Curve synchronous adjustment method

Abstract

The present application relates to the technical field of detection imaging analysis, and provides a curve synchronous adjustment method, comprising the following steps: proportion reference establishment, target curve modification detection, modification information extraction, to-be-modified associated curve determination, determination of adjustment parameters and batch adjustment of curves. The curve adjustment method provided by the present application realizes automatic synchronous adjustment of multiple curves through a batch execution engine, greatly shortens the adjustment time, improves the curve adjustment efficiency, and improves the adjustment accuracy.

Description

Technical Field

[0001] This invention relates to the field of detection imaging analysis technology, and in particular to a method for synchronous curve adjustment. Background Technology

[0002] In testing and analysis, such as scientific research data analysis and industrial testing, when there are multiple related analytical items, it is often necessary to generate multiple curves for comparative analysis. These curves are usually layered and aligned. For example... Figure 1 As shown, this is a data result graph that has been layered and matched. The structure of each curve element is different, and the depth and number of points are also different.

[0003] When analysts learn that the previous reference curve parameters were entered incorrectly, or when parameter adjustments are needed for analysis, they may modify the parameters of a certain curve, such as modifying the depth data of a certain curve. To ensure the alignment between curves, other curves need to be adjusted accordingly.

[0004] In existing technical solutions, when one curve is modified and other curves need to be adjusted synchronously to adapt to new analysis requirements, a manual adjustment mode is used. The operator must first determine the modification range of the target curve as the target curve, and then manually input the adjustment values ​​to complete the modification of a single curve by adjusting the various layer parameters (such as depth, rate value, etc.) of other curves one by one. If there are many curves or the layering is complex, the above steps need to be repeated dozens or even hundreds of times.

[0005] This approach has the following significant drawbacks:

[0006] The operation is cumbersome: each layer of each associated curve requires manual operation, and the steps are highly repetitive, especially when the number of curves exceeds 5 or the number of layers exceeds 3, the operation process is lengthy.

[0007] Time-consuming: The process of manually calculating the proportions and inputting adjustment values ​​one by one is time-consuming. For a dataset containing 10 curves and 4 layers, a single adjustment takes more than 10 minutes, which seriously affects the data processing efficiency.

[0008] High risk of error: When manually calculating the ratio and inputting adjustment values, calculation errors or input deviations are easily caused by fatigue or negligence, resulting in a decrease in curve alignment accuracy, which requires repeated verification and correction. Summary of the Invention

[0009] Based on the problems existing in the prior art, this application provides a curve synchronization adjustment method to solve the problems of low efficiency and high error risk caused by the need for manual modification in the curve adjustment provided by the prior art.

[0010] The curve synchronization adjustment method provided in this application includes the following steps:

[0011] S100. Establish the scale benchmark. Take a set of curves that have been aligned in layers as the benchmark group, identify and extract the initial scale parameters of different layers between the curves in the benchmark group, and form the initial scale matrix.

[0012] S200, Target Curve Modification Detection: Real-time detection of whether any curves have been modified. When a parameter modification of a certain curve is detected, the modified curve is taken as the target curve, and the information of the target curve is recorded.

[0013] S300, Modification Information Extraction: Analyze the modified curve data of the target curve, and extract at least the hierarchical index information, modified value size information, and modification range information involved in the modification.

[0014] S400. Determine the correlation curve to be modified. Based on the information of the target curve and the scale benchmark, determine the correlation curve to be modified through dimensional matching.

[0015] S500. Determine the adjustment parameters. Based on the initial ratio matrix, obtain the adjustment ratio value of each related curve to be modified and the modification information of the target curve, and generate the adjustment parameters for each related curve to be modified.

[0016] S600, Batch Adjust Curves: Start the batch execution engine and synchronously adjust all related curves to be modified according to the generated adjustment parameters.

[0017] The curve synchronization adjustment method provided in this application establishes a proportional benchmark, providing a basis for obtaining a set of curve adjustment parameters among a group of aligned curves through mutual parameter comparison. When a parameter modification is detected on a target curve, by acquiring the modification information and the proportional benchmark, the parameter adjustment values ​​of other related curves in the same standard group as the target curve can be calculated. A batch execution engine enables automatic synchronization adjustment of multiple curves, replacing manual one-by-one operation and significantly shortening the adjustment time, especially in scenarios with a large number of curves, where the efficiency improvement is even more pronounced. This method solves the problem of calculation errors or input deviations due to fatigue or negligence when manually calculating proportions and inputting adjustment values, thus improving the accuracy of the adjustment.

[0018] Preferably, after forming the initial scaling matrix in step S100, the method further includes:

[0019] S110. Dynamic scaling model construction: Introducing time decay factor and hierarchical weight coefficient to construct dynamic scaling model;

[0020] S120, Scale Library Management: Establish a scale library system to manage dynamic scale models.

[0021] Preferably, step S400 specifically involves: based on the dynamic ratio model features in the ratio library system, performing curve type consistency verification, hierarchical structure similarity comparison, and historical adjustment record correlation analysis on the target curve and other related curves.

[0022] Preferably, in step S500, the adjustment ratio value is obtained through dynamic calculation. Specifically, the dynamic calculation involves calling a dynamic ratio model based on the extracted modification information and matching results, and combining the time decay factor and the layer weight coefficient to calculate the adjustment ratio value of each layer of the correlation curve to be modified in real time.

[0023] Preferably, after determining the adjustment parameters in step S500, an anomaly detection step is further included, specifically including:

[0024] Dual threshold detection: Set hard and soft thresholds for anomaly detection; the hard threshold is the limit range based on the physical characteristics of the curve. When the modified value of the adjustment parameter or the calculated adjustment value exceeds the hard threshold, a warning is triggered and the adjustment is rejected; the soft threshold calculates the normal fluctuation range based on the 3σ principle. When the adjustment value deviation exceeds 2σ, a backup proportional value is used for adjustment.

[0025] Preferably, after the anomaly detection step, the following steps are also included:

[0026] Breakpoint recovery: Blockchain technology is used to store key ratio data and adjustment records. When the system experiences an abnormal interruption, it can be traced back to the most recent valid state based on the stored data.

[0027] Anomaly location and alerts: When an anomaly is detected, the source of the anomaly is located through log analysis and an alert is issued.

[0028] Preferably, after step S600, the model further includes step S700, which is a proportional model iterative upgrade, specifically including:

[0029] S710. Data collection and analysis: Collect modification record data, verification result data, and anomaly handling data during the batch adjustment process; analyze the applicability and accuracy of the proportional model through machine learning models; and identify the deviation patterns of the proportional model in different scenarios.

[0030] S720. Model Optimization and Upgrade: Optimize the proportional model based on the analysis results, including adjusting the time decay factor coefficient, updating the hierarchical weight coefficient, and adding at least one of the following: proportional calculation rules for special scenarios.

[0031] This application also provides an electronic device, including a memory and a processor. The memory stores a computer-executable program, and the processor calls the computer-executable program in the memory to implement the above-described curve synchronization adjustment method.

[0032] This application embodiment also provides a storage medium, which is a computer-readable storage medium, and a computer-executable program is stored on the computer-readable storage medium. When the computer-executable program is executed by a processor, it implements the above-described curve synchronization adjustment method.

[0033] This application also provides a computer program product for implementing curve synchronization adjustment, including code for implementing the above-described curve synchronization adjustment method. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a schematic diagram of multiple correlated curves generated in the existing technology detection and analysis.

[0036] Figure 2 This is a schematic diagram of the curve synchronization adjustment method in the embodiments of this application;

[0037] Figure 3 This is a schematic diagram of the target curve detection process provided in an embodiment of this application;

[0038] Figure 4 This is a schematic diagram of the state backtracking process in an embodiment of this application;

[0039] Figure 5 This is a schematic diagram illustrating the parameter adjustment process in an embodiment of this application. Detailed Implementation

[0040] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0041] It is important to note that terms such as "first," "second," "symmetric," and "array" are used only to distinguish between descriptive and positional descriptions and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, features specified with terms such as "first" or "symmetric" may explicitly or implicitly include one or more of that feature; similarly, when the quantity of certain features is not limited by words such as "two" or "three," it should be noted that such features also explicitly or implicitly include one or more features.

[0042] In this invention, unless otherwise explicitly specified and limited, terms such as "installation," "connection," and "fixation" should be interpreted broadly; for example, they can refer to a fixed connection, a detachable connection, or an integral molding; they can refer to a mechanical connection, a direct connection, a welding connection, or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the accompanying drawings and specific circumstances.

[0043] The technical solution of this application will now be described in detail with reference to the accompanying drawings.

[0044] like Figure 2 As shown in the figure, this application provides a method for curve synchronization adjustment, which specifically includes the following steps:

[0045] S100. Establish the scale benchmark. Take a set of curves that have been aligned in layers as the benchmark group, identify and extract the initial scale parameters of different layers between the curves in the benchmark group, and form the initial scale matrix.

[0046] The establishment of the proportional benchmark provided in this application specifically includes the following steps:

[0047] The parameter extraction module selects a set of verified depth data curves as a baseline group, identifies and extracts the initial scale parameters of different layers between each curve, including the layer speed scale and depth scale, to form an initial scale matrix.

[0048] In the embodiments provided in this application, for a group of curves that have been layered and aligned (i.e., whose depth data has been verified), any one of the curves can be selected as the reference curve, and the other curves are calculated with respect to the reference curve. The ratio of the i-th layer of the j-th curve to the reference curve is defined as follows:

[0049]

[0050] Where: R i,j V represents the ratio of the i-th layer of the j-th curve to the baseline curve. i,jV represents the i-th layer parameter value (e.g., depth) for the j-th curve. i,base : The i-th layer parameter value of the baseline curve.

[0051] Suppose we have four curves (A, B, C, D) that have already been aligned and layered, each curve containing three layers, with the following depth data:

[0052] A 100 200 300 B 120 240 360 C 80 160 240 D 110 220 330

[0053] Step S101: Set the reference curve: Select curve A as the reference curve (base = A).

[0054] Step S102: Calculate the layer ratio of each curve to the baseline curve to form a ratio matrix:

[0055] The ratio matrix of curves B and A:

[0056]

[0057] The ratio matrix of curves C and A:

[0058]

[0059] The ratio matrix of curves D to A:

[0060]

[0061] Step S103: Rearrange into a proportional matrix:

[0062]

[0063] (Rows: Layer 1-3, Columns: Curves B-D).

[0064] Then, curve B can be selected as the reference curve, and the ratio matrix between curve B and other curves in the reference group can be calculated. The specific calculation steps can be found in the example with curve A above, and will not be repeated here.

[0065] S200, Target Curve Modification Detection: Real-time detection of any curve modification operations. When a parameter modification of a certain curve is detected, the modified curve is taken as the target curve, and the information of the target curve is recorded.

[0066] The target curve modification detection provided in this application can be found in [reference needed]. Figure 3When a user modifies the parameters of a curve in the operation interface, such as changing the depth of the intermediate layer of curve C001 from 200um to 250um, the data interface layer monitors all curve modification operations in real time. When a change in the parameters (such as depth data) of curve C001 is detected, the curve in which the modification operation occurred is designated as the target curve. The modification information of the target curve is recorded, including the curve identifier, layer index, values ​​before and after the modification, modification time, and other key information.

[0067] S300, Modification Information Extraction: Analyze the modified curve data of the target curve, and extract at least the hierarchical index information, modified value size information, and modification range information involved in the modification.

[0068] S400. Determine the correlation curve to be modified. Based on the information of the target curve and the scale benchmark, determine the correlation curve to be modified through dimensional matching.

[0069] S500. Determine the adjustment parameters. Based on the initial ratio matrix, obtain the adjustment ratio value of each associated curve to be modified and the modification information of the target curve, and generate the adjustment parameters for each associated curve to be modified.

[0070] When the value of the i-th layer parameter of the benchmark curve is from Modified to At that time, the synchronous adjustment formula for the correlation curve is:

[0071]

[0072] in: The adjusted value of the i-th layer of the j-th curve.

[0073] Taking the aforementioned set of reference curves (A, B, C, and D) as an example, assuming the user modifies the intermediate layer depth of reference curve A (from 200μm to 220μm), then:

[0074] Depth of the intermediate layer of curve B after adjustment:

[0075]

[0076] Depth of the intermediate layer after adjustment for curve C:

[0077]

[0078] Depth of the intermediate layer of curve D after adjustment:

[0079]

[0080] Final result after adjustment:

[0081] A 100 220 300 B 120 264 360 C 80 176 240 D 110 242 330

[0082] S600, Batch Adjust Curves: Start the batch execution engine and synchronously adjust all related curves to be modified according to the generated adjustment parameters.

[0083] In the preferred embodiment provided in this application, after establishing the proportional benchmark, step S100 further establishes a dynamic proportional benchmark, specifically including the following steps:

[0084] S110. Dynamic scaling model construction: Introducing time decay factor and hierarchical weight coefficient to construct dynamic scaling model.

[0085] A dynamic scaling model is constructed by introducing a time decay factor and a hierarchical weighting coefficient. The time decay factor is calculated according to the formula R(t)=R0×e^(-kt), where R0 is the initial scaling factor, k is the decay coefficient, which can be set according to the equipment characteristics, and t is the data acquisition interval time, used to correct the scaling deviation caused by equipment drift; the hierarchical weighting coefficient W... i Based on historical data fitting, the dynamic proportion is used to distinguish the importance of different strata. The formula for calculating the dynamic proportion is V. ji ×V templati ×(R ij ×W i ×R(t)), where V ji V is the correction value for the i-th layer of the j-th curve. templati R represents the modified parameter values ​​for the i-th layer of the target curve. ij The initial scale is used as the reference scale. By introducing a time decay factor and a layer weight coefficient, a dynamic scaling model is constructed to achieve real-time correction of the curve scale, solving the problem of insufficient accuracy under equipment drift and complex layering scenarios, and ensuring the long-term synchronous alignment accuracy of the curve.

[0086] The time decay factor is set based on the following factors:

[0087] 1. Measurement stability

[0088] Feature Description: The ability of a device to maintain measurement accuracy during long-term operation. Devices with poor stability exhibit more pronounced drift over time (such as sensor zero-point shift due to temperature changes).

[0089] Typical equipment: precision mass spectrometer, laser rangefinder.

[0090] 2. Response latency characteristics

[0091] Feature Description: The device's response speed to parameter changes. High latency in devices can cause data timestamps to become out of sync with actual physical time during continuous sampling.

[0092] Typical equipment: electrochemical sensors, slow-response thermometers.

[0093] 3. Environmental sensitivity

[0094] Characteristic description: The degree to which the performance of the equipment is affected by the external environment (temperature, humidity, electromagnetic interference, etc.).

[0095] Typical equipment: optical microscope, semiconductor detector.

[0096] 4. Sampling frequency

[0097] Feature description: The number of times the device collects data per unit of time. Devices with low sampling frequencies may experience significant drift during the data acquisition interval.

[0098] Typical equipment: geological exploration equipment, meteorological station sensors.

[0099] How to set the attenuation coefficient (k value)

[0100] The attenuation coefficient k determines the rate at which the ratio decays over time; the larger the value, the faster the ratio correction. The k value needs to be set considering the device characteristics and application scenario. The specific method is as follows:

[0101] 1. Empirical values ​​based on equipment stability testing

[0102] Test method:

[0103] Under constant conditions, continuously collect curve data of the same object using equipment (e.g., 24-hour uninterrupted sampling);

[0104] Calculate the ratio between curves at fixed time intervals (e.g., 1 hour) and compare it with the initial ratio, recording the curve of the ratio deviation changing over time;

[0105] The fitted deviation curve follows an exponential decay model: deviation(t) = initial deviation × e -kt Solve for the value of k.

[0106] Example: A stability test of a mass spectrometer showed that its proportional deviation reached 50% of its initial value after 2 hours. Substituting this into the formula:

[0107] 2. Theoretical calculations based on equipment specifications

[0108] Formula method:

[0109] The hourly drift rate of the equipment can be obtained from the equipment specifications (e.g., "0.1% drift per hour").

[0110] Example: A laser rangefinder's specifications state "0.05% drift per hour," then:

[0111] 3. Adaptive dynamic adjustment

[0112] Method: During system operation, the difference between the currently calculated ratio and the historical ratio is compared in real time, and the k value is dynamically adjusted through a feedback mechanism.

[0113]

[0114] in:

[0115] ΔR: The absolute difference between the current ratio and the historical ratio;

[0116] R: Historical percentage value;

[0117] α: Learning rate (usually taken as 0.1-0.3).

[0118] Advantages: It automatically adapts to factors such as equipment aging and environmental changes, without the need for manual intervention.

[0119] Recommended range of k-values ​​for different device characteristics

[0120]

[0121]

[0122] The following example illustrates how to set the attenuation factor K in a SIMS device:

[0123] Secondary ion mass spectrometers (SIMS) are used for depth analysis of materials. When the equipment is operating continuously, the depth measurement drift is caused by fluctuations in the ion source, so it is necessary to set the attenuation coefficient correction ratio.

[0124] Setup steps

[0125] Equipment characteristic analysis:

[0126] Stability tests showed that after 8 hours of continuous operation, the depth measurement deviation reached 3%.

[0127] Sampling frequency: 2 minutes / point;

[0128] Environmental sensitivity: moderate (requires constant temperature environment).

[0129] k-value calculation:

[0130] Assuming an initial ratio of 1, and a deviation of 3% after 8 hours, the ratio becomes...

[0131] Practical applications:

[0132] Set k = 0.004 in the system;

[0133] When the user adjusts the curve, the system operates according to R(t) = R0 × e -0.004×tDynamic correction ratio (t is the data acquisition interval, in hours).

[0134] By introducing a time decay factor and a layer weight coefficient to construct a dynamic scaling model, the curve scaling is corrected in real time, which solves the problem of insufficient accuracy in equipment drift and complex layering scenarios, and ensures the long-term synchronous alignment accuracy of the curve.

[0135] S120, Scale Library Management: Establish a scale library system to manage dynamic scale models.

[0136] Establish a scale library system that supports the categorized storage (e.g., by curve type, application scenario), querying, editing, and version control of scale models, enabling the reuse and sharing of scale models.

[0137] The following is a proportional database system established for the analysis of a certain material, used to store dynamic proportional models of SIMS (secondary ion mass spectrometry) curves. Its core structure includes three levels:

[0138] Primary classification: by curve type (e.g., depth curve, concentration curve);

[0139] Secondary classification: By application scenario (e.g., semiconductor chip analysis, thin film material analysis);

[0140] Third-level classification: by sample type (e.g., silicon-based wafers, compound semiconductors).

[0141] Category storage example

[0142] 1. Storage path design

[0143] Proportional library system:

[0144]

[0145] 2. Model storage content (taking "Silicon-based wafer depth curve model V1.0" as an example)

[0146] Each model contains:

[0147] Basic information: Model name, creation time, creator, number of applicable curves (e.g., 5 curves), number of layers (e.g., 4 layers);

[0148] Core data: Initial scaling matrix (4 rows × 4 columns, corresponding to the layering ratio of the 4 correlation curves and the baseline curve), attenuation coefficient k value (set to 0.004 for SIMS device characteristics);

[0149] Associated files: original curve data snapshot, adjustment log, verification report (e.g., curve alignment deviation ≤ 1% after adjustment).

[0150] Example of query function

[0151] Users need to retrieve historical scaling models for "depth profile of silicon-based wafers (semiconductor chip analysis scenario)". The query process is as follows:

[0152] Multi-level filtering: In the system interface, select "Depth Curve → Semiconductor Chip Analysis → Silicon-based Wafer" in sequence;

[0153] Conditional filtering: Supports filtering by "creation time (e.g., the last 3 months)" and "applicable number of layers (e.g., 4 layers)" to display models that meet the conditions (e.g., V1.0 and V2.0);

[0154] Preview and Comparison: Click on the model to view details, such as the scale matrix of V1.0, k-value, and historical adjustment effects (with a visual curve comparison chart), to help determine whether it is applicable.

[0155] Editing function example

[0156] If a user finds that the attenuation coefficient k value (0.004) of the "Silicon-based Wafer Depth Curve Model V1.0" is too large on a new device, resulting in a large deviation after adjustment, the editing steps are as follows:

[0157] Model copy: Retrieve V1.0 from the library and select "Copy as new version" (avoid directly modifying the original model);

[0158] Parameter modification: Adjust the k value to 0.003 (based on the stability test results of the new device) and record the reason for the modification ("adapting to the drift characteristics of the new SIMS device (model XXX)");

[0159] Save and apply: Save the file to the original path with the name "V2.0", overwriting the original V2.0 or the new version, and generate a change log (including a comparison of parameters before and after the changes).

[0160] In the preferred embodiment provided in this application, step S400 determines the associated curve to be modified (hereinafter referred to as the associated curve) through multi-dimensional matching. Multi-dimensional matching is based on the proportional model features in the proportional library, and performs various checks on the target curve and other associated curves, including curve type consistency verification, hierarchical structure similarity comparison, and historical adjustment record correlation analysis, to ensure the accuracy of the matching.

[0161] The core of multi-dimensional matching is to ensure the accuracy of the matching between the target curve and the correlation curve through a three-level verification mechanism of "feature extraction → hierarchical verification → correlation analysis". The following is an example of the three-level verification process using the synchronous adjustment of the SIMS curve. The specific process is as follows:

[0162] Level 1 Validation: Curve Type Consistency Validation (Basic Screening):

[0163] 1. Feature Extraction

[0164] Retrieve the features of the scale model to be matched from the scale library, and extract the core attributes of the target curve and the associated curve to be modified:

[0165] Required features:

[0166] Curve type (e.g., "depth curve", "concentration curve", "sputtering rate curve");

[0167] Physical units (such as "A", "μm", "nm" for depth curves, and "atoms / cm" for concentration curves) 3 ”);

[0168] Sample type label (e.g., "silicon-based wafer", "compound semiconductor", "metal thin film").

[0169] 2. Consistency Judgment

[0170] If the target curve is completely identical to a related curve in terms of curve type, unit, and sample label, proceed to the next level of verification;

[0171] If any feature mismatch exists (e.g., the target curve is a "depth curve (μm)" and the associated curve is a "concentration curve (atoms / cm³)"), 3 The associated curve is directly excluded and marked as "mismatch".

[0172] Example as follows:

[0173] Target curve characteristics: Type = Depth curve, Unit = μm, Sample = Silicon-based wafer;

[0174] Correlation curve A characteristics: Type = depth curve, unit = μm, sample = silicon-based wafer → passed first-level verification;

[0175] Correlation curve B characteristics: Type = concentration curve, Unit = atoms / cm 3 Sample = silicon-based wafer → fails, excluded.

[0176] Secondary verification: Hierarchical structure similarity comparison (structural matching):

[0177] For the correlation curves that pass the first-level verification, verify the consistency of their hierarchical structure with the target curve:

[0178] 1. Hierarchical feature extraction

[0179] Number of layers (e.g., if the target curve has 3 layers, the associated curves must also have 3 layers);

[0180] Layer boundary values ​​(e.g., surface-intermediate layer boundary: target curve is 100μm, associated curve must be within the range of [90μm, 110μm], with an allowable error of ±10%).

[0181] Layering trend (such as the direction of change of each layer parameter with depth: the target curve shows an "increment" from the surface layer to the middle layer, and the associated curves must be consistent).

[0182] 2. Similarity Calculation

[0183] The structural similarity is quantified using "hierarchical matching degree":

[0184] If the matching degree is ≥80% (i.e., only 1 level of mismatch is allowed), proceed to the next level of verification;

[0185] If the match rate is less than 80%, it is marked as "low match" and manual confirmation is required to continue.

[0186] Example as follows:

[0187] The target curve is layered: 3 layers, with boundaries = [100μm, 200μm] and a trend of increasing.

[0188] Correlation curve A: 3 layers, boundary = [95μm, 205μm], trend = increasing → layer matching degree = 100% → pass;

[0189] Correlation curve C: 3 layers, boundary = [150μm, 250μm], trend = decreasing → layer matching degree = 0% → low matching, manual intervention.

[0190] Level 3 verification: Correlation analysis of historical adjustment records (empirical verification):

[0191] By analyzing historical data to determine whether there are records of coordinated adjustments between two curves, the reliability of the matching can be enhanced.

[0192] 1. Historical data extraction

[0193] Retrieve from system logs:

[0194] In the last three adjustment records, do the target curve and the associated curve belong to the same curve group?

[0195] Historical adjustment ratio deviation (e.g., after previous synchronous adjustments, whether the stratification ratio deviation between the two is ≤2%);

[0196] Operator tags (e.g., whether it has been historically tagged as a "correlation curve" by the same analyst).

[0197] 2. Relevance Score

[0198] Scoring will be based on the following rules (out of 100):

[0199] Adjustment record within the same group: +20 points each time (maximum 60 points);

[0200] Historical percentage deviation ≤2%: +20 points;

[0201] Manually labeled associations: +20 points.

[0202] A score of 60 or higher is considered "highly relevant" and the match is successful.

[0203] A score of <60 is considered "low relevance," prompting the user to confirm whether to force a match.

[0204] Example:

[0205] Correlation curve A: Adjusted in the same group as the target curve for the last 3 times (60 points), historical deviation 1.5% (20 points) → total score 80 points → high correlation, passed;

[0206] Correlation curve D: No records in the same group (0 points), historical deviation 5% (0 points) → total score 0 points → low correlation, manual confirmation required.

[0207] Matching result output and processing:

[0208] Automatic matching results:

[0209] At the same time, the correlation curves that pass the three-level verification are automatically included in the synchronization adjustment range and marked as "automatic matching passed";

[0210] Curves that fail any level of validation are marked as "mismatch" or "low match" and displayed in a separate list.

[0211] Human intervention mechanism:

[0212] For "low-match" curves, a visual comparison interface is provided (showing the mismatched features, such as the layer boundary difference map);

[0213] Provide a "force match" option and provide feedback to the user, receive user confirmation to perform forced matching, and record the reason for intervention (for subsequent proportional model iterations).

[0214] In step S500, when determining the adjustment parameters, dynamic scaling calculation is required first: based on the extracted modification information and matching results, the dynamic scaling model is called, and combined with the time decay factor and the layer weight coefficient, the scaling value that should be adjusted for each layer of other related curves is calculated in real time to ensure that the calibrated scaling can accurately reflect the current curve status.

[0215] Specific process and formula for dynamic scaling calculation (taking SIMS depth curve synchronous adjustment as an example)

[0216] The core of dynamic scaling is to generate adjustment ratios for each layer of the correlation curve in real time based on modified information, matching results, and dynamic model parameters, ensuring that the calibrated ratios are consistent with the current curve state. The specific process and formulas are as follows:

[0217] I. Parameter Preparation Before Calculation

[0218] 1. Basic parameter extraction

[0219] Obtain the following key parameters from the system:

[0220] Target curve information:

[0221] The target curve's i-th layer's original value before modification: V base old (i);

[0222] Modified value of the i-th layer of the target curve: V base new (i);

[0223] Modification time: t modify (Current timestamp);

[0224] Related curve information:

[0225] The initial scale (extracted from the scale library) of the correlation curve j and the baseline curve (i.e., the target curve selected as the baseline curve): R _init(i,j) (The i-th layer, the j-th correlation curve);

[0226] The last adjustment time of the correlation curve j: t last(j) ;

[0227] Dynamic model parameters:

[0228] Time decay factor coefficient: k (preset according to equipment characteristics, such as 0.004 / h);

[0229] Stratification weight coefficient: W(i) (weight of the i-th stratum, fitted based on historical data, ranging from 0 to 1);

[0230] II. Dynamic Proportion Calculation Process

[0231] Step 1: Calculate the time decay factor (to correct for device drift)

[0232] Calculate the time decay factor R based on the time difference between the baseline curve and the associated curve. t (i,j):

[0233]

[0234] Meaning: If the correlation curve has not been adjusted for a long time (large time difference), R t As the value decreases, the proportional correction range increases, offsetting the impact of equipment drift.

[0235] Unit: t modify -t last(j) The unit is "hour" (matching the unit of k).

[0236] Example as follows:

[0237] Last adjustment time for correlation curve j: t last(j) =2024-07-20 10:00;

[0238] Current modification time: t modify =2024-07-20 14:00 (Time difference = 4 hours);

[0239] k = 0.004 / h, then R t (i,j)=e^{-0.004×4}≈e^{-0.016}≈0.984;

[0240] Step 2: Determine the stratification weight coefficients (highlighting key strata);

[0241] The stratification weight W(i) is preset based on the stratification importance (which can be edited in the proportion library):

[0242] Key layer (such as the intermediate doping layer of a semiconductor chip): W(i) = 0.9ˉ1.0;

[0243] Secondary layers (such as the surface oxide layer): W(i) = 0.3ˉ0.5;

[0244] Example as follows:

[0245] Layer 1 (surface layer): W(1) = 0.4;

[0246] Layer 2 (intermediate layer): W(2) = 1.0 (critical layer);

[0247] Layer 3 (bottom layer): W(3) = 0.6;

[0248] Step 3: Calculate the dynamic adjustment ratio (comprehensive correction);

[0249] The dynamic adjustment ratio R of the i-th layer of the correlation curve j dynamic Formula for (i,j):

[0250] R dynamic (i,j)=R init (i,j)×R t (i,j)×W(i)

[0251] Meaning: Initial ratio R init decay over time (R) t After correction of the layer weight (W), a dynamic ratio that fits the current state is obtained.

[0252] Example:

[0253] Initial ratio R init(2,j) = 1.2 (intermediate layer, initial ratio of correlation curve j to baseline curve)

[0254] Time decay factor R t (2,j)=0.984 (Result of Step 1)

[0255] The hierarchical weight W(2) = 1.0 (result of step 2)

[0256] Dynamic scaling: R dynamic (2,j)=1.2×0.984×1.0≈1.181

[0257] Step 4: Calculate the adjusted value of the correlation curve:

[0258] The adjusted value V of the i-th layer of the correlation curve j j_new (i):

[0259] V j_new (i)=V base_new (i)×R dynamic (i,j)

[0260] Example:

[0261] Modified value V of the intermediate layer of the baseline curve base_new (2) = 250 μm

[0262] Dynamic ratio R dynamic (2,j)=1.181 (Result of step 3)

[0263] Adjusted value of the intermediate layer of the correlation curve j: 250 × 1.181 ≈ 295.25 μm

[0264] III. Complete Calculation Process Example (Multi-layered Scenario)

[0265] Assuming the baseline curve (A) is modified in the intermediate layer (layer 2), the calculation process for the associated curve (B) is shown in the table below:

[0266]

[0267] IV. Explanation of Key Logic

[0268] Time dimension correction: via R t To compensate for drift errors caused by long-term operation of equipment, the longer the time, the greater the correction.

[0269] Spatial dimension optimization: W(i) is used to ensure that the proportion accuracy of key layers is prioritized, so as to avoid the impact of secondary layer errors on the overall results;

[0270] Real-time performance guarantee: all parameters (such as t) modify V base_newAll data are extracted in real time, and the calculation process is completed within 100ms, meeting the requirement for real-time synchronous adjustment.

[0271] Visual matching interface: Displays the matching results of the target curve with other curves and the dynamic ratio calculation process in an interactive table and graphical manner, annotates the matching degree and ratio confidence, and supports operators to manually confirm or adjust the matching relationship and ratio parameters.

[0272]

[0273]

[0274] The curve synchronization adjustment method provided in this application embodiment, after determining the adjustment parameters in step S500, further includes an anomaly detection step, specifically including:

[0275] Dual threshold detection: Set hard and soft thresholds for anomaly detection; the hard threshold is the limit range based on the physical characteristics of the curve. When the modified value of the adjustment parameter or the calculated adjustment value exceeds the hard threshold, a warning is triggered and the adjustment is rejected; the soft threshold calculates the normal fluctuation range using the 3σ principle. When the adjustment value deviation exceeds 2σ, a backup ratio value is used for adjustment. The backup ratio is the average of the last 3 effective ratios.

[0276] Dual threshold detection formula:

[0277] Hard threshold formula (taking depth value as an example): Hard threshold range = [0, device maximum range] (if the range is exceeded, it is judged as abnormal).

[0278] The core formula for soft threshold:

[0279] Soft threshold range = [μ-2σ, μ+2σ]

[0280] (n=30, that is, take the most recent 30 valid data as samples).

[0281] Reserve ratio calculation:

[0282] (R1, R2, and R3 are the percentage values ​​of the three most recent effective adjustments).

[0283] Breakpoint recovery: Blockchain technology is used to store key ratio data and adjustment records. When the system experiences an abnormal interruption, it can be traced back to the most recent valid state based on the stored data, avoiding data loss and error accumulation.

[0284] Abnormal types and judgment criteria

[0285]

[0286]

[0287] For state backtracking, please refer to the appendix. Figure 4 .

[0288] Anomaly location and alerts: When an anomaly is detected, the source of the anomaly is located through log analysis, such as input errors or proportional model failures, and the operator is alerted in an intuitive way to assist in problem resolution.

[0289] The curve synchronization method provided in this application employs dual threshold detection and blockchain-based breakpoint recovery to effectively identify and handle abnormal situations, improve the robustness and reliability of the system, and reduce curve adjustment errors caused by anomalies.

[0290] In the curve synchronization adjustment method provided in the embodiments of this application, such as Figure 5 As shown, step S500 is illustrated below:

[0291] Adjust parameters to generate:

[0292] The total depth of the target curve (reference curve H) was changed from 4.78722 μm to 6 μm;

[0293] The velocity of H in the first layer became 147.89, and in the second layer it became 71.44.

[0294] There are four correlation curves (Al, C, O, Si), and their proportional values ​​after dynamic calibration are shown in the table below;

[0295] The adjustment parameters that need to be generated include: the depth adjustment value for each layer of each curve, and the difference (Δ) before and after adjustment.

[0296] Dynamically calibrated scaling matrix (incorporating time decay and hierarchical weights)

[0297]

[0298]

[0299] Step S600 involves batch adjusting the curves, specifically by using multi-threading to achieve parallel algorithm execution. Example:

[0300] Assume the curves to be adjusted are A, B, C, D, and E (all passed parameter validation), and the server has a 4-core CPU (thread pool initialized with 6 threads):

[0301] Task breakdown: Divided into 3 task groups (Group 1: A, B; Group 2: C; Group 3: D, E);

[0302] Thread allocation: Thread 1 processes group 1, Thread 2 processes group 2, Thread 3 processes group 3;

[0303] Parallel execution:

[0304] Thread 1: First adjust A (layers 1 through 3 are all successful), then adjust B (layer 2 fails, triggering a rollback, and B is marked as "failed");

[0305] Thread 2: Adjust C (all layers successfully);

[0306] Thread 3: Adjustments to D (successful) and E (successful);

[0307] Status monitoring: The UI displays "Total progress: 4 / 5 (Curve B failed)" in real time;

[0308] Summary of results:

[0309] Success curves: A, C, D, E;

[0310] Failure curve: B (Reason: Layer 2 adjustment value exceeds soft threshold);

[0311] Conflict detection: The stratification ratio of A and C deviates by 1.2% (normal), with no conflict.

[0312] The curve synchronization adjustment method provided in this application combines a multi-dimensional matching mechanism and a batch execution engine to achieve accurate matching and efficient batch adjustment of the modified curve and related curves, which greatly improves operational efficiency and reduces the cost of manual intervention.

[0313] The curve synchronization adjustment method provided in this application embodiment, after batch synchronization adjustment, also includes S700 scale model iterative upgrade, specifically including:

[0314] S710. Data collection and analysis: Collect modification record data, verification result data, and anomaly handling data during the batch adjustment process; analyze the applicability and accuracy of the proportional model through machine learning models; and identify the deviation patterns of the proportional model in different scenarios.

[0315] S720. Model Optimization and Upgrade: Based on the analysis results, optimize the proportional model, including adjusting the time decay factor coefficient, updating the stratification weight coefficient, and adding at least one of the following: proportional calculation rules for special scenarios. For scenarios where low-confidence proportions frequently occur, it is recommended to create a new proportional model to achieve dynamic iterative upgrades of the proportional model and continuously improve the performance of curve synchronization adjustment.

[0316] The curve synchronization adjustment method of this application, which is further optimized, utilizes hierarchical intelligent adaptation and model iterative upgrade technology to enable the system to adapt to different curve characteristic scenarios, such as continuous hierarchical and discrete hierarchical models, and its performance continuously optimizes with the increase of usage time. Through automatic verification, interactive optimization, and machine learning-driven model iterative upgrades, the system achieves precise control over the adjustment results and continuous optimization of the proportional model, thereby enhancing its adaptability to different scenarios.

[0317] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in the present invention, and these should all be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for synchronous adjustment of curves, characterized in that, Includes the following steps, S100. Establish the scale benchmark. Take a set of curves that have been aligned in layers as the benchmark group, identify and extract the initial scale parameters of different layers between the curves in the benchmark group, and form the initial scale matrix. S200, Target Curve Modification Detection: Real-time detection of whether any curves have been modified. When a parameter modification of a certain curve is detected, the modified curve is taken as the target curve, and the information of the target curve is recorded. S300, Modification Information Extraction: Analyze the modified curve data of the target curve, and extract at least the hierarchical index information, modified value size information, and modification range information involved in the modification. S400. Determine the correlation curve to be modified. Based on the information of the target curve and the scale benchmark, determine the correlation curve to be modified through dimensional matching. S500. Determine the adjustment parameters. Based on the initial ratio matrix, obtain the adjustment ratio value of each related curve to be modified and the modification information of the target curve, and generate the adjustment parameters for each related curve to be modified. S600, Batch Adjust Curves: Start the batch execution engine and synchronously adjust all related curves to be modified according to the generated adjustment parameters; After step S100 forms the initial scaling matrix, the following steps are also included: S110. Dynamic scaling model construction: Introducing time decay factor and hierarchical weight coefficient to construct dynamic scaling model; The time decay factor is calculated using the formula Calculations are performed where R0 is the initial ratio, k is the attenuation coefficient, and t is the data acquisition interval; the hierarchical weighting coefficient W... i The dynamic ratio is obtained by fitting historical data and used to distinguish the importance of different strata. The formula for calculating the dynamic ratio is as follows: ,in This is the correction value for the i-th layer of the j-th curve. The parameter values ​​are modified for the i-th layer of the target curve. This is the initial ratio; S120, Scale Library Management: Establish a scale library system to manage dynamic scale models.

2. The curve synchronization adjustment method as described in claim 1, characterized in that, Step S400 specifically involves: based on the dynamic scale model features in the scale library system, performing curve type consistency verification, hierarchical structure similarity comparison, and historical adjustment record correlation analysis on the target curve and other related curves.

3. The curve synchronization adjustment method as described in claim 2, characterized in that, In step S500, the adjustment ratio value is obtained through dynamic calculation. Specifically, the dynamic calculation involves calling the dynamic ratio model based on the extracted modification information and matching results, and combining the time decay factor and the layer weight coefficient to calculate the adjustment ratio value of each layer of the correlation curve to be modified in real time.

4. The curve synchronization adjustment method as described in claim 1, characterized in that, After determining the adjustment parameters in step S500, an anomaly detection step is also included, which specifically includes: Dual threshold detection: Set hard and soft thresholds for anomaly detection; the hard threshold is the limit range based on the physical characteristics of the curve. When the modified value of the adjustment parameter or the calculated adjustment value exceeds the hard threshold, a warning is triggered and the adjustment is rejected; the soft threshold calculates the normal fluctuation range based on the 3σ principle. When the adjustment value deviation exceeds 2σ, a backup proportional value is used for adjustment.

5. The curve synchronization adjustment method as described in claim 4, characterized in that, Following the anomaly detection step, the following steps are also included: Breakpoint recovery: Blockchain technology is used to store key ratio data and adjustment records. When the system experiences an abnormal interruption, it can be traced back to the most recent valid state based on the stored data. Anomaly location and alerts: When an anomaly is detected, the source of the anomaly is located through log analysis and an alert is issued.

6. The curve synchronization adjustment method as described in claim 5, characterized in that, Following step S600, the process also includes step S700, which involves iterative upgrading of the scaling model. Specifically, this includes: S710. Data collection and analysis: Collect modification record data, verification result data, and anomaly handling data during the batch adjustment process; analyze the applicability and accuracy of the proportional model through machine learning models; and identify the deviation patterns of the proportional model in different scenarios. S720. Model Optimization and Upgrade: Optimize the proportional model based on the analysis results, including adjusting the time decay factor coefficient, updating the hierarchical weight coefficient, and adding at least one of the following: proportional calculation rules for special scenarios.

7. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer-executable program, and the processor calls the computer-executable program in the memory to implement the curve synchronization adjustment method as described in any one of claims 1 to 6.

8. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, on which a computer-executable program is stored, and when the computer-executable program is executed by a processor, it implements the curve synchronization adjustment method as described in any one of claims 1 to 6.

9. A computer program product for achieving synchronous adjustment of curves, characterized in that, Includes code that implements the curve synchronization adjustment method according to any one of claims 1 to 6.

Images