Verification system and method for material formula through confidence interval

Abstract

The invention discloses a verification system and method for a material formula through a confidence interval, and relates to the technical field of material informatics and intelligent research and development decision. Comprising a data acquisition module used for acquiring candidate formula basic data, historical experiment basic data, environment associated data, material recessive data and equipment state data and preprocessing the acquired data; according to the method, the final error value is obtained through fusion, the adjusted final credible interval is constructed, the problems that an existing material performance prediction tool can only output a point prediction result and lacks a stable credible interval, and engineers are difficult to assess that performance reaches the standard and actually and successfully grasp are solved, the coverage rate is verified through the verification set, the error scale is scaled, and the reliability of the material performance prediction tool is improved. It is ensured that the credible interval meets the preset coverage requirement, successful mastering of performance standard reaching can be quantified, an engineer does not need to depend on experience judgment any more, and the accuracy of performance evaluation is improved.

Description

Technical Field

[0001] This invention relates to the fields of materials informatics and intelligent R&D decision-making technology, specifically to a system and method for verifying material formulations using confidence intervals. Background Technology

[0002] In the field of materials research and development, equipment or tools for predicting and evaluating the performance of material formulations or processes are typically based on data-driven prediction models, such as neural networks, tree models, kernel methods, and empirical equations. These models are used to output predicted results for key properties of materials, such as strength, hardness, and elongation, to help engineers determine the feasibility of formulations or processes and to provide preliminary references for subsequent experimental directions.

[0003] Existing material performance prediction tools have significant shortcomings in practical applications: they can only output point prediction results for performance, lacking stable confidence intervals, making it difficult for engineers to assess the actual success rate of achieving performance targets; when faced with combined requirements for multiple indicators, they cannot quantify the overall success probability; if candidate formulations or processes deviate from historical experience distributions, prediction reliability drops significantly, easily misleading experimental directions; and they do not consider the reality of limited experimental resources, making it impossible to prioritize experiments based on success probability, cost, etc., while also lacking interpretability, making it difficult for engineers to pinpoint the source of uncertainty and optimization directions, resulting in high trial-and-error costs and long R&D cycles. Therefore, this paper proposes a verification system and method for material formulations using confidence intervals to address these issues. Summary of the Invention

[0004] Technical problems to be solved To address the shortcomings of existing technologies, this invention provides a system and method for verifying material formulations using confidence intervals, thus solving the problems mentioned in the background section.

[0005] Technical solution To achieve the above objectives, the present invention provides the following technical solution: a system for verifying material formulations using confidence intervals, such as... Figure 1 As shown, it includes: The data acquisition module is used to collect basic data of candidate formulations, basic data of historical experiments, environmental correlation data, implicit data of materials, and equipment status data. The basic data of candidate formulations includes component proportions and process parameters. The component proportions are calculated by weighing with a high-precision electronic balance with a weighing accuracy of 0.001 grams, calculated by dividing the mass of a single component by the total mass of the formulation. The process parameters include reaction temperature and holding time. The reaction temperature is collected by an intelligent temperature control system with an accuracy of ±0.5 degrees Celsius, and the holding time is collected by a timer with an accuracy of ±1 second. The basic data of historical experiments includes measured performance values, number of test repetitions, single measurement error, standard deviation of process parameter fluctuations, and the range of common component proportions in similar formulations. The measured performance values ​​are determined by selecting the corresponding standard instrument according to the material type. The strength of metallic materials is determined by a universal testing machine with a range of 0 to 1000 MPa and an accuracy of ±1 MPa. The glass transition temperature of plastic materials is determined by a differential scanning calorimeter with an accuracy of ±1 degree Celsius. The number of test repetitions is determined based on performance stability, usually 3-5 times. Performance that is easily affected by the environment is repeated 5 times, and performance with good stability is repeated 3 times, recorded as an integer. The single measurement error is the difference between a single measured performance value and the average value of multiple measurements, with the unit consistent with the performance index; The standard deviation of process parameter fluctuation is calculated statistically from the process parameter data of similar historical formulas. The calculation process is as follows: first, determine the number of process parameter data groups to be included in the statistics, then calculate the difference between the single process parameter value of each group and the average value of all data, square each difference separately and sum them up, divide the sum by (the number of data groups minus one), and finally take the square root of the result. The unit is consistent with the process parameter. The range of common ingredient proportions in similar formulas is calculated based on the ingredient proportion data of similar historical formulas. The calculation process is as follows: first, calculate the average and standard deviation of the ingredient proportion in similar historical formulas; subtract 1.96 from the average and multiply by the standard deviation to obtain the lower limit of the range; and add 1.96 to the average and multiply by the standard deviation to obtain the upper limit of the range. Environmental data include high and low temperature cycling degradation rate, salt spray retention rate, and temperature and humidity difference. The high and low temperature cycling degradation rate is obtained by testing in a high and low temperature cycling test chamber. The initial performance of the material is measured first, and the performance is measured after 100 cycles from -40℃ to 85℃. The result is calculated as (initial performance minus post-cycle performance) divided by the initial performance and then multiplied by 100%. Salt spray retention rate was tested in a salt spray test chamber. After spraying with a 5% sodium chloride solution for 48 hours, the result was calculated as (performance after testing divided by initial performance) multiplied by 100%. The temperature and humidity difference is obtained by continuously collecting data in the target environment for 24 hours using temperature and humidity sensors, calculating the difference between the highest and lowest temperatures, and calculating the difference between the highest and lowest humidity. The implicit data of the materials include the purity of the raw materials, the standard deviation of the particle size distribution, and the glass transition temperature. The purity of the raw materials is determined by high performance liquid chromatography. The standard deviation of particle size distribution was determined using a laser particle size analyzer; the glass transition temperature was measured using a differential scanning calorimeter under a nitrogen atmosphere at a heating rate of 10℃ / min, and the midpoint temperature of the heat flow rate abrupt change interval of the differential scanning calorimeter curve was taken with an accuracy of ±1℃. Equipment status data includes calibration out-of-tolerance values, repeated measurement standard deviation, critical component runtime, and zero-point drift. Calibration out-of-tolerance values ​​are determined by comparison with standard samples, i.e., the difference between the equipment's measurement value of the standard sample and the true value of the standard sample, with units consistent with the measurement parameters. Repeated measurement standard deviation is obtained by repeatedly measuring the same standard sample 10 times. The calculation process is as follows: first, calculate the difference between each group of single measurements and the average of the 10 measurements; square each difference separately and sum them up; divide the sum by (10 minus 1); finally, take the square root of the result. The runtime of key components is obtained by statistical analysis of the equipment control system or by accumulating the runtime of each run using a timer; the zero-point drift is calculated by recording the initial zero-point value after the equipment has been preheated for 30 minutes, recording the zero-point value again after the equipment has been running continuously for 4 hours, and calculating the difference between the two zero-point values. Calibration deviations are checked daily to ensure that the measurement accuracy of the equipment is under real-time monitoring; the standard deviation of repeated measurements is checked every 3 days to avoid short-term equipment fluctuations affecting error calculation; the running time of key components is recorded in real time, and a wear assessment is conducted every 500 hours of cumulative running time; the zero-point drift is checked before each equipment start-up, and the initial zero-point value is recorded after 30 minutes of preheating to ensure that the zero point of the equipment is free of deviation before each experiment. The collected data is then preprocessed. The specific steps for data preprocessing are as follows: First, the total number of samples and the average value of historical data for each class are counted. The difference between a single data value and the average value is calculated, squared, and summed. This sum is divided by the total number of samples minus one, and the square root of the result is taken to obtain the standard deviation. Outliers exceeding a preset multiple of this standard deviation are removed. The historical minimum and maximum values ​​of each class are determined. Using 0 as a fixed lower bound, the original data is subtracted from the historical minimum value, and then divided by the historical maximum value minus the historical minimum value, resulting in standardized data ranging from 0 to 1. The cleaned data is standardized into a unified format and mapped to directly calculable quantitative values. A unique code is assigned to each candidate formulation, and environmental, material, and equipment-related data are all associated with this code and stored together in a relational database. Finally, a structured dataset is output in CSV format, containing core fields such as formulation code, component proportion, process temperature, raw material purity, equipment calibration deviation, and environmental temperature and humidity difference.

[0006] The error calculation module is used to normalize the preprocessed environmental correlation data, material implicit data and equipment status data, and then perform error calculation and correction to obtain the final error value. Based on the final error value, the module constructs and outputs the adjusted final confidence interval. The specific steps for error calculation and correction of the preprocessed environmental correlation data, material implicit data, and equipment status data are as follows: First, the preprocessed structured dataset is divided according to a preset ratio: 6 sets for training, 2 sets for calibration, and 2 sets for validation. The dataset is then stratified according to the component percentage range (e.g., 0-5%, 5-10%) and the process temperature range (e.g., 20-30℃, 30-40℃) to ensure that the sample distribution of the training, calibration, and validation sets is completely consistent with the original structured dataset, thus avoiding the impact of data distribution deviations on subsequent calculations. Next, the calibration set was clustered using the K-means algorithm. The number of clusters K was determined by the elbow rule: the sum of squares within clusters (WCSS) was calculated as K ranged from 2 to 10. When K increased to 5, the WCSS curve showed a clear inflection point. Further increasing the value of K resulted in a decrease in WCSS of less than 10%, so the optimal number of clusters K=5 was determined. At the same time, the clustering effect was verified by the silhouette coefficient. When K=5, the silhouette coefficient was 0.72. The clustering dimensions were component proportion and process temperature: the component proportion was divided into intervals of 5 percentage points, such as 0-5% and 5-10%, and the process temperature was divided into intervals of 10 degrees Celsius, such as 20-30℃ and 30-40℃. This algorithm divided the calibration set into multiple sample groups with similar characteristics.

[0007] First, obtain the predicted and measured performance values ​​for each sample in each sample group. Subtract the measured performance value from the predicted performance value and take the absolute value to obtain the absolute value of the performance residual for a single sample. Then, statistically analyze the distribution of all absolute values ​​of performance residuals in each sample group and take the 95th percentile. Specifically, this means that 95% of the absolute values ​​of performance residuals in that sample group are less than or equal to this value, thus obtaining the basic error value.

[0008] By combining implicit data of materials and introducing parameters such as raw material purity and particle size distribution standard deviation, the implicit correction coefficient of materials is obtained through comprehensive calculation. Specifically, it is based on empirical coefficients based on historical data statistics. For example, by analyzing experimental data of similar materials in the past 3 years, it was found that: for every 1% decrease in raw material purity, the performance error increases by an average of 0.01 times; for every 1 micrometer increase in particle size distribution standard deviation, the performance error increases by an average of 0.005 times. Substituting the raw material purity and particle size distribution standard deviation into the calculation, the implicit correction coefficient of materials is obtained. The equipment status is determined based on the equipment status data, and corresponding corrections are applied to the basic errors according to the determination results. Based on environmental correlation data, parameters such as high and low temperature cycle decay rate, salt spray retention rate, and temperature and humidity difference are introduced, and environmental correction coefficients are obtained through comprehensive calculation. Specifically, the calibration deviation value is determined by comparing with standard samples, that is, the difference between the value of the standard sample measured by the equipment and the true value of the standard sample, with the unit consistent with the measurement parameter. The running time of key components is read in real time from the equipment control system.

[0009] If the calibration deviation is less than or equal to 0.5 and the operating time of the key components is less than or equal to 1000 hours, the equipment condition is judged as excellent; if the calibration deviation is greater than 0.5 and less than or equal to 1.0, or the operating time of the key components is greater than 1000 hours and less than or equal to 2000 hours, the equipment condition is judged as medium; if the calibration deviation is greater than 1.0, or the operating time of the key components is greater than 2000 hours, the equipment condition is judged as poor.

[0010] A correction factor of 1.0 is applied when the condition is good, a correction factor of 1.2 is applied when the condition is medium, and a correction factor of 1.5 is applied when the condition is poor, thus obtaining the equipment condition correction factor.

[0011] The high and low temperature cycle decay rate was obtained by testing in a high and low temperature chamber. The test conditions were 100 cycles from -40 degrees Celsius to 85 degrees Celsius, and the results were expressed as a percentage. The salt spray retention rate was obtained by testing in a salt spray chamber. The test conditions were 48 hours of spraying with a 5% sodium chloride solution, and the results were expressed as a percentage. The temperature and humidity difference was obtained by continuously collecting data from a temperature and humidity sensor for 24 hours and calculating the temperature difference corresponding to the maximum daily temperature and humidity.

[0012] Based on empirical coefficients derived from environmental impact experiments, this coefficient is obtained by analyzing material performance deviations under different environments: for every 1% decrease in high and low temperature cycling decay rate, the performance error increases by an average of 0.008 times; for every 1% decrease in salt spray retention rate, the performance error increases by an average of 0.003 times; for every 1 degree Celsius increase in temperature and humidity difference, the performance error increases by an average of 0.002 times. Substituting these three environmental parameters into the calculation, the environmental correction coefficient is obtained.

[0013] The final error value is obtained by combining the basic error with various correction coefficients. The coverage is verified using the validation set. If the preset target is not met, the final error value is scaled proportionally until the coverage meets the requirements, and the adjusted final confidence interval is output.

[0014] The final error value is obtained by multiplying the basic error value by the material implicit correction factor, then by the equipment condition correction factor, and finally by the environmental correlation correction factor.

[0015] Determine the range between the predicted performance value and the final error value. Count the number of samples whose measured performance values ​​in the validation set fall within this range. Divide this number by the total number of samples in the validation set to obtain the coverage rate. The result is expressed as a percentage. The preset coverage target is 90%.

[0016] If the calculated coverage is less than 90%, calculate the scaling factor, which is equal to 90% divided by the actual coverage. Multiply the scaling factor by the original final error value to obtain the adjusted final error value. Validate the adjusted coverage again using the validation set. If it still does not reach 90%, repeat the scaling steps until the coverage meets the preset target of 90%. Use the performance prediction value plus or minus the adjusted final error value as the final confidence interval.

[0017] The interval adjustment module is used to calculate the similarity and material influence value between candidate formulations and historical samples. The specific steps to obtain the similarity are as follows: All components of the candidate formulations were examined, and the mass of each component was weighed using a high-precision electronic balance. The ratio of the mass of a single component to the total mass of the formulation was calculated to obtain the component proportion of the candidate formulation. The process temperature of the candidate formulations was collected using an intelligent temperature control system, and the holding time was collected using a timer to obtain the process parameters of the candidate formulations. Historical sample data of similar materials were extracted from the historical experimental database to obtain the component proportion and process parameters of the historical samples. The component proportion and process parameters were normalized to ensure that they were on the same dimension. The distance evaluation value between the two was calculated based on a fixed feature dimension, and the similarity was obtained by combining the maximum historical distance. The specific steps for calculating the distance assessment value are as follows: First, calculate the difference between the proportion of corresponding components in the candidate formulation and the historical sample, square each difference and sum them; then calculate the difference between the corresponding process parameters of the two, square each difference and sum them; add the two sums together, and take the square root of the sum to obtain the distance evaluation value. The distance assessment value is obtained as follows: In the formula, This represents the distance assessment value. The larger the value, the greater the overall difference between the candidate formulation and historical samples in terms of component proportions and process parameters. Indicates the number of ingredient types. Indicates the candidate formulation number The proportion of each component Indicates the first historical sample The proportion of each component This indicates the number of process parameter types, i.e., the number of process parameters involved in the candidate formulation, and determines the number of calculation items for each process parameter dimension. Indicates the candidate formulation number One process parameter, Indicates the first historical sample Each process parameter.

[0018] The specific steps to obtain the similarity are as follows: Distance assessment values ​​between all historical samples are extracted from the historical experimental database. The maximum value among these distance assessment values ​​is taken as the maximum historical distance. The similarity is obtained by combining the distance assessment values ​​between the candidate recipe and the historical samples with the maximum historical distance.

[0019] The similarity is obtained in the following ways: In the formula, The similarity score ranges from 0 to 1. The larger the value, the higher the similarity between the candidate formulation and historical samples, and the stronger the reliability of subsequent evaluation and adjustment operations based on historical samples. Indicates the distance assessment value. This represents the maximum historical distance, i.e., the maximum value of the distance assessment between all historical samples, expressed in units of 1 and 2. Consistent.

[0020] The specific steps to obtain the material influence value are as follows: The material is placed in a high-low temperature chamber, and the cycling conditions are set from -40°C to 85°C, with 100 consecutive cycles. Key material properties, such as strength, are tested before and after the cycles. The ratio of the post-cycle performance value to the pre-cycle performance value is calculated, and then multiplied by 100% to obtain the high-low temperature cycle decay rate. The thermal behavior curve of the material is measured using a differential scanning calorimeter, and the glass transition temperature is determined based on the inflection point of the curve. The maximum target environmental temperature is determined according to the actual application scenario of the material; for example, 85°C for automotive materials and 40°C for interior decoration materials. Calculate the difference between the glass transition temperature and the highest temperature of the target environment, and take the absolute value of the difference to obtain the temperature difference; Multiply the temperature difference by 0.5 to obtain the influence value of the temperature difference; subtract the high and low temperature cycle attenuation rate from 100, and multiply the result by 0.3 to obtain the influence value of the attenuation rate; subtract the sum of the influence values ​​of the temperature difference and the attenuation rate from 100 to obtain the influence value of the material. Among them, 0.5 and 0.3 are empirical coefficients based on the statistical data of material environmental adaptability experiments, and are dimensionless.

[0021] The specific steps for adjusting the range based on the similarity comparison with the preset threshold and in conjunction with the material influence value are as follows: Two similarity thresholds are set: a first threshold and a second threshold. The first threshold corresponds to the 95th percentile of the historical sample similarity distribution, and the second threshold corresponds to the 99th percentile of the historical sample similarity distribution. The calculated similarity is compared with the first and second thresholds respectively. Specifically, the similarity data between all historical samples is extracted from the historical experimental database, and the distribution of this data is statistically analyzed. The 95th percentile of the distribution is taken as the first threshold, and the 99th percentile of the distribution is taken as the second threshold. The preset score is set to 60 points, the first ratio is 1.05, and the second ratio is 1.1. The adjusted final error value is obtained.

[0022] When the similarity is greater than or equal to the first threshold, it indicates that the candidate formulation is highly similar to historical samples, and the historical data provides strong support for the formulation. Therefore, there is no need to adjust the interval width, and the final credible interval remains unchanged.

[0023] When the similarity is between the first and second thresholds, it indicates that the candidate formulation has a moderate degree of similarity to historical samples, and further judgment is needed by combining the material influence value. If the material's influence value is greater than or equal to the preset score, it means that the material's overall performance is good. The interval width is then adjusted according to the first ratio to appropriately expand the interval coverage.

[0024] If the material's impact value is less than the preset score, it indicates that the material's overall performance is relatively weak. The interval width will be adjusted according to the second ratio to further expand the interval coverage to address potential risks.

[0025] When the similarity is less than the second threshold, it indicates that the candidate formulation has a low degree of similarity with historical samples, and the existing historical data is insufficient to reliably evaluate the formulation. A rejection conclusion is output, and a suggestion is made to supplement the historical data of the characteristic region of the formulation, such as the proportion range of specific components, experimental data under specific process parameter conditions, etc.

[0026] If the condition remains unchanged, the initial confidence interval is directly used as the target confidence interval.

[0027] For cases where the width needs to be adjusted, the adjusted interval width is calculated according to the corresponding ratio to obtain the target confidence interval.

[0028] In the case of rejection, the confidence interval is not output; only the rejection conclusion and data supplementation prompts are output.

[0029] Finally, the determined target confidence interval or related conclusions will be output.

[0030] The stability assessment module is used to compare the basic data of candidate formulations with historical experimental data in real time to obtain the stability judgment result. The specific steps to obtain the stability determination result are as follows: The composition information of all components is extracted from the candidate formulation's basic data, including the percentage of each component in the formulation and its corresponding type. Formulation data of the same type as the candidate formulations are selected from the historical experimental database, and the percentage data of each corresponding component in these historical formulations is obtained. Statistical analysis is used to determine the range of common component percentages for similar formulations. Simultaneously, key process parameters, such as reaction temperature, holding time, and pressure, are extracted from the candidate formulation's basic data; parameter data for similar processes are extracted from the historical experimental data for subsequent calculation of the standard deviation of process parameter fluctuations. The proportion of each ingredient in the candidate formulation was compared one by one with the typical proportion range of the corresponding ingredient in similar formulations from historical experimental data. If the proportions of all ingredients in a candidate formula fall within the corresponding normal range, it indicates that the ingredient ratios of the formula conform to the common ranges of similar historical formulas, and the proportions are deemed reasonable.

[0031] If the proportion of at least one component in a candidate formulation exceeds the corresponding normal range, it indicates that the composition ratio of the formulation is significantly different from that of similar historical formulations, and is therefore deemed to have insufficient reasonableness in terms of component proportions.

[0032] Based on historical experimental data, data on similar process parameters in the same dimension as the candidate formulation process parameters are extracted. For example, if the process parameter of the candidate formulation is the reaction temperature, only the reaction temperature data of all similar formulations in the historical data are extracted. The specific requirements are: the number of data groups involved in the statistics is ≥300, the law of large numbers is satisfied, and the standard deviation calculation results are ensured to be reliable. The standard deviation of the historical process parameter fluctuation under this dimension is obtained through statistical calculation.

[0033] Subsequently, the target process parameter value is extracted from the candidate formulation basic data. The difference between this parameter value and the average value of the same historical process parameters is calculated, and the absolute value of the calculated difference is the process parameter fluctuation value.

[0034] Finally, using the process parameter fluctuation value as the dividend and the historical process parameter fluctuation standard deviation as the divisor, a division operation is performed. The result is then multiplied by 100% to convert it into a percentage form. The final result is the impact rate of process parameter fluctuation on performance.

[0035] If the lower threshold is set to 50%, and the impact rate of process parameter fluctuations on performance is less than or equal to 50%, it indicates that the process parameters of the candidate formulation deviate little from those of similar historical processes, and the impact on performance is within an acceptable range, thus the process is considered stable.

[0036] If the impact rate of process parameter fluctuations on performance is greater than 50%, it indicates that the process parameters of the candidate formulation deviate significantly from those of similar historical processes, which may have a significant impact on performance, and is therefore judged as an unstable process.

[0037] A comprehensive analysis was conducted, combining the results of the component rationality assessment and the process stability assessment: When the proportions are reasonable and the process is stable, it means that the composition ratio and process parameters of the formula are within a reasonable range, the overall stability is good, and it is judged to be of qualified stability.

[0038] When the proportions are reasonable but the process is unstable, or when the proportions are not reasonable but the process is stable, it indicates that there are certain problems with the formulation in terms of ingredient ratios or process parameters, but it is not completely out of control. The overall stability is at a moderate level and is judged as moderate stability.

[0039] When the proportions are not reasonable and the process is unstable, it indicates that there are obvious problems with the ingredient ratios and process parameters of the formula, resulting in poor overall stability, and the formula is judged to be unqualified for stability.

[0040] Based on the above comprehensive assessment, the final stability determination result is obtained.

[0041] Using the stability determination result as an auxiliary reference, combined with the similarity and preset threshold comparison results and the material influence value, the initial confidence interval is adjusted according to the specific steps of subsequent interval adjustment, and finally the adjusted target confidence interval is output.

[0042] The credibility assessment module is used to obtain a basic compliance assessment value based on the target credibility interval through random sampling. Specifically, the number of random samplings can be set to 300. Combined with the stability judgment result, a feasibility assessment value is obtained. The basic compliance assessment value and the feasibility assessment value are fused to obtain the final credibility assessment value. The product fusion is based on historical data verification, which can give priority to reflecting the key impact of stability on the success of the experiment and output decision suggestions.

[0043] The specific steps for obtaining the final reliable assessment value and outputting decision recommendations are as follows: Obtain parameters related to interval width, obtain the final confidence interval of the current candidate formulation performance index from the error calculation module, calculate the final confidence interval width, extract the confidence interval width of similar formulations from the historical experimental database, statistically obtain the historical maximum interval width, and then calculate the interval relative width term. The similarity between the candidate formula and historical samples is obtained from the interval adjustment module. The final error value of the current candidate formula is obtained. The final error value of the same formula is extracted from the historical experimental database. The historical maximum error is obtained statistically. Then, the measurement error ratio is calculated. The above parameters are fused and calculated according to weights to obtain the final reliable evaluation value. The final reliable evaluation value is obtained as follows: In the formula, This represents the final credible evaluation value, and is achieved through... Will The conservative score is mapped to 0-100, where 0.95 is a conservative scaling factor. Based on historical data validation, this makes the evaluation results conservative, ensuring a high success rate. When the initial evaluation value is 0.8, multiplying by 0.95 results in a conservative score of 76, corresponding to an actual experimental success rate ≥85%; when the initial evaluation value is 0.7, the conservative score is 66.5, corresponding to an actual experimental success rate ≥75%, ensuring a relatively stable evaluation result. The value ranges from 0 to 1, with higher values ​​indicating a more suitable candidate formulation. Indicates the final confidence interval width. Indicates the historical maximum interval width. This step calculates the relative width of the interval. Indicates similarity. This represents the final error value. This indicates the largest historical error. This step calculates the measurement error percentage. , as well as These are the width weight factor, similarity weight factor, and error weight factor, respectively, and the sum of the three weight factors is 1, for example, 0.4, 0.3, and 0.3. Three processes were set up for testing. The parameters selected for the candidate processes were: reaction temperature 150℃, holding time 30 minutes, pressure 2.5MPa, and heating rate 10℃ / min.

[0044] The parameters used in the comparative process 1 were: reaction temperature 180℃, holding time 20 minutes, pressure 3.0MPa, and heating rate 15℃ / min.

[0045] The parameters used in the comparative process 2 were: reaction temperature 120℃, holding time 40 minutes, pressure 2.0MPa, and heating rate 5℃ / min.

[0046] Group Interval relative width item Similarity Measurement error percentage Final credible assessment value Candidate process 0.75 0.85 0.8 0.785 Comparison Process 1 0.5 0.6 0.5 0.525 Comparison Process 2 0.6 0.7 0.65 0.63 Based on the data in the table, it can be seen that the candidate process outperforms comparison processes 1 and 2 in terms of relative interval width, similarity, and measurement error percentage. The final confidence evaluation value reaches 0.785, significantly higher than 0.525 for comparison process 1 and 0.63 for comparison process 2. This indicates that the candidate process performs better overall in terms of performance reliability, interval stability, and error controllability. Furthermore, the candidate process corresponds to Candidate A, comparison process 1 to Candidate B, and comparison process 2 to Candidate C. Evaluation thresholds are set as follows: a high threshold of 80 and above (final confidence evaluation value greater than or equal to 80); a medium threshold of 60 to below 80 (final confidence evaluation value greater than or equal to 60 and less than 80); a low threshold of 40 to below 60 (final confidence evaluation value greater than or equal to 40 and less than 60); and a low threshold of below 40 (final confidence evaluation value less than 40). The above thresholds are determined based on the correspondence between "evaluation value and actual experimental success rate" in historical experimental data. For example, when the final credible evaluation value reaches 80, the average actual experimental success rate of similar formulations in history reaches more than 90%.

[0047] Based on the magnitude of the final confidence assessment value, the corresponding experimental priority suggestions are output: If the final credibility assessment value reaches the high threshold, it indicates that the formulation has high credibility and experimental value, and priority experimental recommendations are output.

[0048] If the final credibility assessment value is in the middle threshold range, it indicates that the formulation has a certain credibility and experimental value, but its priority is lower than the former, and alternative experimental suggestions are output.

[0049] If the final credibility assessment value is in the low threshold range, it means that the credibility of the formulation is low and the experimental risk is high. An exploratory experiment suggestion is output, recommending that a small-scale, exploratory experiment be conducted.

[0050] If the final confidence assessment value is lower than the low threshold, it indicates that the confidence of the formula is extremely low and the possibility of the experiment being successful is very small. The output suggests postponing the experiment and recommending that the formula be optimized and improved first.

[0051] Simultaneously, combining the calibration deviation and zero-point drift values ​​from the equipment status data: If the calibration error exceeds the allowable range, a suggestion will be made to recalibrate the relevant equipment, and a standard sample can be used for calibration.

[0052] If the zero-point drift is too large, the system will output a suggestion to perform zero-point calibration on the device.

[0053] In addition, based on the deviations of the candidate formulation's component proportions and process parameters from historical reasonable ranges: If the proportion of a component exceeds the historical reasonable range, a suggestion will be provided to adjust it to a reasonable range.

[0054] If the process parameters deviate from the historical reasonable range, suggestions for optimizing the process parameters to bring them back to a reasonable range will be output.

[0055] Online update module: When new experimental data is generated, the measured value is compared with the predicted interval during the evaluation. If the measured value falls within the interval, the data is added to the calibration set and validation set in the historical database, and the estimate of the basic error value is updated incrementally. If the measured value falls outside the interval, it is marked as an out-of-domain sample and added to the specific out-of-domain sample set in the historical database for subsequent optimization of similarity threshold and interval expansion logic.

[0056] The validation set coverage (target ≥ 90%) and rejection rate (target ≤ 10%) are automatically calculated monthly. If the coverage is below 88% for two consecutive months, the scaling factor of the final error value is readjusted to between 1.05 and 1.1. After adjustment, the validation set is used again to ensure that the coverage returns to above 90%. If the rejection rate is above 12% for two consecutive months, the second threshold is lowered, specifically from the 99th percentile to the 98th percentile, to reduce the misclassification rate of out-of-domain samples.

[0057] After removing the range adjustment module, the fluctuation range of the final confidence evaluation value of the candidate process increased from ±5% to ±12%; after removing the stability evaluation module, the matching degree between the experimental recommendations and the actual success rate decreased from 85% to 60%, verifying the supporting role of each module in the evaluation reliability.

[0058] A method for verifying material formulations using confidence intervals, such as... Figure 2 As shown, it includes the following steps: Step 1: Collect basic data of candidate formulations, basic data of historical experiments, environmental correlation data, implicit data of materials, and equipment status data, and preprocess the collected data; Step 2: Perform error calculation and correction on the preprocessed environmental correlation data, material implicit data and equipment status data to obtain the final error value, and construct and output the adjusted final confidence interval based on the final error value; Step 3: Calculate the similarity and material influence value between the candidate formulation and historical samples. Based on the comparison results of the similarity and the preset threshold, and combined with the material influence value, adjust the range and output the target confidence range. Step 4: Compare the basic data of the candidate formulation with the basic data of historical experiments in real time to obtain the stability determination results; Step 5: Based on the target confidence interval, obtain the basic compliance assessment value through random sampling, combine it with the stability judgment result to obtain the feasibility assessment value, integrate the basic compliance assessment value and the feasibility assessment value to obtain the final confidence assessment value, and output decision recommendations.

[0059] Beneficial effects The present invention has the following beneficial effects: (1) The verification system and method for material formulation by confidence intervals, by fusing to obtain the final error value and constructing the adjusted final confidence interval, solves the problems of existing material performance prediction tools that can only output point prediction results, lack stable confidence intervals, and make it difficult for engineers to assess the actual success of performance compliance. Furthermore, by verifying the coverage rate and scaling the error scale through the validation set, the confidence interval is ensured to meet the preset coverage requirements, making the success of performance compliance quantifiable. Engineers no longer need to rely on experience to make judgments, thus improving the accuracy of performance evaluation.

[0060] (2) This verification system and method for material formulation through confidence intervals solves the problem that existing tools cannot quantify the overall success probability when facing the joint requirements of multiple indicators by obtaining the final credible evaluation value. Furthermore, it transforms the joint evaluation results of multiple indicators into an intuitive credibility score, allowing engineers to quickly grasp the overall grasp of the simultaneous achievement of multiple indicators without complicated conversion, thus reducing the difficulty of multi-indicator evaluation.

[0061] (3) This verification system and method for material formulations by confidence intervals solves the problem that existing tools have a significant drop in prediction reliability and are prone to misleading experimental directions when candidates exceed the historical experience distribution by calculating the similarity between candidate formulations and historical samples and the material influence value. Furthermore, the system ensures that candidate predictions within the historical experience range are reliable by adjusting the differential interval. At the same time, the out-of-domain rejection mechanism can avoid experimental direction deviations caused by unreliable predictions and reduce invalid experiments.

[0062] (4) This system and method for verifying material formulations through confidence intervals outputs decision recommendations based on the final confidence assessment value obtained. It solves the problem that existing tools do not take into account the reality of limited experimental resources and cannot prioritize experiments according to the success probability. Furthermore, it optimizes the feasibility assessment value by combining the stability judgment result, so that the experiment ranking takes into account both the success probability and the formulation stability. Engineers can prioritize the allocation of resources to candidates with high confidence and high stability, avoid wasting resources on low-confidence experiments, and improve the utilization rate of experimental resources.

[0063] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0064] Figure 1 This is a structural diagram of a material formulation verification system based on confidence intervals according to the present invention. Figure 2 This is a flowchart of a method for verifying material formulations using confidence intervals according to the present invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] Example 1 To verify the practical application effect and operability of the material formulation verification system and method based on confidence intervals described in this invention, this embodiment takes the development of rubber formulations for automotive seals as the application scenario, and conducts full-process verification around the core needs of quantifying the reliable performance of rubber materials and making experimental priority decisions.

[0067] Based on the mechanical performance requirements of rubber materials for automotive seals, and referring to industry standards such as GB / T528-2009 "Determination of Tensile Stress-Strain Properties of Vulcanized Rubber or Thermoplastic Rubber" and GB / T531.1-2008 "Test Method for Indentation Hardness of Vulcanized Rubber or Thermoplastic Rubber Part 1: Shore Hardness Tester Method", the material performance targets for this embodiment are set as follows: strength not less than 20 MPa, hardness within the range of 65 ± 2A, and elongation not less than 420%.

[0068] To compare the impact of different process parameters and component ratios on the reliability of rubber performance compliance, three groups of rubber formulations with differentiated characteristics were selected for evaluation (covering conventional processes, high-temperature short-shelf-life processes, and low-temperature long-shelf-life processes). The core process parameters and key component ratios of each group of formulations are as follows: The material performance targets are set as follows: strength not less than 20 MPa, hardness within the range of 65 ± 2A, and elongation not less than 420%.

[0069] Three rubber formulations were selected for evaluation. The core process parameters of each formulation are as follows: Candidate A has a reaction temperature of 150℃, a holding time of 30 minutes, a carbon black content of 32%, and a vulcanizing agent content of 1.2%; Candidate B has a reaction temperature of 160℃, a holding time of 25 minutes, a carbon black content of 35%, and a vulcanizing agent content of 1.0%; Candidate C has a reaction temperature of 180℃, a holding time of 20 minutes, a carbon black content of 40%, and a vulcanizing agent content of 0.8%.

[0070] Through a complete process of data collection, error calculation, interval adjustment, and reliability scoring, the evaluation conclusions for each group of formulations were obtained: Candidate A's performance indicators had a reliability interval of strength between 23 MPa and 27 MPa, hardness between 65A and 67A, and elongation between 425% and 440%, with a reliability score of 76. It is recommended to prioritize this experiment. Candidate B's performance indicators had a reliability interval of strength between 21 MPa and 25 MPa, hardness between 64A and 66A, and elongation between 410% and 425%, with a reliability score of 65. It is recommended to fine-tune and re-evaluate. Candidate C's performance indicators had a reliability interval of strength between 18 MPa and 22 MPa, hardness between 60A and 64A, and elongation between 390% and 410%. It had no reliability score and should be rejected. It is recommended to supplement sampling or conduct simulations first.

[0071] Regarding interpretability details, taking candidate A as an example, the sources of uncertainty are decomposed as follows: Model error accounts for 30%, because the performance prediction model simplifies the complex kinetic process of rubber crosslinking reaction, resulting in a systematic deviation of an average of 1.2 MPa between the predicted and measured strength values; Measurement error accounts for 25%, with the hardness tester calibration deviation being 0.3A, which is within the allowable range of 0.5A, but the repeatability standard deviation of elongation measurement is 5%, accounting for 25% of the total uncertainty; Out-of-domain expansion accounts for 20%, with the similarity between candidate A and historical samples being 0.85, which is lower than the interval expansion starting point (95th percentile of historical similarity distribution), requiring a moderate widening of the interval to cover potential performance fluctuations; Data sparsity accounts for 25%, with only 45 historical samples in the 1.2% vulcanizing agent interval, fewer than the recommended 50, which limits the accuracy of similarity calculation.

[0072] The three samples with the highest similarity to candidate A were extracted from the historical database, and their actual performance was examined: The first sample contained 31.8% carbon black and 1.1% vulcanizing agent, with process parameters of reaction temperature 148℃ and holding time 28 minutes. The measured strength was 25MPa, hardness 66A, and elongation 430%, and all 10 tests conducted met the standards; The second sample contained 32.2% carbon black and 1.3% vulcanizing agent, with process parameters of reaction temperature 152℃ and holding time 32 minutes. The measured strength was 26MPa, hardness 67A, and elongation 428%, and 7 out of 8 tests conducted met the standards; The third sample contained 31.5% carbon black and 1.2% vulcanizing agent, with process parameters of reaction temperature 150℃ and holding time 30 minutes. The measured strength was 24.5MPa, hardness 65.5A, and elongation 425%, and all 12 tests conducted met the standards.

[0073] Targeted fine-tuning suggestions: The lower limit of the elongation range for candidate B, 410%, is lower than the target value of 420%. It is recommended to extend the holding time from 25 minutes to 28 minutes. Referring to the holding time of the second adjacent sample of 32 minutes, it is expected that the elongation range can be improved to between 420% and 435% after the adjustment. The reaction temperature of candidate C, 180℃, exceeds the historical reasonable range of 140℃ to 170℃, and the carbon black content of 40% exceeds the historical reasonable range of 28% to 35%. It belongs to the out-of-range sample. It is recommended to conduct simulation experiments under the conditions of reaction temperature of 170℃ and carbon black content of 35% to supplement the performance data in this area before evaluation.

[0074] Experimental decision-making logic: Candidate A's performance indicators are mostly within the target range, with a lower limit of strength (23 MPa > 20 MPa), hardness (65A to 67A) meeting the requirements of the upper and lower 2A ranges, and a lower limit of elongation (425% > 420%). Furthermore, the three most similar historical samples all successfully met the targets, so this experiment is prioritized. Candidate B only has an elongation range lower limit (410%) lower than the target value of 420%. Performance can be quickly optimized by fine-tuning the heat treatment time, so a re-evaluation after fine-tuning is recommended. Candidate C's strength, hardness, and elongation range lower limits are all below the target requirements: strength (18 MPa < 20 MPa), hardness (60A < 63A), and elongation (390% < 420%). Moreover, its process parameters significantly exceed the reasonable historical range, classifying it as an out-of-range sample. Direct experimentation carries high risk, so direct experimentation is rejected, and data supplementation is required first.

[0075] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0076] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A system for verifying material formulations using confidence intervals, characterized in that, include: The data acquisition module is used to collect basic data of candidate formulations, basic data of historical experiments, environmental correlation data, implicit data of materials and equipment status data, and to preprocess the collected data. The error calculation module is used to calculate and correct the errors in the preprocessed environmental correlation data, material implicit data and equipment status data to obtain the final error value. Based on the final error value, the module constructs and outputs the adjusted final confidence interval. The interval adjustment module is used to calculate the similarity and material influence value between the candidate formula and historical samples. Based on the comparison results of similarity and preset threshold and combined with the material influence value, the interval is adjusted and the target reliable interval is output. The stability assessment module is used to compare the basic data of candidate formulations with historical experimental data in real time to obtain the stability judgment result. The credibility assessment module is used to obtain a basic compliance assessment value based on the target credibility interval through random sampling, obtain a feasibility assessment value by combining the stability judgment result, and obtain a final credibility assessment value by merging the basic compliance assessment value and the feasibility assessment value, and output decision recommendations.

2. The system for verifying material formulations using confidence intervals according to claim 1, characterized in that: The basic data for the candidate formulations include component proportions and process parameters; Historical experimental data includes measured performance values, number of test repetitions, single measurement error, standard deviation of process parameter fluctuations, and the range of common ingredient proportions in similar formulations. Environmental data include high and low temperature cycling degradation rate, salt spray retention rate, and temperature and humidity difference; Implicit material data includes raw material purity, particle size distribution standard deviation, and glass transition temperature; Equipment status data includes calibration deviation, standard deviation of repeated measurements, running time of key components, and zero drift.

3. The system for verifying material formulations using confidence intervals according to claim 1, characterized in that: The specific steps for preprocessing the collected data are as follows: Acquire basic data of candidate formulations, basic data of historical experiments, environmental correlation data, implicit data of materials, and equipment status data. Clean the data to remove outliers and eliminate outliers that exceed a preset multiple of the standard deviation. Standardize the cleaned data into a unified format and map it into directly calculable quantitative values. Perform correlation matching on the standardized data to bind the correspondence between the formulation and the corresponding environmental, material, and equipment data to obtain a structured dataset.

4. The system for verifying material formulations using confidence intervals according to claim 1, characterized in that: The specific steps for error calculation and correction of the preprocessed environmental correlation data, material implicit data, and equipment status data are as follows: The preprocessed data is formed into a structured dataset and divided into a training set, a calibration set, and a validation set according to a preset ratio. Based on the clustering of the calibration set according to the component proportion range and the process temperature range, a specific quantile of the absolute value of the performance residual of each class is calculated as the basic error value; By combining implicit data of materials and introducing parameters such as raw material purity and particle size distribution standard deviation, the implicit correction coefficient of materials is obtained through comprehensive calculation. The equipment status is determined based on the equipment status data, and corresponding corrections are applied to the basic errors based on the determination results; Based on environmental correlation data, parameters such as high and low temperature cycle attenuation rate, salt spray retention rate and temperature and humidity difference are introduced, and environmental correction coefficients are obtained through comprehensive calculation. The final error value is obtained by combining the basic error with various correction coefficients. The coverage is verified using the validation set. If the preset target is not met, the final error value is scaled proportionally until the coverage meets the requirements, and the adjusted final confidence interval is output.

5. The system for verifying material formulations using confidence intervals according to claim 1, characterized in that: The specific steps for calculating the similarity between candidate formulations and historical samples are as follows: Obtain the component proportions and process parameters of candidate formulations and historical samples, calculate the distance evaluation value between them based on fixed feature dimensions, and obtain the similarity by combining the maximum historical distance.

6. The system for verifying material formulations using confidence intervals according to claim 1, characterized in that: The specific steps for obtaining the material influence value are as follows: Obtain the high and low temperature cycling decay rate, glass transition temperature, and the highest temperature of the target environment. Calculate the temperature difference between the glass transition temperature and the highest temperature of the target environment. Use this temperature difference and the high and low temperature cycling decay rate to obtain the material influence value through comprehensive calculation.

7. The system for verifying material formulations using confidence intervals according to claim 1, characterized in that: The specific steps for adjusting the range based on the similarity comparison result with the preset threshold and the material influence value are as follows: Two similarity thresholds are set: a first threshold and a second threshold. The first threshold corresponds to the high percentile of the historical sample similarity distribution, and the second threshold corresponds to the even higher percentile of the historical sample similarity distribution. The calculated similarity is then compared with the first threshold and the second threshold respectively. If the similarity is greater than or equal to the first threshold, the final confidence interval remains unchanged. If the similarity is between the first threshold and the second threshold and the material influence value is greater than or equal to the preset score, the interval width is adjusted according to the first ratio. If the similarity is between the first threshold and the second threshold and the material influence value is less than the preset score, the interval width is adjusted according to the second ratio. If the similarity is less than the second threshold, output a rejection conclusion and prompt for supplementary historical data of the basic data of the formula. Finally, output the adjusted target confidence interval.

8. The system for verifying material formulations using confidence intervals according to claim 1, characterized in that: The specific steps for obtaining the stability determination result are as follows: Extract the component proportions from the candidate formulation's basic data and compare them with the common component proportion ranges of similar formulations in historical experimental basic data. If the component proportions are within the common range, the component proportions are deemed reasonable; otherwise, the component proportions are deemed insufficiently reasonable. Extract process parameters from candidate formulation basic data, compare them with the standard deviation of process parameter fluctuations in historical experimental basic data, calculate the impact rate of process parameter fluctuations on performance, and determine the process is stable if the impact rate is less than or equal to the lower threshold; otherwise, determine the process is unstable. If the proportions are reasonable and the process is stable, it is determined that the stability is qualified; If the component is reasonable but the process is unstable, or if the component is not reasonable but the process is stable, it is judged as having moderate stability. If the proportions are not reasonable and the process is unstable, the stability is deemed unqualified, and the overall stability judgment result is obtained.

9. The system for verifying material formulations using confidence intervals according to claim 1, characterized in that: The specific steps by which the credibility assessment module obtains the final credibility assessment value and outputs decision recommendations are as follows: Perform random sampling a preset number of times within the target confidence interval, and statistically analyze the percentage of times multiple indicators simultaneously hit the preset target to obtain the basic compliance assessment value; Based on the stability assessment results output by the stability assessment module, different levels of feasibility assessment values ​​are assigned to different assessment scenarios; if the stability assessment result is qualified, the highest level of feasibility assessment value is assigned. If the rating is medium, assign a medium-level feasibility assessment value. If the result is unqualified, the lowest level of feasibility assessment value will be assigned. The final credible assessment value is obtained by integrating the basic compliance assessment value and the feasibility assessment value. The experiment priority is output based on the final credibility assessment value: if the credibility score reaches the high threshold, the priority experiment suggestion is output; if it is in the medium threshold range, the alternative experiment suggestion is output; if it is in the low threshold range, the exploratory experiment suggestion is output. If the value is below the lower threshold, a suggestion to postpone the experiment will be output. At the same time, based on the calibration deviation and zero drift in the equipment status data, a device calibration suggestion will be output. Based on the proportion of candidate formulation components and the deviation of process parameters from the historical reasonable range, a parameter adjustment scheme will be output.

10. A method for verifying material formulations using confidence intervals, applied to the material formulation verification system using confidence intervals as described in any one of claims 1-9, characterized in that: Includes the following steps: Step 1: Collect basic data of candidate formulations, basic data of historical experiments, environmental correlation data, implicit data of materials, and equipment status data, and preprocess the collected data; Step 2: Perform error calculation and correction on the preprocessed environmental correlation data, material implicit data and equipment status data to obtain the final error value, and construct and output the adjusted final confidence interval based on the final error value; Step 3: Calculate the similarity and material influence value between the candidate formulation and historical samples. Based on the comparison results of the similarity and the preset threshold, and combined with the material influence value, adjust the range and output the target confidence range. Step 4: Compare the basic data of the candidate formulation with the basic data of historical experiments in real time to obtain the stability determination results; Step 5: Based on the target confidence interval, obtain the basic compliance assessment value through random sampling, combine it with the stability judgment result to obtain the feasibility assessment value, integrate the basic compliance assessment value and the feasibility assessment value to obtain the final confidence assessment value, and output decision recommendations.

Images