Method for detecting cup separation failure of outdoor SMD (Surface Mount Device) packaged lamp bead

Abstract

The invention, which relates to the LED packaging field, discloses a detection method for cup-off failure of an outdoor SMD packaging lamp bead, comprising the following steps: S1, carrying out a thrust test on an SMD packaging structure to obtain thrust test data, and carrying out an environment simulation test to obtain an environment simulation test result; s2, cleaning the data, removing abnormal values and repeated data, and performing normalization processing on the cleaned data to obtain normalized feature data; s3, extracting key features according to the preprocessed feature data; s4, a weight is distributed for each key feature, the sum of the weights is 1, and a comprehensive score is calculated and used for evaluating the cup-off failure risk of the lamp beads; according to the method, a thrust test and an environment simulation test are combined, multi-dimensional key features such as a thrust peak value, a stripping speed mean value, a performance reduction rate under an environment simulation condition and a microstructure feature index are extracted, and a comprehensive evaluation system is constructed.

Description

Technical Field

[0001] The present invention relates to LED packaging technology, and particularly to a detection method for the cup detachment failure of outdoor SMD packaged lamp beads. Background Art

[0002] As one of the most mature LED packaging technologies, SMD packaging is widely used in the LED packaging industry. The bracket / chip / underfill / encapsulant, as important raw material components, affect the optical performance of LED products from multiple angles and to varying degrees. Among them, the performance of the encapsulant is affected by various factors, such as the PPA material, the ratio of epoxy resin glue to curing agent, and the selection and proportion of the filler powder, which will directly change the optical and thermal properties of the encapsulant, and thus affect the quality of LED products.

[0003] Currently, in the context of the continuous pursuit of product quality improvement, the comprehensive selection of raw materials has become particularly crucial. Nevertheless, an effective detection method for the bonding strength between the encapsulant and the bracket is still lacking, resulting in the frequent occurrence of cup detachment failure of outdoor LED lamp beads under harsh environments such as high temperature, high humidity, high salt spray, and strong light irradiation. This problem not only affects the long-term stability and durability of the product but also increases the maintenance cost of the product.

[0004] At present, the industry urgently needs a detection method for the cup detachment failure of outdoor SMD packaged lamp beads, which should be able to provide data support for the selection of various packaging schemes during the R & D stage of LED lamp beads, and then ensure the performance stability of the finally promoted products through a series of airtightness, salt spray test, cold punching test, and breathing test (energized) reliability verification. Summary of the Invention

[0005] The purpose of the present invention is to provide a detection method for the cup detachment failure of outdoor SMD packaged lamp beads, so as to solve the problem in the prior art that there is still a lack of an effective detection method for the bonding strength between the encapsulant and the bracket, resulting in the frequent occurrence of cup detachment failure of outdoor LED lamp beads under harsh environments such as high temperature, high humidity, high salt spray, and strong light irradiation.

[0006] To achieve the above purpose, the present invention provides the following technical solution: A detection method for the cup detachment failure of outdoor SMD packaged lamp beads, comprising the following steps:

[0007] S1. Perform a thrust test on the SMD packaging structure, obtain thrust test data, and perform an environmental simulation test to obtain environmental simulation test results, including performance data under high temperature, high humidity, high salt spray, and strong light irradiation;

[0008] S2. Clean the data, remove outliers and duplicate data, and perform normalization processing on the cleaned data to obtain normalized characteristic data;

[0009] S3. Extract key features based on the preprocessed feature data;

[0010] S4. Assign weights to each key feature, with the sum of the weights being 1, and calculate the comprehensive score for evaluating the cup-off failure risk of the lamp beads;

[0011] S5. Optimize the encapsulation scheme and material selection through an improved genetic algorithm;

[0012] S6. Confirm the effectiveness of the selected encapsulation scheme and materials through reliability verification of the selected encapsulation scheme and materials.

[0013] Further, the thrust test adopts a mechanical peeling method to evaluate the bonding strength between the encapsulating glue and the PPA shell through a thrust tester; record the real-time thrust curve during the test, and extract the thrust peak and the average peeling speed as key features.

[0014] Further, the data cleaning step uses statistical methods to identify and remove outliers, and the cleaned data is standardized through the min-max normalization method to ensure subsequent calculations are carried out on the same scale for each feature.

[0015] Further, the key features include the thrust peak, the average peeling speed, the performance degradation rate under simulated environmental conditions, and the microtopography feature index.

[0016] Further, the weights are assigned using the expert scoring method or the entropy weight method.

[0017] Further, the improved genetic algorithm uses a real-number coding method to represent different combinations of encapsulation schemes and materials. The selection operation in the genetic algorithm adopts a selection strategy based on ranking, the crossover operation adopts an adaptive crossover probability, and the mutation operation adopts a dynamic mutation probability; the individual with the highest fitness function value is output as the optimal encapsulation scheme and material selection; among them, the fitness function is calculated by the following formula:

[0018] F = k·ln(S + 1)+(1 - k)·S;

[0019] Where F is the fitness value, S is the comprehensive score, and k is a non-linear adjustment coefficient, 0 < k < 1.

[0020] Further, the calculation of the comprehensive score introduces a dynamic adjustment mechanism and a non-linear weighting factor, and the specific calculation formula is:

[0021]

[0022] Among them, S represents the comprehensive score, ω iis the weight of the i-th key feature, χ i is the value of the i-th key feature after normalization, α i is the adjustment coefficient of the i-th feature, n is the total number of key features, f i (χ i ) is a non-linear function:

[0023]

[0024] where μ is the average value of n features;

[0025] α i is a dynamic adjustment coefficient, which is adjusted according to real-time environmental conditions:

[0026] Temperature-related adjustment:

[0027] Humidity-related adjustment: α i = 1 + m(RH - RH 0 );

[0028] where k and m are adjustment coefficients, T 0 and RH 0 are the reference temperature and humidity.

[0029] Furthermore, the improved genetic algorithm introduces a reinforcement learning mechanism to dynamically adjust the weights of key features, and continuously optimizes the weight allocation strategy through the reinforcement learning algorithm; the specific update formula is:

[0030] w i (t+1) = w i (t) + η·δ i ;

[0031] where w i (t) is the weight of the i-th feature at the t-th step, η is the learning rate, and δ i is the reward signal based on the detection result.

[0032] Furthermore, the reliability verification includes airtightness test, salt spray test, cold punching test and breathing test.

[0033] Furthermore, it also includes a verification step: verifying the accuracy and reliability of the method by comparing with experimental data, and using statistical analysis methods to evaluate the consistency between the detection results and the actual experimental results during the verification process.

[0034] Compared with the prior art, a detection method for the de-cup failure of outdoor SMD packaged lamp beads provided by the present invention combines a thrust test and an environmental simulation test to extract multi-dimensional key features such as the peak thrust, average peeling speed, performance degradation rate under environmental simulation conditions, and microscopic morphology characteristic index, and constructs a comprehensive evaluation system;

[0035] By introducing a dynamic adjustment coefficient, the weights of the key features can be dynamically adjusted according to real-time environmental conditions, significantly improving the accuracy and adaptability of the evaluation results;

[0036] By adopting an improved genetic algorithm combined with a reinforcement learning mechanism, the weight allocation strategy of the key features is dynamically optimized, significantly improving the optimization efficiency of the packaging scheme and material selection;

[0037] By enhancing the importance of the key features through a non-linear function, the contribution of the high-score features to the comprehensive score is made greater, further improving the sensitivity and accuracy of the evaluation results;

[0038] By dynamically adjusting the weights of the key features through a reinforcement learning mechanism, the problem of fixed weights in traditional methods is avoided, significantly improving the intelligence level of the detection method. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings.

[0040] Figure 1 It is a flowchart provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the present invention will be further described in detail below with reference to the drawings.

[0042] Please refer to Figure 1 , a detection method for the de-cup failure of outdoor SMD packaged lamp beads, comprising the following steps:

[0043] S1. Perform a thrust test on the SMD package structure to obtain thrust test data, providing a direct basis for mechanical failure; and perform an environmental simulation test to obtain environmental simulation test results, including performance data under high temperature, high humidity, high salt fog, and strong light irradiation, simulating a real outdoor scenario, avoiding the limitations of single-environment testing, providing basic data for subsequent analysis, and reflecting the performance of the lamp beads under different environmental conditions;

[0044] S2. Clean the data, remove outliers and duplicate data, and normalize the cleaned data to obtain normalized feature data, ensuring data consistency and accuracy and avoiding the impact of dimensional differences on the analysis results;

[0045] S3. Extract key features based on the preprocessed feature data, identify the main factors affecting the cup detachment failure, and provide a core basis for subsequent evaluation;

[0046] S4. Assign weights to each key feature, with the sum of weights being 1, and calculate the comprehensive score for evaluating the cup detachment failure risk of the lamp beads;

[0047] S5. Optimize the encapsulation scheme and material selection through an improved genetic algorithm to find the optimal encapsulation scheme and material combination, improving the efficiency and accuracy of the detection method;

[0048] S6. Verify the reliability of the selected encapsulation scheme and materials to confirm the effectiveness of the selected encapsulation scheme and materials, ensuring the stability and reliability of the optimized scheme in practical applications.

[0049] In one embodiment of the present invention, the thrust test adopts a mechanical peeling method to evaluate the bonding force between the encapsulation glue and the PPA housing through a thrust tester; during the test process, the real-time thrust curve is recorded, and the thrust peak value and the average peeling speed are extracted as key features.

[0050] In one embodiment of the present invention, the thrust test and data acquisition can be obtained through the following steps:

[0051] S1.1: Apply an axial thrust to the SMD encapsulation structure and record the thrust value F(t), where t is the time variable;

[0052] S1.2: Collect the thrust-displacement curve data, extract the thrust peak value F_max and the peeling speed v(t); S1.3: Synchronously collect the video data of the cup detachment process through a high-speed imaging device.

[0053] In one embodiment of the present invention, the environmental simulation and performance monitoring can be obtained through the following steps:

[0054] S1.3: Conduct an accelerated life test under the conditions of high temperature (85°C ± 2°C), high humidity (95% RH ± 2%), high salt spray (NaCl concentration 5% ± 0.1%), and strong light irradiation (1000 lux ± 50 lux);

[0055] S1.4: Real-time monitor and record the following parameters:

[0056] Light decay parameter L(t): The ratio of the initial brightness L0 to the current brightness Lt;

[0057] Electrical parameter V(t): The variation of the working voltage over time;

[0058] Temperature rise parameter ΔT(t): The temperature rise on the surface of the package;

[0059] S1.5: Obtain the three-dimensional morphological evolution data of the packaging structure using micro-CT technology.

[0060] In one embodiment of the present invention, the data cleaning step uses statistical methods to identify and remove outliers. The cleaned data is standardized by the min-max normalization method to ensure subsequent calculations are performed on the same scale for each feature;

[0061] Specifically, identify outliers through statistical methods (such as the 3σ principle) to avoid interference from test errors; map data with different dimensions to [0, 1] through the min-max method to eliminate differences in parameter units.

[0062] Screen core features sensitive to cup removal failure from the massive data to reduce the model complexity. Determine key parameters through physical mechanism analysis and data-driven methods. In one embodiment of the present invention, the key features include the peak thrust, average peeling speed, performance degradation rate under environmental simulation conditions, and micro-morphology feature index;

[0063] Specifically, core indicators such as the peak thrust, average peeling speed, and performance degradation rate can be screened through correlation analysis (such as the Pearson coefficient), and a failure mode map can be established by combining micro-morphology features (such as the colloid crack index). Extract features with high variance contribution rate through principal component analysis (PCA), such as retaining the principal components with a cumulative contribution rate > 85%, and preferentially retain parameters directly related to cup removal failure (such as colloid binding force, corrosion resistance).

[0064] In one embodiment of the present invention, the assignment of weights uses the expert scoring method or the entropy weight method.

[0065] In one embodiment of the present invention, the improved genetic algorithm uses real number coding to represent different packaging schemes and material combinations; adopts a ranking-based selection strategy to avoid the "premature" phenomenon that may occur in the traditional roulette wheel selection method; the crossover operation uses an adaptive crossover probability and dynamically adjusts the crossover probability according to the population diversity; the mutation operation uses a dynamic mutation probability;

[0066] Output the individual with the highest fitness function value as the optimal packaging scheme and material selection, where the fitness function is calculated by the following formula:

[0067] F = k·ln(S + 1)+(1 - k)·S;

[0068] Where F is the fitness value, S is the comprehensive score, k is the non-linear adjustment coefficient, and 0 < k < 1;

[0069] Specifically, in the prior art, a simple linear fitness function is usually used, which cannot effectively distinguish high-score individuals. To balance the linear and non-linear relationships and improve the flexibility of fitness evaluation, the comprehensive score S is converted into the fitness value F, which is used to select the optimal encapsulation scheme and material combination in the improved genetic algorithm. The non-linear adjustment coefficient k is introduced to balance the linear and non-linear relationships, avoid over-reliance on a single index, and improve the search efficiency and robustness of the genetic algorithm;

[0070] Among them, the linear part (1 - k)·S retains the basic trend of the comprehensive score, and the non-linear part k·ln(S + 1) enhances the sensitivity to high scores, making it easier to select high-score individuals. The combination of the two takes into account both the absolute value of the score and introduces a non-linear incentive mechanism. Assuming that the comprehensive score is S = 0.8 and the non-linear adjustment coefficient k = 0.5, F = 0.5·ln(0.8 + 1) + (1 - 0.5)·0.8, F ≈ 0.7466. Through the calculation of the above formula, the fitness value not only reflects the level of the comprehensive score but also introduces a non-linear incentive mechanism, improving the flexibility of fitness evaluation, making the optimization process more in line with actual needs, and solving the problem that the fitness function in the traditional genetic algorithm is too simple to effectively distinguish the advantages and disadvantages of individuals.

[0071] In an embodiment of the present invention, the improved genetic algorithm introduces a reinforcement learning mechanism to dynamically adjust the weights of key features, and continuously optimizes the weight allocation strategy through the reinforcement learning algorithm; the specific update formula is:

[0072] w i (t+1) = w i (t) + η·δ i ;

[0073] Where, w i (t) is the weight of the i-th feature at the t-th step, η is the learning rate, and δ i is the reward signal based on the detection result;

[0074] Specifically, in the traditional genetic algorithm, the weights of key features (such as peak thrust, salt spray corrosion rate) are usually fixed (such as preset by expert experience or entropy weight method). However, the outdoor environment (temperature, humidity, salt spray concentration) is highly dynamic, and fixed weights cannot meet the requirements of multiple working conditions;

[0075] Therefore, the improved genetic algorithm generates a reward signal based on test results (such as the passing rate of salt spray test and the stability of cold stamping test), forming a closed loop of "algorithm-generated solution → experimental verification → weight adjustment → re-optimization", promoting the evolution of the algorithm towards high reliability. For example, if the lamp beads in a certain batch fail due to colloid delamination, the reward signal will reduce the weight of "average peeling speed" and increase the weight of "microscopic morphology feature index", guiding the algorithm to select materials with stronger interfacial bonding force.

[0076] The traditional genetic algorithm is prone to falling into local optima (such as over-relying on a certain material combination). Reinforcement learning dynamically adjusts the search direction through the reward signal, and the reward signal δ i reflects the quality of the detection result. If the detection result is highly consistent with the actual experimental result, a positive reward is given; if there is a large deviation between the detection result and the actual experimental result, a negative reward is given. For example, assume that in a certain detection, the probability of the lamp beads failing to pass the cup-off test predicted by the detection method is 80%, and the actual experimental result shows that the lamp beads indeed failed to pass the cup-off test. At this time, the system will give a relatively high positive reward (such as +0.8) to encourage the algorithm to continue to optimize the weight allocation of this feature. On the contrary, if the predicted probability by the detection method is 80%, but the failure does not actually occur, a negative reward (such as -0.8) is given to prompt the algorithm to reduce its attention to this feature. In this way, the reward signal can not only reflect the accuracy of the detection result but also guide the algorithm to dynamically adjust the weight allocation strategy under different environmental conditions, thereby improving the overall performance and adaptability of the detection method.

[0077] The learning rate η in the above algorithm controls the speed of weight update. The weight update formula realizes online optimization, solves the problem that the weights are fixed in the traditional evaluation method and cannot be optimized according to the detection results, and improves the accuracy and adaptability of the method.

[0078] In an embodiment of the present invention, a dynamic adjustment mechanism and a non-linear weighting factor are introduced into the calculation of the comprehensive score. The specific calculation formula is:

[0079]

[0080] where S represents the comprehensive score, ω i is the weight of the i-th key feature, χ i is the value of the i-th key feature after normalization, α i is the adjustment coefficient of the i-th feature (used for fine-tuning according to the importance or particularity of the feature), n is the total number of key features, and f i (χ i ) is a non-linear function:

[0081]

[0082] where μ is the average value of n features;

[0083] α i is a dynamic adjustment coefficient, which is adjusted according to real-time environmental conditions:

[0084] Temperature-related adjustment:

[0085] Humidity-related adjustment: α i = 1 + m(RH - RH 0 );

[0086] where k and m are adjustment coefficients, T 0 and RH 0 are the reference temperature and humidity.

[0087] Specifically, since the traditional comprehensive score calculation often uses linear weighting and cannot reflect the non-linear relationship between features, therefore, by introducing the non-linear function f i (χ i ) and the dynamic adjustment coefficient α i , the dynamic adjustment coefficient α i is adjusted according to real-time environmental conditions (such as temperature, humidity), ω i reflects the importance of each key feature, the non-linear function f i (χ i ) enhances the importance of key features, the dynamic adjustment coefficient α i is adjusted according to changes in environmental conditions, making the evaluation results closer to the actual environment, the evaluation results more dynamic and targeted, solving the problem of fixed weights in traditional evaluation methods and inability to adapt to environmental changes, and improving the sensitivity and accuracy of evaluation results to changes in environmental conditions;

[0088] In the prior art, fixed weight allocation is usually used and cannot be adjusted according to environmental conditions. By introducing the dynamic adjustment coefficient α i , the weights are adjusted according to real-time temperature and humidity, making the scoring system automatically optimized with environmental changes, making the weight allocation more flexible and intelligent, solving the problem of fixed weights in traditional evaluation methods and inability to adapt to environmental changes, and improving the sensitivity and accuracy of evaluation results to changes in environmental conditions.

[0089] In order to verify the effectiveness of the optimized encapsulation scheme and materials, in an embodiment of the present invention, the selected encapsulation scheme and materials are subjected to the following reliability verification:

[0090] Air tightness test: Evaluate the sealing performance of the encapsulation structure;

[0091] Salt spray test: Verify the corrosion resistance of the encapsulation material in a high salt spray environment;

[0092] Cold impact test: To test the stability of the encapsulation structure under sudden temperature changes;

[0093] Breathing test: To evaluate the airtightness and durability of the encapsulation structure under humidity changes;

[0094] Specifically, by performing the above-mentioned reliability verification on the selected encapsulation solutions and materials, it is ensured that the optimized encapsulation solutions and materials have good reliability and durability in practical applications.

[0095] In an embodiment of the present invention, a verification step is further included: verifying the accuracy and reliability of the method by comparing with experimental data, and using statistical analysis methods (such as regression analysis, error analysis) during the verification process to evaluate the consistency between the detection results and the actual experimental results;

[0096] Specifically, by calculating the correlation coefficient (R 2 value), the strength of the linear relationship between the detection results and the actual results is quantified; by calculating the mean absolute error (MAE) and the root mean square error (RMSE), the deviation degree between the detection results and the actual results is measured. Through these analyses, the accuracy and reliability of the detection method can be quantified, and data support can be provided for the further optimization of the method.

[0097] In summary, the present invention constructs a comprehensive and efficient outdoor SMD encapsulation lamp bead de-cup failure detection system through multi-dimensional data acquisition, feature extraction, improved genetic algorithm optimization, and dynamic adjustment mechanism. This method can not only accurately evaluate the de-cup failure risk of the lamp beads, but also optimize the encapsulation solutions and material selection, and is applicable to the quality control requirements in large-scale industrial production. Compared with the prior art, the present invention significantly improves the comprehensiveness, dynamics, and intelligence level of the detection, and provides strong technical support for the reliability design of outdoor LED encapsulation products.

[0098] Only some exemplary embodiments of the present invention have been described by way of illustration above. Undoubtedly, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above-mentioned drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for detecting failure of outdoor SMD packaged lamp beads to be out of the cup, characterized in that: The following steps are involved: S1. Perform a thrust test on the SMD packaging structure to obtain thrust test data, and perform an environmental simulation test to obtain environmental simulation test results; S2, cleaning the data, removing outliers and duplicate data, and normalizing the cleaned data to obtain normalized feature data; S3, extracting key features based on the preprocessed feature data; S4. Assign weights to each key feature, the sum of which is 1, and calculate a comprehensive score to assess the risk of lamp bead failure due to cup removal; S5. Optimizing packaging scheme and material selection through improved genetic algorithm; S6. Confirm the effectiveness of the selected packaging solutions and materials by performing reliability verification on them.

2. A method for detecting failure of outdoor SMD packaged lamp beads to be out of cup according to claim 1, characterized in that: The thrust test adopts a mechanical stripping method to evaluate the bonding strength between the encapsulation glue and the PPA shell through a thrust testing machine; during the test, a real-time thrust curve is recorded, and the thrust peak value and the mean value of the stripping speed are extracted as key features.

3. A method for detecting failure of outdoor SMD packaged lamp beads to be out of cup according to claim 1, characterized in that: The data cleaning step uses statistical methods to identify and remove outliers, and the cleaned data is standardized using the min-max normalization method to ensure that each feature is calculated at a uniform scale.

4. A method for detecting failure of outdoor SMD packaged lamp beads to be out of cup according to claim 1, characterized in that: The key features include thrust peak value, stripping velocity average value, performance degradation rate under environmental simulation conditions and micro-morphology feature index.

5. The method for detecting failure of outdoor SMD packaged lamp beads to be out of the cup according to claim 1, characterized in that: The weights are allocated using an expert scoring method or an entropy weight method.

6. A method for detecting failure of outdoor SMD packaged lamp beads to be out of the cup according to claim 1, characterized in that: The improved genetic algorithm adopts real number coding to represent different packaging solutions and material combinations. The selection operation in the genetic algorithm adopts a ranking-based selection strategy, the crossover operation adopts an adaptive crossover probability, and the mutation operation adopts a dynamic mutation probability; the individual with the highest fitness function value is output as the optimal packaging solution and material selection; wherein the fitness function is calculated by the following formula: F = k·ln(S+1)+(1-k)·S; Where F is the fitness value, S is the comprehensive score, k is the nonlinear adjustment coefficient, 0<k<1.

7. A method for detecting failure of outdoor SMD packaged lamp beads to be out of cup according to claim 1, characterized in that: The calculation of the comprehensive score introduces a dynamic adjustment mechanism and a nonlinear weighting factor. The specific calculation formula is: Among them, S represents the comprehensive score, ω i is the weight of the i-th key feature, χ i is the normalized value of the i-th key feature, α i is the adjustment coefficient of the ith feature, n is the total number of key features, f i (x i ) is a nonlinear function: Where μ is the average value of n features; α i To dynamically adjust the coefficient, adjust it according to the real-time environmental conditions: Temperature related adjustments: Humidity related adjustments: α i =1+m(RH-RH0); Among them, k and m are adjustment coefficients, T0 and RH0 are reference temperature and humidity.

8. A method for detecting failure of outdoor SMD packaged lamp beads to be out of the cup according to claim 7, characterized in that: The improved genetic algorithm introduces a reinforcement learning mechanism to dynamically adjust the weights of key features, and continuously optimizes the weight allocation strategy through the reinforcement learning algorithm; the specific update formula is: w i (t+1) =w i (t) +n·d i ; Among them, w i (t) is the weight of the i-th feature at the t-th step, η is the learning rate, δ i is the reward signal based on the detection result.

9. A method for detecting failure of outdoor SMD packaged lamp beads to be out of the cup according to claim 1, characterized in that: The reliability verification includes air tightness test, salt spray test, cold shock test and breathing test.

10. A method for detecting failure of outdoor SMD packaged lamp beads to be out of the cup according to claim 1, characterized in that: It also includes a verification step: verifying the accuracy and reliability of the method by comparing it with experimental data, and using statistical analysis methods to evaluate the consistency of the test results with the actual experimental results during the verification process.

Images