information are extracted and subjected to first-order multidimensional screening in a rich element resource library to obtain a set of rich element combinations to be verified. Based on dual objectives, the virtual representation of the set of rich element combinations to be verified is reverse-tuned to obtain a set of rich element combinations to be verified and finely tuned. The dual objectives include that the performance similarity between the rich element combination to be verified and the scarce element meets a preset performance requirement, and that the distribution concentration between the virtual representations of the set of rich element combinations to be verified is less than a preset distribution concentration requirement. Sample preparation and sputtering experiments are performed on the set of rich element combinations to be verified and finely tuned to obtain a sample set of rich element combinations to be verified and a set of sputtering experiment results. When the set of sputtering experiment results is deemed to have passed the experiment, the sample set of rich element combinations to be verified and finely tuned to be verified is tested and evaluated according to a preset performance test type, and the target rich element combination is determined based on the test evaluation results.

[0007] A second aspect of this application provides an adaptive testing optimization system for element substitution, the system comprising: The rich element combination set acquisition module is used to extract the performance data and type of scarce elements in the target element replacement scene information, and perform first-order multi-dimensional screening in the rich element resource library to obtain a set of rich element combinations to be verified; the reverse fine-tuning module is used to perform virtual representation reverse fine-tuning on the set of rich element combinations to be verified based on dual objectives to obtain a set of fine-tuned rich element combinations to be verified, wherein the dual objectives include the performance similarity between the rich element combination to be verified and the scarce element performance meeting a preset performance requirement and the distribution concentration between the virtual representations of the set of rich element combinations to be verified being less than a preset distribution concentration requirement; the sputtering experiment module is used to traverse the set of fine-tuned rich element combinations to be verified to perform sample preparation and sputtering experiments to obtain a set of fine-tuned rich element combinations to be verified and a set of sputtering experiment results; the test evaluation module is used to test and evaluate the set of fine-tuned rich element combinations to be verified according to a preset performance test type when the set of sputtering experiment results is a pass, and determine the target fine-tuned rich element combination based on the test evaluation results.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: The process involves extracting performance data and types of scarce elements from target element replacement scenario information, and performing first-order multidimensional screening in a rich element resource library to obtain a set of rich element combinations to be verified. Based on dual objectives, the set of rich element combinations to be verified is then subjected to virtual representation and reverse fine-tuning to obtain a set of fine-tuned rich element combinations to be verified. Sample preparation and sputtering experiments are conducted to obtain a sample set of fine-tuned rich element combinations to be verified and a set of sputtering experiment results. When the sputtering experiment results set passes the experiment, the sample set of fine-tuned rich element combinations to be verified is tested and evaluated according to a preset performance test type, and the target fine-tuned rich element combination is determined based on the test evaluation results. This achieves the technical effect of optimizing the replacement test of scarce elements with rich elements, improving R&D efficiency, reducing R&D losses, and ensuring the performance and differentiation potential of rich element combinations. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0010] Figure 1 A schematic flowchart of an adaptive testing and optimization method for element replacement provided in an embodiment of this application; Figure 2 This is a schematic diagram of an adaptive testing and optimization system structure for element replacement provided in an embodiment of this application.

[0011] Figure labeling: Rich element combination set acquisition module 10, reverse fine-tuning module 20, sputtering experiment module 30, test and evaluation module 40. Detailed Implementation

[0012] This application provides an adaptive testing optimization method and system for element substitution, which addresses the technical problems in the prior art such as low testing efficiency, high R&D costs, and difficulty in guaranteeing the performance and lack of differentiation of rich element combinations.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] Example 1, as Figure 1 As shown, this application provides an adaptive testing optimization method for element substitution, the method comprising: Step S100: Extract the performance data and type of scarce elements from the target element replacement scene information, and perform first-order multi-dimensional filtering in the rich element resource library to obtain a set of rich element combinations to be verified.

[0015] Specifically, the process involves clarifying the element replacement scenarios and, in conjunction with the application domain, first defining mature element combinations—element formulations already applied in batches in the target scenario with complete performance data; and then identifying new element combinations to be validated, which are the new formulations planned for replacement, while simultaneously outlining the core indicators that the new combinations must meet. Next, from the production database, mature element combinations with similar application scenarios and core performance requirements to the new element combinations are selected, and their complete production parameters and performance data are extracted to construct a similar mature combination dataset. Then, performance data and types of scarce elements are extracted from the target element replacement scenario information. Using the similar mature combination dataset as a reference, a first-order multi-dimensional screening is conducted in the rich element resource library from both material and process dimensions. First, a set of rich elements with similar materials and a set of rich elements with similar processes are determined and included in the candidate rich element pool. Then, elements in the pool are randomly combined in terms of quantity and proportion, and after screening, a set of rich element combinations to be validated is obtained.

[0016] Step S200: Based on dual objectives, perform reverse fine-tuning of the virtual representation of the set of rich element combinations to be verified to obtain a set of rich element combinations to be verified and fine-tuned. The dual objectives include that the performance similarity between the rich element combination to be verified and the scarce element meets a preset performance requirement, and that the distribution concentration between the virtual representation of the set of rich element combinations to be verified is less than a preset distribution concentration requirement.

[0017] Specifically, based on dual objectives, the virtual representation of the set of rich element combinations to be verified is fine-tuned in reverse, thereby obtaining the set of rich element combinations to be fine-tuned. The dual objectives are: first, to ensure that the performance similarity between the rich element combinations to be verified and the scarce elements meets the preset performance requirements, ensuring that the performance does not deviate from the core standard after replacement; second, to ensure that the distribution concentration among the virtual representations of the set of rich element combinations to be verified is lower than the preset distribution concentration requirement, avoiding excessive similarity between combinations and retaining the possibility of differentiation. In practice, it is first determined whether the virtual representation meets these dual objectives. If not, the target deviation is determined, and fine-tuning is performed according to the preset adjustment range. After multiple verifications and adjustments, the final set of rich element combinations to be fine-tuned is obtained.

[0018] Step S300: Traverse the set of elements-rich combinations to be verified for sample preparation and sputtering experiments to obtain the set of elements-rich combination samples to be verified and the set of sputtering experiment results.

[0019] Specifically, for each element-rich combination in the set to be verified and finely tuned, small-batch samples are first prepared using vacuum melting and target processing to obtain the corresponding element-rich or alloy targets. Next, under standardized process conditions, sputtering experiments are conducted on each prepared sample to test indicators such as film deposition rate, density, and compositional uniformity, and the data during the experiment are recorded. After traversing the entire set of element-rich combinations to be verified, all prepared samples are integrated to form a sample set of element-rich combinations to be verified and finely tuned. Simultaneously, the sputtering experimental data of each sample are summarized to obtain a sputtering experimental results set.

[0020] Step S400: When the sputtering experiment result set is a pass, the sample set of the element-rich combination to be verified is tested and evaluated according to the preset performance test type, and the target element-rich combination is determined based on the test evaluation results.

[0021] Specifically, the sputtering experiment results are first evaluated. If the results are satisfactory, meaning the sample meets the basic compatibility requirements of the sputtering process, then all samples of the element-rich combination to be verified that have passed the experiments are comprehensively tested and evaluated according to the preset performance test types, including resistivity performance test, transmittance performance test, reflectivity performance test, hardness performance test, and adhesion performance test. After all performance tests are completed, the test data are comprehensively weighted and scored, and the sample with the highest comprehensive score and all performance indicators meeting the standards is selected. Finally, the element-rich combination corresponding to the sample is determined as the target element-rich combination for fine-tuning.

[0022] In one possible implementation, step S400 further includes: Step S410: The preset performance test types include resistivity performance test, transmittance performance test, reflectivity performance test, hardness performance test and adhesion performance test.

[0023] Specifically, the preset performance test types cover the detection of key indicators in multiple dimensions. Among them, resistivity performance test is used to measure the conductivity of materials by measuring the resistance characteristics of current conduction in materials; transmittance performance test focuses on the degree of light transmission of materials, detecting the proportion of light of a specific wavelength passing through the material; reflectivity performance test measures the proportion of incident light reflected by the material surface, reflecting the reflective characteristics of the material; hardness performance test uses corresponding testing methods to determine the material's ability to resist local deformation, such as indentation and scratches; and adhesion performance test is used to evaluate the strength of the bond between the film layer and the substrate, and to determine whether it is easy to fall off.

[0024] In one possible implementation, step S100 further includes: Step S110: Using the performance data and types of scarce elements as constraints, perform first-order multidimensional screening of the rich element resource library from the material dimension and the process dimension respectively to determine the material-similar rich element set and the process-similar rich element set.

[0025] Step S120: Add the material-similar rich element set and the process-similar rich element set to the candidate rich element pool.

[0026] Step S130: Combine candidate rich elements in the candidate rich element pool with random numbers and random proportions to obtain a candidate rich element combination set.

[0027] Step S140: Perform stepwise directional iterative screening on the candidate rich element combination set to obtain the rich element combination set to be verified.

[0028] Specifically, constrained by the performance data and types of scarce elements, a first-order multidimensional screening of the rich element resource library is conducted from both material and process dimensions. In the material dimension, a material-level attention mechanism is used to establish affinity: first, the period and group of the scarce element are determined based on the periodic table; then, the period difference between rich and scarce elements is calculated, with 0 for the same period and 1 for adjacent periods, with interval period differences increasing according to the actual interval number; the group difference is also calculated, with 0 for the same main group and 1 for adjacent main groups, and subgroups and main groups calculated based on the difference in group ordinal numbers. Distance weights are set according to element characteristics, with a period difference weight of 0.3 and a group difference weight of 0.4, as group has a more significant impact on the chemical properties of elements. Finally, the atomic structures of rich and scarce elements are simulated using the B3LYP functional in density functional theory (DFT) combined with the 6-31G(d) basis set to obtain valence charges. Subshell structures, such as the number of valence electrons, subshell electron distribution, and orbital hybridization types (e.g., sp³ and sp² hybridization), are analyzed. The degree of overlap in valence electron configurations is calculated, such as the proportion of identical valence electrons and the matching ratio of subshell electron distributions. Electron cloud distribution similarity is obtained by analyzing the overlapping area ratio of electron cloud contour maps. Subsequently, a material-level attention weight is introduced, assigning 0.5 weight to both electronic structure similarity and periodic table adjacency quantification results. This weight is optimized through training with historical element replacement data to ensure a correlation coefficient ≥ 0.85 with actual material properties. The electron configuration matching degree, electron cloud distribution similarity, and the quantification results of periodic differences and group differences are then compared. Weighted fusion is performed, where the period difference quantification value = 1 / (1 + period difference) and the group difference quantification value = 1 / (1 + group difference). The formula is: affinity = 0.5 × (0.6 × electron configuration matching degree + 0.4 × electron cloud distribution similarity) + 0.5 × (0.3 × period difference quantification value + 0.4 × group difference quantification value). Affinity ≥ 0.7 is selected. This threshold is verified through multiple material replacement experiments to ensure that the selected rich elements are similar to the scarce elements in core material properties such as chemical activity and bonding ability, forming a set of material-similar rich elements. In terms of process, the processing feasibility is calculated using a process-level attention mechanism: first, the thermodynamic and process parameters of the rich and scarce elements, such as melting point and boiling point, are measured. The vapor pressure was measured using differential scanning calorimetry, and the sputtering yield was calculated using a static method. The sputtering yield was calculated as the ratio of target mass loss to sputtering power per unit time during the sputtering experiment. The differences in each parameter were normalized, with the normalized value calculated as 1 - |rich element parameter value - scarce element parameter value| / scarce element parameter value. Similarity thresholds were set: normalized values ​​for melting point and boiling point ≥ 0.8, normalized values ​​for vapor pressure ≥ 0.75, and normalized values ​​for sputtering yield ≥ 0.7. Furthermore, the similarity of process behaviors such as melt flowability and cooling crystallization rate in smelting and sputtering such as target etching uniformity and thin film deposition rate was quantified and scored. The similarity score for each process behavior was calculated by weighting the parameter normalized values, with the weight for smelting behavior set to 0.4. The sputtering behavior weight was set to 0.6, as sputtering is a core process in the manufacturing of coating targets. Finally, through process-level attention weighting, a weight of 0.7 was assigned to the similarity score of core process behaviors, and a weight of 0.3 was assigned to the scores of non-core process behaviors. The weights were calibrated using process stability test data. The scores of each item were weighted and fused to obtain the processing feasibility. Rich elements with a processing feasibility ≥ 0.72 were selected to determine the set of rich elements with process similarity.

[0029] After determining the rich element sets for material similarity and process similarity, all rich elements in these two sets are added to the candidate rich element pool according to their respective element identifiers, affinity values ​​in the material dimension, processing feasibility values ​​in the process dimension, and other relevant information. This allows the candidate rich element pool to integrate rich elements selected from both material and process aspects, providing a basic resource reserve for subsequent rich element combination operations.

[0030] For the candidate rich element pool, a random algorithm is set to randomly select the number of rich elements in the pool, such as 2 to 5, and combine them in random proportions. For example, 3 rich elements may be randomly selected and combined in a 1:2:1 ratio, or 2 rich elements may be randomly selected and combined in a 3:1 ratio, thus generating a variety of rich element combinations. Finally, a candidate rich element combination set containing many different rich element combinations is obtained, thereby enriching the number and types of samples for subsequent screening.

[0031] After obtaining the candidate rich element combination set, a screening process is first conducted based on the rich element combination screening space. This screening space is constructed with process similarity as the horizontal axis and material similarity as the vertical axis. Each candidate rich element combination corresponds to a coordinate point in the space. Then, a rich element combination screening subspace is defined with preset process similarity thresholds and preset material similarity thresholds as constraints. Only candidate rich element combinations whose coordinate points fall within this subspace are retained. Subsequently, stepwise directional iterative screening is performed within the screening subspace: first, the center coordinate point and the optimal coordinate point of the screening subspace are extracted, that is, the coordinate point with the highest process similarity and material similarity. The direction from the center coordinate point to the optimal coordinate point is taken as the iteration direction. The center coordinate point is subjected to mean drift iteration according to a preset iteration step size to determine the iteration coordinate point. It is continuously judged whether the iteration neighborhood density of the center coordinate point is greater than or equal to the iteration neighborhood density of the iteration coordinate point. If not, the iteration continues until a preset number of iterations is reached or the iteration neighborhood density of the current iteration is less than the iteration neighborhood density of the previous iteration. Finally, based on the target coordinate point obtained from the iteration and the preset iteration step size, combinations that meet the requirements are extracted from the candidate rich element combination set to form a rich element combination set to be verified.

[0032] In one possible implementation, step S140 further includes: Step S141: Construct a rich element combination screening space based on the candidate rich element combination set, wherein the horizontal axis of the rich element combination screening space is the process similarity, the vertical axis is the material similarity, and each coordinate point corresponds to a candidate rich element combination.

[0033] Step S142: In the rich element combination screening space, a rich element combination screening subspace is constructed with preset process similarity threshold and preset material similarity threshold as constraints.

[0034] Step S143: Perform stepwise directional iterative screening in the rich element combination screening subspace to obtain the set of rich element combinations to be verified.

[0035] Specifically, based on the previously determined set of candidate rich element combinations, a rich element combination screening space is constructed for screening analysis. The screening space is a two-dimensional space. The horizontal axis represents process similarity. The process similarity value comes from the quantification results of the processing feasibility of each candidate rich element combination and rare element in the process dimension. Processing feasibility is obtained by weighted fusion of the similarity of melting behavior and sputtering behavior between rich and rare elements. The similarity of melting behavior is obtained by measuring thermodynamic parameters such as melting point, boiling point, and heat of fusion, calculating the normalized value of parameter difference, and setting a score according to a threshold. The similarity of sputtering behavior is obtained by measuring parameters such as vapor pressure and sputtering yield, calculating the normalized value of parameter difference, and setting a score according to a threshold. The vertical axis represents material similarity. The material similarity value corresponds to the quantification results of the affinity of each candidate rich element combination and rare element in the material dimension. Affinity is obtained by weighted fusion of the periodic table adjacency and electronic structure similarity between rich and rare elements. Periodic table adjacency is obtained by calculating the basic value by statistically analyzing the period difference and group difference and setting distance weight. Electronic structure similarity is obtained by comparing electron configuration, calculating the electron configuration matching degree, and combining quantum chemical methods to analyze the similarity of electron cloud distribution. Each candidate rich element combination will correspond to a unique coordinate point in this two-dimensional screening space based on its corresponding process similarity value and material similarity value. This presents all candidate rich element combinations in the space in the form of coordinate points, providing an intuitive and quantifiable analytical basis for subsequent screening operations.

[0036] First, two core constraints are defined: a preset process similarity threshold and a preset material similarity threshold. The preset process similarity threshold corresponds to the minimum acceptable standard for the feasibility of processing rare elements in terms of the process dimension of candidate rich element combinations, while the preset material similarity threshold corresponds to the minimum acceptable standard for the affinity of candidate rich element combinations in terms of the material dimension of rare elements. Then, in the rich element combination screening space, all coordinate points whose process similarity values ​​on the horizontal axis reach or exceed the preset process similarity threshold, and whose material similarity values ​​on the vertical axis reach or exceed the preset material similarity threshold, are selected. The region containing these coordinate points that meet both threshold requirements is defined as the rich element combination screening subspace. This subspace retains only candidate rich element combinations whose process and material similarity both meet the basic acceptable standards, narrowing the scope and focusing the target for subsequent iterative screening.

[0037] From the constructed rich-element combination screening subspace, the center coordinate point and the optimal coordinate point of the subspace are extracted. The center coordinate point is determined by the average process similarity and average material similarity of the coordinate points corresponding to all candidate rich-element combinations within the subspace. The optimal coordinate point is the coordinate point within the subspace where both process similarity and material similarity reach their highest values. Then, using the direction from the center coordinate point to the optimal coordinate point as the iteration direction, a mean-shifting iteration operation is performed on the center coordinate point according to a pre-set iteration step size to determine the iterated coordinate point. Next, it is determined whether the iteration neighborhood density of the initial center coordinate point is greater than or equal to the iteration neighborhood density of the iterated coordinate point. The iteration neighborhood density refers to the number of coordinate points corresponding to other candidate rich-element combinations within a certain range around the coordinate point. If the iteration neighborhood density of the initial center coordinate point is not greater than or equal to the iteration neighborhood density of the iterated coordinate point, the mean-shifting iteration is performed again along the same iteration direction, and this process is repeated until the preset iteration stopping condition is met. The preset iteration stopping conditions include reaching a pre-set number of iterations, or the iteration neighborhood density of a coordinate point obtained in a certain iteration being less than the iteration neighborhood density of the coordinate point obtained in the previous iteration. When the iteration stopping conditions are met, the final target coordinate point is determined. Then, based on the target coordinate point and the pre-set iteration step size, all coordinate points whose distance from the target coordinate point is within the iteration step size range are extracted from the rich element combination screening subspace. The candidate rich element combinations corresponding to these coordinate points together constitute the set of rich element combinations to be verified.

[0038] In one possible implementation, step S143 further includes: Step S1431: Extract the center coordinate point and the optimal coordinate point of the rich element combination screening subspace.

[0039] Step S1432: Using the direction from the center coordinate point to the optimal coordinate point as the iteration direction, perform mean drift iteration on the center coordinate point according to the preset iteration step size to determine the iteration coordinate point.

[0040] Step S1433: Determine whether the iterative neighborhood density of the center coordinate point is greater than or equal to the iterative neighborhood density of the iterative coordinate point. If not, continue to perform mean drift iteration on the iterative coordinate point according to the iterative direction until the preset iteration stop condition is met and the target coordinate point is obtained.

[0041] Step S1434: Based on the target coordinate point and the preset iteration step size, extract the rich element combination screening subspace to obtain the set of rich element combinations to be verified.

[0042] Specifically, for the constructed rich-element combination screening subspace, the mean process similarity and mean material similarity of the coordinate points corresponding to all candidate rich-element combinations within the subspace are calculated. The coordinate points corresponding to these two means are determined as the center coordinate points of the screening subspace, thus reflecting the overall distribution center of the candidate combinations within the subspace. Next, according to a preset weight ratio, such as a process similarity weight of 0.6 and a material similarity weight of 0.4, the process similarity and material similarity of each coordinate point are weighted and calculated to obtain the weighted comprehensive value of each coordinate point. Then, the coordinate point with the largest weighted comprehensive value is selected from all coordinate points and determined as the optimal coordinate point of the screening subspace. This point represents the position of the candidate rich-element combination with the best overall performance in terms of process and material similarity within the subspace.

[0043] The rules for setting the iteration direction are clearly defined. Starting from the central coordinate point extracted within the rich element combination screening subspace, and ending at the optimal coordinate point, the direction of the line connecting these two points is the direction of this iteration. This direction ensures that the iteration process always advances towards the region with better overall performance in terms of process and material similarity. Next, according to a pre-set iteration step size (e.g., 0.1 for both process and material similarity dimensions), a mean-shifting iteration operation is performed on the central coordinate point: the difference between the central and optimal coordinate points in both process and material similarity dimensions is calculated. Then, based on the iteration step size, the movement range in each dimension is determined, and the central coordinate point is moved along the iteration direction by this range to obtain a new coordinate point. Finally, by verifying whether the process and material similarity of this new coordinate point conforms to the adjustment logic corresponding to the iteration step size, its validity is confirmed, and this valid new coordinate point is ultimately determined as the iteration coordinate point.

[0044] The definition of iterative neighborhood density is clarified: it refers to delineating a certain range of neighborhood around a coordinate point, such as a region centered on the coordinate point with radii of 0.05 for both process similarity and material similarity. The number of coordinate points corresponding to other candidate element-rich combinations contained within this neighborhood is counted; this number is the iterative neighborhood density of that coordinate point. Next, the iterative neighborhood densities of the initial center coordinate point and the iterative coordinate point are compared: if the iterative neighborhood density of the center coordinate point is greater than or equal to that of the iterative coordinate point, it indicates that the concentration of candidate combinations within the neighborhood has not improved under the current iteration direction, and iteration in the current direction can be paused; if the iterative neighborhood density of the center coordinate point is less than that of the iterative coordinate point, it indicates that continuing along this iteration direction may yield a better concentration region of candidate combinations, and the current iterative coordinate point should be used as a new starting point, and mean-shifting iteration should be performed again according to the previously determined iteration direction and preset iteration step size to generate a new iterative coordinate point. Then repeat the neighborhood density comparison and iteration operation until the preset iteration stop condition is met, the preset maximum number of iterations is reached, or the iteration neighborhood density of the new coordinate point generated in a certain iteration is less than the iteration neighborhood density of the coordinate point in the previous iteration. At this time, the iteration stops, and the coordinate point obtained in the last iteration is determined as the target coordinate point.

[0045] On the horizontal axis of the rich element combination screening subspace, the extraction interval for process similarity is formed by extending a distance corresponding to a preset iteration step size in both directions of increasing and decreasing process similarity, centered on the process similarity value corresponding to the target coordinate point. Similarly, on the vertical axis, the extraction interval for material similarity is formed by extending a distance corresponding to a preset iteration step size in both directions of increasing and decreasing material similarity, centered on the material similarity value corresponding to the target coordinate point. The extraction intervals of the two axes together constitute a rectangular extraction area, which is the target range for this subspace extraction. Next, the coordinate points corresponding to all candidate rich element combinations within the rich element combination screening subspace are traversed, and it is determined whether the process similarity and material similarity of each coordinate point fall within the process similarity extraction interval and the material similarity extraction interval, respectively. Finally, all candidate rich element combinations corresponding to coordinate points that simultaneously meet the requirements of both intervals are selected, and these combinations are organized and summarized to form the final set of rich element combinations to be verified.

[0046] In one possible implementation, step S1433 further includes: The preset iteration stopping condition is that the preset number of iterations and / or the iteration neighborhood density of the current iteration is less than the iteration neighborhood density of the previous iteration.

[0047] Specifically, the preset iteration stop conditions include two triggerable scenarios, either of which will terminate the iteration process. The first scenario is a preset number of iterations. When the number of mean drift iterations of coordinate points within the rich element combination screening subspace reaches a predetermined fixed number, the iteration stops regardless of the current iteration effect, ensuring the iteration process does not proceed indefinitely. The second scenario is a change in the iteration neighborhood density. The iteration neighborhood density of the coordinate points obtained in the current iteration is compared with that of the coordinate points obtained in the previous iteration. If the iteration neighborhood density of the current iteration is less than that of the previous iteration, it indicates that continuing iteration along the current direction can no longer increase the concentration of candidate rich element combinations within the neighborhood; instead, it decreases, and the iteration must also stop in this case. By setting these two conditions, both excessive resource consumption during iteration can be avoided, and the effectiveness of the iteration results can be guaranteed.

[0048] In one possible implementation, step S200 further includes: Step S210: Determine whether the virtual representation of the rich element combination set to be verified satisfies the dual objectives. If not, determine the objective deviation.

[0049] Step S220: Based on the target deviation, fine-tune the set of rich element combinations to be verified according to a preset adjustment method to obtain a set of rich element combinations to be verified after one adjustment.

[0050] Step S230: Perform dual-target verification on the virtual representation of the rich element combination set to be verified once. After multiple verifications and adjustments, obtain the rich element combination set to be verified.

[0051] Specifically, virtual simulation software is used to perform performance simulation calculations on the set of rich element combinations to be verified. Similarity data for each combination and the scarce element in key performance dimensions such as resistivity, transmittance, and reflectivity are obtained. Simultaneously, statistical tools are used to calculate the coefficients of variation for all combinations in terms of element ratios and performance parameters, thereby quantifying the performance similarity and combination difference in the virtual representation. Next, a pre-defined dual-objective threshold database is accessed to extract the acceptable thresholds for performance similarity (e.g., ≥90%) and combination difference (e.g., coefficient of variation ≥0.1). The calculated mean performance similarity is compared with the performance thresholds, and the combination difference is compared with the difference thresholds. If the mean performance similarity or combination difference does not reach the threshold, the dual objective is not met. In this case, deviation calculation is initiated: the difference between the performance similarity threshold and the actual mean is multiplied by a performance weight (e.g., 0.6), and the difference between the combination difference threshold and the actual value is multiplied by a difference weight (e.g., 0.4). The weighted result is the target deviation.

[0052] The target deviation data is retrieved to determine whether the deviation primarily stems from insufficient performance similarity or inadequate combination difference. Simultaneously, preset adjustment methods are extracted from a pre-defined parameter library, and element proportions are adjusted according to fixed preset adjustment ranges in the rich element combination set to be verified. For example, each adjustment adjusts key elements affecting performance by ±5% and auxiliary elements affecting difference by ±3%. Subsequently, for each combination in the rich element combination set to be verified, if the target deviation mainly stems from insufficient performance similarity, the proportions of elements highly correlated with the performance of scarce elements, such as metal elements affecting resistivity, are increased or decreased according to the adjustment range based on performance simulation data to improve performance similarity. If the deviation mainly stems from excessively low combination difference, elements with similar proportions in each combination are adjusted differentially by the adjustment range to widen the element proportion differences between combinations and enhance difference. After completing the fine-tuning of element proportions for all combinations, the adjusted combinations are integrated to form the first-adjustment rich element combination set to be verified.

[0053] Following the dual-objective verification logic and implementation method, virtual simulation software is used to calculate the performance similarity between each combination in the set of rich element combinations to be verified and the scarce element. Data statistical tools are used to analyze the differences (dispersion coefficients) between combinations, and these are compared with preset performance similarity and difference thresholds. If the verification results show that the set meets the dual objectives—that is, the performance similarity meets the standard and the difference meets the requirements—it is directly identified as the set of rich element combinations to be verified for fine-tuning. If not, the deviation calculation process is restarted. After determining the new target deviation, the element ratio is adjusted according to a preset range, and the set of rich element combinations to be verified for fine-tuning is fine-tuned again, resulting in the set of rich element combinations to be verified for secondary adjustment. This cycle of "virtual representation calculation → dual-objective verification → deviation calculation → element ratio fine-tuning" is repeated until a set of rich element combinations after one adjustment passes the dual-objective verification. At this point, the loop stops, and the adjusted set is finally identified as the set of rich element combinations to be verified for fine-tuning.

[0054] In one possible implementation, step S220 further includes: The default adjustment method is to adjust the element ratio of the rich element combination set to be verified according to the default adjustment range.

[0055] Specifically, during the adaptive testing and optimization process of element replacement, when fine-tuning the set of rich element combinations to be validated is required, the preset adjustment method involves adjusting the element proportions according to a preset adjustment range. First, a preset adjustment range is extracted from the system parameter configuration. For example, a ±5% adjustment range is set for elements affecting core performance, and a ±3% adjustment range is set for auxiliary elements, clearly defining the adjustment range limits for different types of elements. Next, based on the target deviation of the set of rich element combinations to be validated, the main source of deviation is determined. If the deviation stems from substandard performance caused by a certain type of element, the proportion of that type of element is increased or decreased by a preset range. For example, if performance similarity is insufficient, the proportion of elements with high performance correlation to scarce elements is increased. If the deviation stems from excessively low differences between combinations, elements with similar proportions in each combination are adjusted differentially by a preset range. For example, the proportion of a certain element is increased in some combinations, and decreased in others. During the adjustment process, it is ensured that the total element proportion of each combination remains 100% to avoid imbalance. Finally, the element proportions of all rich element combinations to be validated are fine-tuned, forming the adjusted set of rich element combinations.

[0056] In one possible implementation, step S400 further includes: Step S410: Obtain the production feedback monitoring window.

[0057] Step S420: The target fine-tuned rich element combination is prepared into materials in the production feedback monitoring window, and the prepared materials are evaluated for quality. If the quality evaluation result is less than the preset quality requirement, a test warning instruction is generated.

[0058] Specifically, the criteria for defining the production feedback monitoring window are clarified. Based on the target material production characteristics corresponding to the fine-tuned element-rich combination, the monitoring time range and production batch range are determined. The time range is typically selected to cover a period that can complete the entire production cycle, such as three consecutive working days, ensuring that all key stages of material preparation are included. The production batch range is set to a statistically significant number, such as 6-10 consecutive production batches, to avoid the randomness of data from a single batch. Subsequently, by analyzing key nodes in the production process, such as alloy batching, casting, rolling, and finished product inspection, the key production links and data types that need to be tracked within the monitoring window are identified, such as process parameters at each stage, and quality inspection data of semi-finished and finished products. Finally, a monitoring window that can accurately capture production feedback information of the target element-rich combination is defined, providing a clear data collection boundary for subsequent quality assessment.

[0059] During the defined production feedback monitoring window, material preparation is carried out using the target fine-tuning rich element combination according to the standard process of target material manufacturing. From alloy batching, melting and casting, rolling to forming, each step strictly follows the process parameters suitable for the rich element combination. After the material preparation is completed, based on the preset quality assessment indicators, such as the density, purity, surface smoothness, and internal density of the target material, a comprehensive quality inspection is carried out on each batch of finished products using testing methods. All test data are recorded and the quality assessment results are calculated. The calculated quality assessment results are compared with the preset quality requirements. If the quality assessment results of one or more batches of materials do not meet the preset quality requirements, it is determined that there is a deviation in the previous test optimization for the target fine-tuning rich element combination. At this time, a test warning instruction needs to be generated, clearly indicating that it is necessary to return to the early stage of element substitution adaptability testing and redo the virtual characterization reverse fine-tuning, sputtering experiment and other test optimization work to ensure that the quality of the subsequently prepared materials meets the production requirements.

[0060] Example 2, based on the same inventive concept as the adaptive testing optimization method for element substitution in the foregoing examples, such as... Figure 2 As shown, this application provides an adaptive testing optimization system for element substitution. The system and method embodiments in this application are based on the same inventive concept. The system includes: The rich element combination set acquisition module 10 is used to extract the performance data and type of scarce elements in the target element replacement scene information, and perform first-order multi-dimensional filtering in the rich element resource library to obtain the rich element combination set to be verified.

[0061] The reverse fine-tuning module 20 is used to perform reverse fine-tuning of the virtual representation of the set of rich element combinations to be verified based on dual objectives, so as to obtain a set of rich element combinations to be verified and fine-tuned. The dual objectives include that the performance similarity between the rich element combination to be verified and the scarce element meets a preset performance requirement, and that the distribution concentration between the virtual representation of the set of rich element combinations to be verified is less than a preset distribution concentration requirement.

[0062] The sputtering experiment module 30 is used to traverse the set of elements rich in the fine-tuning to be verified to prepare samples and perform sputtering experiments, so as to obtain the set of samples of elements rich in the fine-tuning to be verified and the set of sputtering experiment results.

[0063] The test evaluation module 40 is used to test and evaluate the sample set of the fine-tuned rich element combination to be verified according to a preset performance test type when the sputtering test result set is a pass test, and to determine the target fine-tuned rich element combination based on the test evaluation results.

[0064] Furthermore, the system is also used to implement the following functions: The preset performance test types include resistivity performance test, transmittance performance test, reflectivity performance test, hardness performance test, and adhesion performance test.

[0065] Furthermore, the system is also used to implement the following functions: Constrained by the performance data and types of scarce elements, the rich element resource library is subjected to first-order multidimensional screening from both material and process dimensions to determine the material-similar rich element set and the process-similar rich element set; the material-similar rich element set and the process-similar rich element set are added to the candidate rich element pool; the candidate rich element pool is combined with a random number and random proportion of candidate rich elements to obtain the candidate rich element combination set; the candidate rich element combination set is subjected to stepwise directional iterative screening to obtain the rich element combination set to be verified.

[0066] Furthermore, the system is also used to implement the following functions: A rich element combination screening space is constructed based on the candidate rich element combination set, wherein the horizontal axis of the rich element combination screening space represents process similarity and the vertical axis represents material similarity, and each coordinate point corresponds to a candidate rich element combination; in the rich element combination screening space, a rich element combination screening subspace is constructed with preset process similarity thresholds and preset material similarity thresholds as constraints; stepwise directional iterative screening is performed in the rich element combination screening subspace to obtain the set of rich element combinations to be verified.

[0067] Furthermore, the system is also used to implement the following functions: Extract the center coordinate point and the optimal coordinate point of the rich element combination screening subspace; using the distance from the center coordinate point to the optimal coordinate point as the iteration direction, perform mean drift iteration on the center coordinate point according to a preset iteration step size to determine the iteration coordinate point; determine whether the iteration neighborhood density of the center coordinate point is greater than or equal to the iteration neighborhood density of the iteration coordinate point; if not, continue to perform mean drift iteration on the iteration coordinate point according to the iteration direction until the preset iteration stopping condition is met to obtain the target coordinate point; based on the target coordinate point and the preset iteration step size, extract the rich element combination screening subspace to obtain the set of rich element combinations to be verified.

[0068] Furthermore, the system is also used to implement the following functions: The preset iteration stopping condition is that the preset number of iterations and / or the iteration neighborhood density of the current iteration is less than the iteration neighborhood density of the previous iteration.

[0069] Furthermore, the system is also used to implement the following functions: Determine whether the virtual representation of the rich element combination set to be verified satisfies the dual objectives. If not, determine the objective deviation. Based on the objective deviation, fine-tune the rich element combination set to be verified according to a preset adjustment method to obtain a rich element combination set to be verified after one adjustment. Perform dual objective verification on the virtual representation of the rich element combination set to be verified after one adjustment again. After multiple verification adjustments, obtain the rich element combination set to be verified after fine-tuning.

[0070] Furthermore, the system is also used to implement the following functions: The default adjustment method is to adjust the element ratio of the rich element combination set to be verified according to the default adjustment range.

[0071] Furthermore, the system is also used to perform the following functions: acquire a production feedback monitoring window; prepare materials for the target fine-tuned rich element combination in the production feedback monitoring window, and evaluate the quality of the prepared materials; if the quality evaluation result is less than the preset quality requirements, generate a test warning command.

[0072] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0073] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0074] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for adaptive testing and optimization of element substitution, characterized in that, The method includes: Extract the performance data and type of scarce elements in the target element replacement scene information, and perform first-order multi-dimensional filtering in the rich element resource library to obtain a set of rich element combinations to be verified. Based on dual objectives, the virtual representation of the set of rich element combinations to be verified is reverse-fine-tuned to obtain a set of rich element combinations to be verified and fine-tuned. The dual objectives include that the performance similarity between the rich element combination to be verified and the scarce element meets a preset performance requirement, and that the distribution concentration between the virtual representation of the set of rich element combinations to be verified is less than a preset distribution concentration requirement. The sample preparation and sputtering experiments are carried out by traversing the set of element-rich combinations to be verified, so as to obtain the sample set of element-rich combinations to be verified and the sputtering experiment result set. When the sputtering experiment results set is a pass, the sample set of the element-rich combination to be verified is tested and evaluated according to the preset performance test type, and the target element-rich combination to be fine-tuned is determined based on the test evaluation results.

2. The adaptive testing and optimization method for element replacement as described in claim 1, characterized in that, The preset performance test types include resistivity performance test, transmittance performance test, reflectivity performance test, hardness performance test, and adhesion performance test.

3. The adaptive testing and optimization method for element replacement as described in claim 1, characterized in that, Extract the performance data and type of scarce elements from the target element replacement scene information, and perform first-order multi-dimensional filtering in the rich element resource library to obtain a set of rich element combinations to be verified, including: Using the performance data and types of scarce elements as constraints, the rich element resource library is screened from the material dimension and the process dimension respectively to determine the material-similar rich element set and the process-similar rich element set. Add the material-similar rich element set and the process-similar rich element set to the candidate rich element pool; The candidate rich element pool is combined with a random number and a random proportion of candidate rich elements to obtain a set of candidate rich element combinations; The candidate rich element combination set is subjected to stepwise directional iterative screening to obtain the rich element combination set to be verified.

4. The method for adaptive testing and optimization of element substitution as described in claim 3, characterized in that, The candidate rich element combination set is subjected to stepwise directional iterative screening to obtain the rich element combination set to be verified, including: A rich element combination screening space is constructed based on the candidate rich element combination set, wherein the horizontal axis of the rich element combination screening space is the process similarity, the vertical axis is the material similarity, and each coordinate point corresponds to a candidate rich element combination. In the rich element combination screening space, a rich element combination screening subspace is constructed with preset process similarity threshold and preset material similarity threshold as constraints; A stepwise directional iterative screening is performed in the rich element combination screening subspace to obtain the set of rich element combinations to be verified.

5. The method for adaptive testing and optimization of element substitution as described in claim 4, characterized in that, A stepwise directional iterative screening is performed in the rich element combination screening subspace to obtain the set of rich element combinations to be verified, including: Extract the center coordinates and optimal coordinates of the rich element combination screening subspace; Using the direction from the center coordinate point to the optimal coordinate point as the iteration direction, and according to the preset iteration step size, the center coordinate point is subjected to mean drift iteration to determine the iteration coordinate point; Determine whether the iterative neighborhood density of the center coordinate point is greater than or equal to the iterative neighborhood density of the iterative coordinate point. If not, continue to perform mean drift iteration on the iterative coordinate point according to the iterative direction until the preset iteration stop condition is met to obtain the target coordinate point. Based on the target coordinates and the preset iteration step size, the rich element combination screening subspace is extracted to obtain the set of rich element combinations to be verified.

6. The method for adaptive testing and optimization of element substitution as described in claim 5, characterized in that, The preset iteration stopping condition is that the preset number of iterations and / or the iteration neighborhood density of the current iteration is less than the iteration neighborhood density of the previous iteration.

7. The adaptive testing and optimization method for element substitution as described in claim 1, characterized in that, Based on dual objectives, the virtual representation of the set of rich element combinations to be verified is reverse-fine-tuned to obtain a set of rich element combinations to be verified and fine-tuned. The dual objectives include: the performance similarity between the rich element combinations to be verified and the scarce elements meets a preset performance requirement; and the distribution concentration between the virtual representations of the set of rich element combinations to be verified is less than a preset distribution concentration requirement. Determine whether the virtual representation of the rich element combination set to be verified satisfies the dual objectives; if not, determine the objective deviation. Based on the target deviation, the set of rich element combinations to be verified is fine-tuned according to a preset adjustment method to obtain a set of rich element combinations to be verified after one adjustment. The virtual representation of the rich element combination set to be adjusted once is verified again with dual objectives. After multiple verifications and adjustments, the rich element combination set to be fine-tuned is obtained.

8. The method for adaptive testing and optimization of element substitution as described in claim 7, characterized in that, The default adjustment method is to adjust the element ratio of the rich element combination set to be verified according to the default adjustment range.

9. The method for adaptive testing and optimization of element substitution as described in claim 1, characterized in that, include: Access the production feedback monitoring window; The target fine-tuned element-rich combination is prepared into materials in the production feedback monitoring window, and the quality of the prepared materials is evaluated. If the quality evaluation result is less than the preset quality requirement, a test warning instruction is generated.

10. An adaptive testing and optimization system for element substitution, characterized in that, The system is used to implement the adaptive testing optimization method for element substitution as described in any one of claims 1-9, the system comprising: The rich element combination set acquisition module is used to extract the performance data and type of scarce elements in the target element replacement scene information, and perform first-order multi-dimensional filtering in the rich element resource library to obtain the rich element combination set to be verified. The reverse fine-tuning module is used to perform reverse fine-tuning of the virtual representation of the set of rich element combinations to be verified based on dual objectives, so as to obtain the set of rich element combinations to be verified and fine-tuned. The dual objectives include the performance similarity between the rich element combination to be verified and the scarce element meeting the preset performance requirements, and the distribution concentration between the virtual representation of the set of rich element combinations to be verified being less than the preset distribution concentration requirements. The sputtering experiment module is used to traverse the set of elements-rich combinations to be verified for sample preparation and sputtering experiments, and to obtain the set of sample sets of elements-rich combinations to be verified for sputtering experiments and the set of sputtering experiment results. The test evaluation module is used to test and evaluate the sample set of the element-rich combination to be verified according to a preset performance test type when the sputtering test result set is a pass test, and to determine the target element-rich combination to be fine-tuned based on the test evaluation results.

Images