A sensor-based quality monitoring method and system for nonferrous metal forgings

Abstract

The present invention relates to the field of data integration technology, and specifically to a sensor-based quality monitoring method and system for non-ferrous metal forgings, comprising the following steps: collecting grain size, dislocation density, and precipitated phase data, classifying them by processing section, process stage, and material batch, adjusting the data index sequence, constructing attribution rules, and forming integrated values ​​of microstructure parameters. In the present invention, by carefully screening stress mutation points and combining them with morphological combination intervals, the scientific division of stress response and the accurate identification of abnormal areas are ensured. The calculation of curvature change rate and inflection point spacing is used to improve the detection accuracy of abnormal points and reduce misjudgments. The combination of detailed analysis of stress growth trends and cluster analysis optimizes the defect identification process, improves the pertinence of monitoring and the accuracy of quality assessment, makes the quality monitoring of non-ferrous metal forgings more comprehensive, and enhances the reliability and applicability of the detection system.

Description

Technical Field

[0001] The present invention relates to the technical field of data integration, and in particular to a sensor-based nonferrous metal forging quality monitoring method and system. Background Art

[0002] The field of data integration technology encompasses the collection, storage, management, conversion, and analysis of multi-source, heterogeneous data to achieve standardization, structuring, and efficient utilization of data. The core content of this technical field involves sensor data acquisition, data format conversion, database management, data query optimization, and data fusion strategies. Data integration is widely used in industrial monitoring, intelligent manufacturing, healthcare, smart cities, and other fields, ensuring data consistency, integrity, and availability by building an efficient data management system. In industrial monitoring scenarios, data integration not only involves the standardized processing of information collected by physical sensors, but also includes data storage and retrieval based on database structures, as well as the use of specific data organization methods for quality assessment and monitoring decision support.

[0003] Among them, the sensor-based quality monitoring method for non-ferrous metal forgings refers to the use of sensors to collect physical parameters during the forging production process and process this data to monitor and evaluate the quality of the forgings. This method covers the acquisition of sensor signals, data format conversion, database structured storage, and data analysis. Specifically, it uses pressure sensors, temperature sensors, and deformation sensors to obtain stress, temperature, and deformation data during the forging process. Data preprocessing steps perform format normalization and noise elimination, and the data is indexed and stored in conjunction with the database structure. During the data retrieval stage, a feature comparison-based method is used for quality analysis to detect whether non-ferrous metal forgings have internal defects or external deviations.

[0004] Existing technologies for quality monitoring suffer from insufficient parameter matching accuracy. Data classification methods fail to effectively reflect the microstructural characteristics of different regions, resulting in an unclear mapping relationship between stress states and microscopic properties. Data consistency is low, failing to effectively mitigate noise interference, impacting the precise identification of abnormal areas. Stress mutation point screening relies on a fixed threshold and fails to integrate morphological factors for comprehensive evaluation, limiting the accuracy of identifying some mutation areas. Anomaly detection methods fail to fully consider the curvature change rate and inflection point spacing, potentially leading to misjudgments or omissions, impacting the accurate identification of defective areas. Stress growth trend analysis lacks data overlay calculations, resulting in a lack of precise positioning of stress accumulation areas, impacting the targeted nature of quality monitoring. Abnormal area determination relies primarily on a single physical parameter, failing to incorporate factors such as dislocation distribution and morphological changes, making it difficult to distinguish between different types of defects and reducing the sophistication of quality assessment. Insufficient multidimensional parameter analysis capabilities prevent the formation of a comprehensive quality anomaly monitoring system, limiting the ability to judge quality under complex working conditions. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a sensor-based quality monitoring method for non-ferrous metal forgings.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a sensor-based non-ferrous metal forging quality monitoring method, comprising the following steps:

[0007] S1: Collect grain size, dislocation density and precipitated phase data, classify them by processing section, process stage and material batch, adjust the data index order, build attribution rules, and form integrated values ​​of microstructural parameters;

[0008] S2: calling the integrated value of the microstructure parameter, collecting the stress data of the spindle sensor, matching the potential difference density by section, and normalizing it, classifying it according to the grain size and stress difference, screening the amplification mutation points, establishing the morphological combination relationship, and generating the stress mapping response interval value;

[0009] S3: calling the stress mapping response interval value, combining it with the edge point displacement data, calculating the curvature change rate, analyzing the inflection point spacing, screening abnormal points and superimposing the stress increase, determining the curvature abnormal area, and obtaining the curvature abnormal concentration value;

[0010] S4: Call the curvature anomaly concentration value, analyze the forging contour stress monitoring data, calculate the stress growth ratio per unit time, construct a growth curve, combine the stress mapping response value superposition analysis, screen the stress continuous growth section and locate the change focus area, and obtain the stress growth focus interval value.

[0011] As a further solution of the present invention, the integrated value of the microstructure parameters includes grain size parameters, dislocation density parameters, precipitation phase distribution parameters, regional classification index and stress matching index; the stress mapping response interval value includes equivalent stress data, dislocation density normalization value, stress difference classification value, amplification mutation point and morphological combination interval; the curvature anomaly concentration value includes curvature change rate, inflection point spacing, anomaly point set, stress amplification amount and structural response value; the stress growth focus interval value includes stress growth ratio, growth curve, baseline interval, continuous growth segment and change focus area.

[0012] As a further solution of the present invention, the steps for obtaining the integrated value of the microstructure parameter are:

[0013] S101: Based on the processing section number, process stage sequence and material batch identification, grain size data, dislocation density data and precipitated phase distribution data are collected, these data are combined and classified according to the processing section number, process stage sequence and material batch identification, structural parameters are called and grouped according to section characteristics to obtain a structural grouping parameter set;

[0014] S102: Based on the structure grouping parameter set, for the structure grouping parameter set, calculate the data parameter differences between different groups and construct data attribution rules using the formula:

[0015] ;

[0016] Calculate the stress difference coefficient between each group, adjust the data index order based on the regional classification standard, and obtain the stress matching index;

[0017] in, Representative Group With grouping The stress difference coefficient between and Represent the average stress values ​​of the two groups, Represents the influencing factor of the precipitation phase within the group, where represents the distribution influence coefficient of the precipitated phase, Represents the influence weight of the precipitated phase, represents the total number of precipitated phases;

[0018] S103: Based on the stress matching index, the data index order is adjusted to match the regional classification standard, the stress and morphology matching index between the structural data groups is formed, the overall microstructure parameter integrated value is calculated, and the microstructure parameter integrated value is obtained.

[0019] As a further solution of the present invention, the step of obtaining the stress mapping response interval value is:

[0020] S201: calling the integrated value of the microstructure parameter, collecting equivalent stress data output by the main shaft pressure sensor of the press forming machine, matching dislocation density data according to the segment number, adjusting the data format of the equivalent stress data and the dislocation density data, identifying the accuracy of the data correspondence, and obtaining a stress and dislocation matching data set;

[0021] S202: Based on the stress and dislocation matching data set, the data difference is calculated and normalized, and the data is classified according to the grain size value and the stress difference. The classified data is compared and analyzed for changes within the segment based on the amplification coefficient and the coupling distribution position, using the formula:

[0022] ;

[0023] Calculate the stress increase mutation values ​​in different sections, screen the increase mutation points, and obtain the increase mutation point set;

[0024] in, Representative section With segment The stress increase mutation value between and Separate sections and The mean equivalent stress of and Separate sections and The dislocation density value, is the normalization adjustment factor, and Separate sections and The distribution ratio of precipitated phase;

[0025] S203: Based on the set of amplification mutation points, a morphological combination interval is established according to the position, and the equivalent stress variation range within each morphological combination interval is calculated, and the data index sequence is adjusted to finally obtain the stress mapping response interval value.

[0026] As a further solution of the present invention, the step of obtaining the curvature anomaly concentration value is:

[0027] S301: calling the stress mapping response interval value, collecting edge point displacement sensor data and processing time series data, extracting displacement change values, and establishing a displacement sequence that changes with time, calculating the curvature change rate of continuous points, and obtaining a curvature change rate sequence;

[0028] S302: Based on the curvature change rate sequence, the distance between adjacent inflection points is calculated and compared with the set inflection point distance range, and curvature change points that meet the characteristics are screened. The abnormal point set is selected according to the change rate trend, using the formula:

[0029] ;

[0030] Calculate the curvature change rate anomaly index of each point, filter out the anomaly points, and obtain the anomaly point set;

[0031] in, represents the abnormal index of curvature change rate, Represents the displacement speed of the current adjacent point, Represents the displacement velocity of adjacent points, T i Represents the current corresponding time point and Represents the corresponding time points of adjacent points, represents the cumulative effect of curvature mutation, where is the local curvature value, is the corresponding inflection point spacing, To calculate the number of sample points;

[0032] S303: Based on the abnormal point set, combined with the stress increase and structural response value corresponding to the abnormal point, the curvature change frequency of each point is calculated, and the abnormal area is statistically analyzed to obtain the curvature abnormal concentration value.

[0033] As a further solution of the present invention, the step of obtaining the stress growth focus interval value is:

[0034] S401: Call the curvature abnormal concentration value, obtain the stress monitoring sensor data of the forging contour, extract the stress value at each time point, and construct a time series data set to calculate the stress growth ratio per unit time. Suppose a certain time point The corresponding stress value is , adjacent time points The corresponding stress value is , then the stress growth ratio per unit time is calculated as follows:

[0035] ;

[0036] Calculate the stress growth ratios at all time points and form a growth ratio sequence to obtain a stress growth ratio sequence;

[0037] in, Represents the stress growth ratio per unit time, Representative time point The corresponding stress value at Represents the current time point The corresponding stress value at Representative time point , Represents the current time point , represents a time interval;

[0038] S402: Based on the stress growth ratio sequence, a growth curve is constructed, and the growth ratio is compared with the segment difference baseline interval to screen abnormal growth segments. The trend change of the growth ratio is calculated according to the time series using the formula:

[0039] ;

[0040] Calculate the growth trend indicators for each time period, filter out the segments with continuous growth, and obtain the set of segments with continuous growth;

[0041] in, Representative time period The stress growth trend value, is the stress growth ratio, is the section difference baseline, is the number of sampling points, represents the normalized calculation of the local stress variation amplitude, where is the local stress increase, is the corresponding time interval, is the number of statistical samples;

[0042] S403: In combination with the stress mapping response interval value, a corresponding value superposition analysis is performed on the screened continuous growth segment set, and the change focus area is located, the stress increase trend in the area is calculated, and the abnormal area is focused and located to obtain the stress growth focus interval value.

[0043] As a further embodiment of the present invention, the method further includes: S5: calling the stress growth focus interval value, collecting regional potential difference density and structural trend data, calculating the change ratio, and making a judgment based on the stress mutation amplitude to identify the stress concentration point, establish a regional feature set, and combine the grain size and potential difference distribution sequence for cross comparison, merge defect types, and generate quality monitoring results;

[0044] The quality monitoring results include distribution concentration points, regional feature sets, change ratios, stress mutation amplitudes and dislocation distribution sequences;

[0045] The steps for obtaining the quality monitoring results are:

[0046] S501: Call the stress growth focus interval value, collect the potential difference density data and structural trend data of the corresponding area, arrange the data in time series, calculate the mutation amplitude value and morphological change density, calculate the potential difference density of each measurement point, and set a certain time point. The potential difference density at , adjacent time points The potential difference density at , then the potential density change ratio is calculated as follows:

[0047] ;

[0048] Calculate the change ratio of all time points to obtain the potential difference density change ratio;

[0049] in, represents the ratio of potential density change, Representative time point The potential difference density, Representative time point The potential difference density, represents the measurement time point, represents adjacent time points, represents a time interval;

[0050] S502: Based on the potential difference density change ratio, a combined judgment is performed according to the change ratio and the stress mutation amplitude to identify the distribution concentration points and establish a regional feature set, using the formula:

[0051] ;

[0052] Calculate and obtain the characteristic values ​​of the distribution points and construct the regional feature set;

[0053] in, Representative measurement point The characteristic index value of Represents the stress mutation amplitude of each measuring point, Represents the ratio of potential difference density change at the corresponding measurement point, represents the balancing factor, preventing the denominator from approaching zero, represents the eigenvalues ​​of local morphological parameters, Represents the morphological data of the corresponding area, Represents the number of measurement points, Represents the number of morphological data points;

[0054] S503: Combined with the regional feature set, cross-comparison is performed on the grain size and dislocation distribution sequence to determine the defect type within the region, statistically analyze the distribution of measurement points in the regional feature set, calculate the comparison deviation of abnormal points, analyze the type of structural defects, and obtain quality monitoring results.

[0055] Sensor-based quality monitoring system for non-ferrous metal forgings, including:

[0056] The microstructure analysis module obtains grain size data, dislocation density data, and precipitate phase distribution data, classifies the data according to the processing section number, process stage sequence, and material batch identification, establishes a section feature index, calculates the interval difference between grain size data and dislocation density data, constructs attribution rules, adjusts the data index order, forms a matching relationship, calculates the stress and morphological parameter matching ratio, and obtains the integrated value of the microstructure parameters;

[0057] The stress response matching module calls the integrated value of the microstructure parameters, matches the equivalent stress data and the dislocation density data, calculates the difference normalization, screens the amplification mutation point, and obtains the stress mapping response interval value;

[0058] The curvature fluctuation screening module calls the stress mapping response interval value, analyzes the curvature change rate of the displacement data and the processing time series data, screens the mutation area, and obtains the curvature abnormal concentration value;

[0059] The stress growth trend positioning module calls the curvature anomaly concentration value, analyzes the growth ratio of the stress monitoring data, screens the continuous growth area, and obtains the stress growth focus interval value;

[0060] The defect area aggregation module calls the stress growth focus interval value, analyzes the change ratio of dislocation density data and structural trend data, screens the concentrated area, and obtains the defect area feature set.

[0061] Compared with the prior art, the advantages and positive effects of the present invention are:

[0062] In the present invention, by carefully screening stress mutation points and combining them with morphological combination intervals, the scientific division of stress response and the accurate identification of abnormal areas are ensured. The calculation of curvature change rate and inflection point spacing is used to improve the detection accuracy of abnormal points and reduce misjudgments. The combination of detailed analysis of stress growth trends and cluster analysis optimizes the defect identification process, improves the targeted monitoring and the accuracy of quality assessment, makes the quality monitoring of non-ferrous metal forgings more comprehensive, and enhances the reliability and applicability of the detection system. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a flow chart of the main steps of the present invention;

[0064] Figure 2 This is a flow chart for obtaining the integrated value of the microstructure parameters of the present invention;

[0065] Figure 3 This is a flow chart for obtaining stress mapping response interval values ​​according to the present invention;

[0066] Figure 4 This is a flow chart for obtaining the curvature anomaly concentration value of the present invention;

[0067] Figure 5 This is a flow chart for obtaining the stress growth focus interval value of the present invention;

[0068] Figure 6 The figure is a flow chart for obtaining the quality monitoring results of the present invention. DETAILED DESCRIPTION

[0069] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0070] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0071] See also Figure 1 , a sensor-based nonferrous metal forging quality monitoring method, comprising the following steps:

[0072] S1: Microstructure measurement and regional parameter collation, collecting grain size data, dislocation density data and precipitate distribution data, combining and classifying them according to processing section number, process stage sequence and material batch identification, calling structural parameters and grouping them according to section characteristics, building attribution rules based on parameter differences between data, adjusting data index order according to regional classification standards, forming stress and morphology basis matching indexes between structural data groups, and generating integrated values ​​of microstructure parameters;

[0073] S2: Construction of stress response characteristics and mapping interval analysis: call the integrated value of microstructure parameters, collect the equivalent stress data output by the pressure forming machine spindle pressure sensor, match the potential difference density data according to the segment number, and perform normalized calculation and processing on the difference. Perform structural classification according to the grain size value and stress difference value, compare the change segment based on the amplification coefficient and the coupling distribution position, screen the amplification mutation point, and establish the morphological combination interval according to the position to generate the stress mapping response interval value;

[0074] S3: Morphological curvature fluctuation detection and outlier inference: call the stress mapping response interval value, collect edge point displacement sensor data and processing time series data, calculate the curvature change rate of continuous points, and compare the distance between adjacent inflection points with the set inflection point distance range. Select the outlier point set according to the change rate trend, combine the corresponding stress increase value of the outlier point with the structural response value for superposition judgment, screen the area with abnormal curvature change frequency, and obtain the curvature anomaly concentration value;

[0075] S4: Analyze the continuous growth trend of stress and locate key areas. Call the curvature abnormal concentration value, obtain the stress monitoring sensor data of the forging contour, calculate the stress growth ratio per unit time according to the time series, and construct a growth curve. Compare the growth ratio with the segment difference baseline interval. Combined with the stress mapping response interval value, perform corresponding value superposition analysis, screen the continuous growth segment and locate the change focus area, and obtain the stress growth focus interval value;

[0076] S5: Identify clustered structures of abnormal areas and merge defect types, call the stress growth focus interval value, collect the corresponding regional difference density and structural trend data, calculate the change ratio based on the mutation amplitude value and morphological change density, make a combined judgment based on the change ratio and stress mutation amplitude, identify the distribution concentration points and establish a regional feature set, combine the grain size and dislocation distribution sequence for cross-comparison, and generate quality monitoring results.

[0077] The integrated values ​​of microstructure parameters include grain size parameters, dislocation density parameters, precipitation phase distribution parameters, regional classification index and stress matching index; the stress mapping response interval values ​​include equivalent stress data, normalized value of dislocation density, stress difference classification value, amplification mutation point and morphological combination interval; the curvature anomaly concentration value includes curvature change rate, inflection point spacing, anomaly point set, stress amplification amount and structural response value; the stress growth focus interval value includes stress growth ratio, growth curve, baseline interval, continuous growth section and change focus area; the quality monitoring results include distribution concentration point, regional feature set, change ratio, stress mutation amplitude and dislocation distribution sequence.

[0078] See also Figure 2 , the steps to obtain the integrated value of microstructure parameters are:

[0079] S101: Based on the processing section number, process stage sequence and material batch identification, grain size data, dislocation density data and precipitated phase distribution data are collected, these data are combined and classified according to the processing section number, process stage sequence and material batch identification, structural parameters are called and grouped according to section characteristics to obtain a structural grouping parameter set;

[0080] The process of collecting data on grain size, dislocation density, and precipitate distribution begins with meticulous microscopic examination of physical samples from different processing sections to determine the grain size distribution, dislocations within the crystal, and the specific morphology and distribution of precipitates. For example, when analyzing materials for aircraft engine turbine blades, operators take samples from different processing batches and observe and record these structural features using scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Data for each batch includes the average grain size, the number of dislocations per square millimeter, and the type and distribution density of precipitates. This analysis yields detailed data related to the processing section number, process stage sequence, and material batch identification. This data is then organized and recorded by process number and material batch. This data classification and recording is crucial for subsequent material property analysis. For example, when recording data, not only the average grain size but also the proportion of grains in each size range is recorded in detail. This allows for subsequent data comparison to identify differences between material batches, fluctuations in material properties, and the acquisition of structural grouping parameter sets.

[0081] S102: Based on the structure grouping parameter set, calculate the data parameter differences between different groups for the structure grouping parameter set, and construct data attribution rules using the formula:

[0082] ;

[0083] Calculate the stress difference coefficient between each group, adjust the data index order based on the regional classification standard, and obtain the stress matching index;

[0084] in, Representative Group With grouping The stress difference coefficient between and Represent the average stress values ​​of the two groups, Represents the influencing factor of the precipitation phase within the group, where represents the distribution influence coefficient of the precipitated phase, Represents the influence weight of the precipitated phase, represents the total number of precipitated phases;

[0085] For the structural grouping parameter set, the data parameter differences between different groups are analyzed by calculating the differences in the average grain size, dislocation density and precipitate phase distribution density of each group. For example, in the production process of automobile engine piston materials, the raw materials provided by different suppliers may have different internal structural characteristics due to differences in manufacturing processes. By comparing and calculating these difference values, the quality of the supplier's materials can be evaluated.

[0086] Calculation steps:

[0087] Get material batch data:

[0088] Two different material supply batches were selected, marked as batch A and batch B, and sample data was collected. Five material samples were randomly selected from each batch for analysis, and the grain size (μm), dislocation density ( m ) and precipitate distribution density (%). For example, the grain sizes of the five samples of batch A are 12.3μm, 13.1μm, 11.8μm, 12.6μm, and 12.9μm, respectively; the grain sizes of the five samples of batch B are 14.2μm, 14.7μm, 13.9μm, 14.3μm, and 14.5μm, respectively. At the same time, the dislocation densities of batch A are 4.2, 4.8, 4.5, 4.3, and 4.7, respectively, while those of batch B are 5.3, 5.6, 5.1, 5.4, and 5.2, respectively. The precipitate distribution densities of batch A are 3.5%, 3.2%, 3.7%, 3.4%, and 3.6%, respectively, and those of batch B are 2.9%, 3.1%, 2.8%, 3.0%, and 3.2%, respectively.

[0089] Calculate the average grain size, dislocation density, and precipitate distribution density of batch A and batch B:

[0090] The average grain size of batch A is calculated to be (12.3+13.1+11.8+12.6+12.9) / 5=12.54μm, the average grain size of batch B is (14.2+14.7+13.9+14.3+14.5) / 5=14.32μm, and the average dislocation density of batch A is (4.2+4.8+4.5+4.3+4.7) / 5=4.5× m , the average dislocation density of batch B is (5.3+5.6+5.1+5.4+5.2) / 5=5.32× m The average precipitate phase distribution density of batch A is (3.5+3.2+3.7+3.4+3.6) / 5=3.48%, and the average precipitate phase distribution density of batch B is (2.9+3.1+2.8+3.0+3.2) / 5=2.96%.

[0091] Calculate the stress difference coefficient between groups

[0092] Using the formula:

[0093] ;

[0094] in, , , , , set the influence weight of precipitated phase , then calculate:

[0095] ;

[0096] ;

[0097] ;

[0098] Get the stress matching index:

[0099] because Within the set threshold range (0.5-2.5), it means that there are moderate differences in the microstructural parameters of batches A and B, which are suitable for specific engineering needs. , it means that the material properties are quite different and not suitable for the same purpose. The material matching strategy needs to be readjusted. , it means that the structural characteristics of the two batches are basically the same and can be used interchangeably without affecting the engineering stability, and finally the stress difference coefficient and stress matching index are obtained.

[0100] S103: Based on the stress matching index, the data index order is adjusted to match the regional classification standard, the stress and morphology matching indexes between the structural data groups are formed, and the integrated value of the overall microstructure parameters is calculated to obtain the integrated value of the microstructure parameters;

[0101] Adjust the data index order to form stress and morphological matching indexes between structural data groups. For example, in the steel supply of large structures such as bridges or high-rise buildings, the distribution and use of materials can be optimized by analyzing the stress matching indexes of different batches of steel. The specific steps of adjusting the data index order include sorting according to the stress difference coefficient of each batch to ensure that the materials used in the structure have similar physical properties, which helps to reduce structural problems caused by uneven material properties. In this way, engineers can manage the use of materials more effectively and ensure the quality of the project. Finally, the integrated value of the microstructure parameters obtained by calculation not only provides engineers with a quantitative data reference, but also helps to make more reasonable decisions in the design and construction stages. The calculation and acquisition of the integrated value of the microstructure parameters.

[0102] See also Figure 3 , the steps to obtain the stress mapping response interval value are:

[0103] S201: calling the integrated value of the microstructure parameter, collecting the equivalent stress data output by the main shaft pressure sensor of the press forming machine, matching the dislocation density data according to the segment number, adjusting the data format of the equivalent stress data and the dislocation density data, identifying the accuracy of the data correspondence, and obtaining the stress and dislocation matching data set;

[0104] Call the integrated value of the microstructure parameter, collect the equivalent stress data output by the pressure forming machine spindle pressure sensor, and match the dislocation density data according to the segment number. First, extract the grain size, dislocation density and precipitate phase distribution data of different processing segments from the integrated value of the microstructure parameter, and obtain the processing number and processing parameters of each segment. Then, collect the equivalent stress data recorded by the pressure forming machine spindle pressure sensor during the material deformation process. The data is usually stored in the form of a continuous signal and must be segmented into a time series. The signal must be split and numbered according to different processing segments to ensure that the data can be matched with the processing segment number. For example, in the hot pressing process of high-precision titanium alloy parts, the stress data under different temperature and pressure conditions need to be segmented according to the time axis to keep it consistent with the microstructure data of the material. Next, the dislocation density data is called. This data comes from the material microstructure detection experiment. The electron backscatter diffraction (EBSD) technology is usually used to measure the dislocation density of different grain areas. The measured dislocation density data is matched with the stress data according to the processing section number, corresponding to the microstructure and external load information of each sample section, providing complete stress-structure matching data for subsequent analysis, and obtaining the stress and dislocation matching data set.

[0105] S202: Based on the stress and dislocation matching data set, the data difference is calculated and normalized, and the data is classified according to the grain size value and the stress difference. The classified data is compared and analyzed for changes within the segment based on the amplification coefficient and the coupling distribution position, using the formula:

[0106] ;

[0107] Calculate the stress increase mutation values ​​in different sections, screen the increase mutation points, and obtain the increase mutation point set;

[0108] in, Representative section With segment The stress increase mutation value between and Separate sections and The mean equivalent stress of and Separate sections and The dislocation density value, is the normalization adjustment factor, and Separate sections and The distribution ratio of precipitated phase;

[0109] Calculate the data difference and perform normalization calculation on the difference. Perform data classification according to the grain size value and stress difference. Compare and analyze the changes in the segment of the classified data based on the amplification coefficient and the coupling distribution position. The process first calculates the stress difference of each segment. Suppose a segment The mean stress of , adjacent segments The mean stress of , then the stress difference is calculated as:

[0110] ;

[0111] Assume that the mean equivalent stress values ​​of two sections are and , then calculate:

[0112] ;

[0113] Simultaneously calculate the dislocation density difference To measure the micro stress changes inside the material, if the section and The dislocation densities are m and m ,but:

[0114] ;

[0115] Subsequently, the stress and dislocation density differences are normalized to eliminate the influence of the numerical magnitude between different samples. The normalization calculation formula is:

[0116] ;

[0117] in, is the normalization adjustment factor, assuming , , , then calculate:

[0118] ;

[0119] ;

[0120] like Exceeding a certain threshold, e.g. , then the point is determined to be an amplification mutation point, otherwise it is considered to be a stable change interval, and finally the amplification mutation point set is obtained.

[0121] S203: Based on the set of amplification mutation points, a morphological combination interval is established according to the position, and the equivalent stress variation range within each morphological combination interval is calculated, and the data index sequence is adjusted to finally obtain the stress mapping response interval value;

[0122] The morphological combination interval is established according to the position, and the equivalent stress variation range within each morphological combination interval is calculated. The process first determines the adjacent stress mutation area according to the amplification mutation point set, takes the amplification mutation point as the boundary of the morphological combination interval, divides it into multiple intervals, and calculates the stress range of each interval. Let The initial stress value of each interval is The final stress value is , the stress interval is calculated as follows:

[0123] ;

[0124] If the initial stress value and the final stress value within a certain morphological combination interval are and , then calculate:

[0125] ;

[0126] Among them, if In a certain range, for example , then the morphological combination interval is marked as a stable interval, otherwise it is marked as an unstable interval. For example, in the stamping process of aviation structural parts, if the equivalent stress variation within a certain morphological combination interval exceeds 2.0 MPa, it is necessary to adjust the process parameters to optimize the material deformation process, and finally adjust the data index order to obtain the stress mapping response interval value.

[0127] See also Figure 4 , the steps to obtain the curvature anomaly concentration value are:

[0128] S301: calling the stress mapping response interval value, collecting edge point displacement sensor data and processing time series data, extracting displacement change values, and establishing a displacement sequence that changes with time, calculating the curvature change rate of continuous points, and obtaining a curvature change rate sequence;

[0129] Collect edge point displacement sensor data and processing time series data, extract displacement change values, establish a displacement sequence that changes with time, and calculate the curvature change rate of continuous points. First, call the stress mapping response interval value to obtain the stress distribution information of each processing section, and obtain edge point displacement sensor data. The sensor is installed at the key position of the processing equipment, such as around the stamping point of the sheet metal bending machine or the key stress area of ​​the large mechanical component. The sensor records the displacement of the edge point in real time and stores the data as a time series signal. To ensure data quality, data cleaning is first performed to remove abnormal data introduced by mechanical vibration or measurement error. Subsequently, the processing time series data is obtained and matched with the displacement sensor data on the time axis to ensure that the collected data corresponds to the correct processing stage. On this basis, the curvature change rate of the continuous points is calculated. Suppose the adjacent displacement data within a certain time period is 、 、 , then the steps for calculating the curvature change rate are as follows:

[0130] Compute the Euclidean distance between adjacent points:

[0131] ;

[0132] ;

[0133] Calculate the rate of change of curvature:

[0134] ;

[0135] Substitute the numerical value and calculate , store it into the curvature change rate sequence, and obtain the curvature change rate sequence.

[0136] S302: Based on the curvature change rate sequence, the distance between adjacent inflection points is calculated and compared with the set inflection point distance range. Curvature change points that meet the characteristics are screened and an abnormal point set is selected according to the change rate trend. The formula is:

[0137] ;

[0138] Calculate the curvature change rate anomaly index of each point, filter out the anomaly points, and obtain the anomaly point set;

[0139] in, represents the abnormal index of curvature change rate, Represents the displacement speed of the current adjacent point, represents the displacement velocity of adjacent points, Represents the current corresponding time point and Represents the corresponding time points of adjacent points, represents the cumulative effect of curvature mutation, where is the local curvature value, is the corresponding inflection point spacing, To calculate the number of sample points;

[0140] Calculate the distance between adjacent inflection points and compare it with the set inflection point distance range to screen the curvature change points that meet the characteristics. Select the abnormal point set according to the change rate trend. Suppose that the curvature change inflection point data of a workpiece during the processing is 、 、 , set the inflection point spacing threshold to 2.5mm, and calculate the spacing between each adjacent inflection point:

[0141] Calculate the inflection point spacing:

[0142] ;

[0143] Suppose the coordinates of two inflection points are and ,calculate:

[0144] ;

[0145] If the calculated inflection point distance is less than the threshold , then it is determined that the point may be an outlier.

[0146] Calculate the curvature change rate anomaly index:

[0147] ;

[0148] in, Represents the displacement speed of the current adjacent point, Represents the displacement velocity of adjacent points, T i Represents the current corresponding time point and Represents the corresponding time points of adjacent points, represents the cumulative effect of curvature mutation, where is the local curvature value, is the corresponding inflection point spacing, To calculate the number of sample points, we can get ,like , then the point is determined to be an outlier, and the outlier point set is finally obtained.

[0149] S303: Based on the set of abnormal points, the curvature change frequency of each point is calculated by combining the stress increase and the structural response value corresponding to the abnormal point, and the abnormal area is statistically analyzed to obtain the curvature abnormal concentration value;

[0150] Combined with the stress increase and structural response value corresponding to the abnormal point, the curvature change frequency of each point is calculated, and the abnormal area is statistically analyzed. Assuming that the stress increase of a point is 3.2MPa, and the stress increase of its adjacent point is 1.8MPa, the calculation is:

[0151] Calculate the change in stress increase:

[0152] ;

[0153] Substitute into the calculation:

[0154] ;

[0155] like MPa, it is determined that there may be an abnormality in the area corresponding to the point.

[0156] Calculate the frequency of curvature changes:

[0157] ;

[0158] in, is the frequency of curvature change, is the curvature value of the outlier point, is the total processing time. Assuming the total processing time is 60s, calculate:

[0159] ;

[0160] If the number of curvature changes per unit time is greater than the set threshold of 0.025, the area is marked as a curvature anomaly concentration area, and the curvature anomaly concentration value is finally obtained.

[0161] See also Figure 5 , the steps to obtain the stress growth focus interval value are:

[0162] S401: Call the curvature abnormal concentration value, obtain the stress monitoring sensor data of the forging contour, extract the stress value at each time point, and construct a time series data set to calculate the stress growth ratio per unit time. Suppose a certain time point The corresponding stress value is , adjacent time points The corresponding stress value is , then the stress growth ratio per unit time is calculated as follows:

[0163] ;

[0164] Calculate the stress growth ratios at all time points and form a growth ratio sequence to obtain a stress growth ratio sequence;

[0165] in, Represents the stress growth ratio per unit time, Representative time point The corresponding stress value at Represents the current time point The corresponding stress value at Representative time point , Represents the current time point , represents a time interval;

[0166] Obtain stress monitoring sensor data at the contour of the forging, extract stress values ​​at each time point, and construct a time series data set to calculate the stress growth ratio per unit time. First, call the curvature anomaly concentration value, which is derived from the statistical analysis of curvature changes in the previous step and can reflect the morphological anomaly in a specific area. Then, obtain stress monitoring sensor data installed at the contour of the forging. These sensors are usually distributed in areas with significant stress, such as corners, edges, or impact points. Read continuous stress measurement values ​​through the data interface and perform time series sorting to ensure the integrity and continuity of the data. Next, extract stress values ​​at each time point and construct a time series data set. The data set is arranged in ascending order of time for subsequent calculations. On this basis, calculate the stress growth ratio per unit time. Suppose a certain time point The corresponding stress value is , adjacent time points The corresponding stress value is , then the stress growth ratio per unit time is calculated as follows:

[0167] ;

[0168] Assume that the time series data of a test point is: MPa, MPa, s, then calculate:

[0169] ;

[0170] Calculate the stress growth ratios at all time points, form a growth ratio sequence, and obtain the stress growth ratio sequence.

[0171] S402: Based on the stress growth ratio sequence, a growth curve is constructed, and the growth ratio is compared with the segment difference baseline interval to screen the abnormal growth segments. The trend change of the growth ratio is calculated according to the time series using the formula:

[0172] ;

[0173] Calculate the growth trend indicators for each time period, filter out the segments with continuous growth, and obtain the set of segments with continuous growth;

[0174] in, Representative time period The stress growth trend value, is the stress growth ratio, is the section difference baseline, is the number of sampling points, represents the normalized calculation of the local stress variation amplitude, where is the local stress increase, is the corresponding time interval, is the number of statistical samples;

[0175] Construct a growth curve chart and compare the growth ratio with the interval of the segment difference baseline to screen out abnormal growth segments. First, use the stress growth ratio sequence to draw a stress growth trend curve, which can intuitively show the change in stress change rate. Then, set the segment difference baseline. The baseline is determined by historical data or experimental tests, which represents the range of change in the stress growth rate of each segment under normal circumstances. By comparing the growth ratio with the baseline range, screen out abnormal growth segments that exceed the preset interval. On this basis, calculate the trend change of the growth ratio according to the time series using the formula:

[0176] ;

[0177] Calculate the growth trend indicators for each time period and filter out the segments with continuous growth, among which, Representative time period The stress growth trend value, is the stress growth ratio, is the section difference baseline, is the number of sampling points, represents the normalized calculation of the local stress variation amplitude, where is the local stress increase, is the corresponding time interval, To count the number of samples, for example, in an experiment, let MPa / s, MPa / s, , MPa, s, calculate:

[0178] ;

[0179] ;

[0180] ;

[0181] If the calculated Exceeding a certain threshold, such as , then the segment is determined to be a continuously growing segment, and finally a continuously growing segment set is obtained.

[0182] S403: Based on the stress mapping response interval value, perform corresponding value superposition analysis on the screened continuous growth segment set, locate the change focus area, calculate the stress increase trend in the area, focus on and locate the abnormal area, and obtain the stress growth focus interval value;

[0183] Perform corresponding value superposition analysis on the selected continuous growth segment set and locate the focus area of ​​change. First, call the stress mapping response interval value, which represents the stress change situation in different areas. Through data matching, the stress change trend of the continuous growth segment is superimposed with the stress characteristics in the mapping interval. In this process, the stress increase trend of each growth segment is calculated. If at a certain time point The corresponding stress increase is , then use the calculation formula:

[0184] ;

[0185] in, Represents the stress trend value of the current segment, represents the stress value at the previous time point, is the sum of stress increases at each point in the current section, is the total number of calculation points in the segment. If the stress increase data for a certain period is MPa, calculation:

[0186] ;

[0187] ;

[0188] If the calculation results show that the stress value increases continuously and the increase trend exceeds a certain set threshold, the area is determined to be a stress growth focus area, and the stress growth focus interval value is finally obtained.

[0189] See also Figure 6 , the steps to obtain the quality monitoring results are:

[0190] S501: Call the stress growth focus interval value, collect the potential difference density data and structural trend data of the corresponding area, arrange the data in time series, calculate the mutation amplitude value and morphological change density, calculate the potential difference density of each measurement point, and set a certain time point. The potential difference density at , adjacent time points The potential difference density at , then the potential density change ratio is calculated as follows:

[0191] ;

[0192] Calculate the change ratio of all time points to obtain the potential difference density change ratio;

[0193] in, represents the ratio of potential density change, Representative time point The potential difference density, Representative time point The potential difference density, represents the measurement time point, represents adjacent time points, Represents a time interval.

[0194] Collect the potential difference density data and structural trend data of the corresponding area, arrange them in time series, and calculate the mutation amplitude value and morphological change density. First, calculate the potential difference density of the measurement point. Assume that at a certain time point The potential difference density at , adjacent time points The potential difference density at , then the potential density change ratio can be calculated as follows:

[0195] ;

[0196] The calculation steps of the embodiment are as follows:

[0197] Set measurement time point s, s;

[0198] The potential difference density data corresponding to the measurement point is m , m ;

[0199] Calculate the time interval:

[0200] ;

[0201] Calculate the potential difference density change ratio:

[0202] ;

[0203] Calculate the change ratios at all time points to obtain the potential difference density change ratios.

[0204] S502: Based on the potential difference density change ratio, a combined judgment is made based on the change ratio and the stress mutation amplitude to identify the distribution concentration points and establish a regional feature set using the formula:

[0205] ;

[0206] Calculate and obtain the characteristic values ​​of the distribution points and construct the regional feature set;

[0207] in, Representative measurement point The characteristic index value of Represents the stress mutation amplitude of each measuring point, Represents the ratio of potential difference density change at the corresponding measurement point, represents the balancing factor, preventing the denominator from approaching zero, represents the eigenvalues ​​of local morphological parameters, Represents the morphological data of the corresponding area, Represents the number of measurement points, Represents the number of morphological data points.

[0208] Based on the combined judgment of the change ratio and the stress mutation amplitude, the distribution concentration points are identified and the regional feature set is established. The following calculation method is used:

[0209] ;

[0210] The calculation steps of the embodiment are as follows:

[0211] Set the number of regional feature points , number of morphological data points ;

[0212] Assume that the stress mutation amplitude at the measuring point is MPa, MPa, MPa;

[0213] Set the potential difference density change ratio m / s, m / s, m / s;

[0214] Set the balance factor To prevent the denominator from approaching zero;

[0215] Let the local morphological parameter eigenvalue , ;

[0216] Assume that the morphological data of the corresponding area , ;

[0217] Calculate the characteristic index value:

[0218] ;

[0219] Calculate the first part:

[0220] ;

[0221] Calculate the second part:

[0222] ;

[0223] The operation obtains the characteristic values ​​of the distribution points and constructs the regional feature set.

[0224] S503: Combine the regional feature set to cross-compare the grain size and dislocation distribution sequence, determine the defect type within the region, count the distribution of measurement points in the regional feature set, calculate the comparison deviation of abnormal points, analyze the type of structural defects, and obtain quality monitoring results;

[0225] First, the grain size data of the measurement points are screened to determine the grain size differences between adjacent measurement points. The data are then arranged in a time series to observe the fluctuations in grain size between consecutive measurement points. If the grain size continuously decreases within a short period of time, it indicates that uneven microstructural shrinkage may have occurred in the area, which requires further analysis based on the dislocation distribution. During the analysis, the grain size data of multiple locations within the measurement area are first obtained, and the dislocation density information of different measurement points is recorded to ensure the integrity of the data. For each measurement point, the spatial distribution characteristics of its grain size are statistically analyzed and compared with the grain size of its adjacent measurement points. The magnitude of the grain size change in different areas is calculated. If the grain size change trend in a certain area is significantly different from the overall distribution, further anomaly analysis based on the dislocation density data is required. For example, within a measurement area, if the grain size of some measurement points continues to shrink while the grain size of other areas remains stable, this may indicate that stress concentration or abnormal plastic deformation of the material exists in the area, which requires confirmation based on the dislocation distribution. Further analysis of the dislocation distribution is conducted. By measuring the changing trend of the dislocation density, it is determined whether there is local stress concentration in the area. If the dislocation density increases significantly in a short period of time, it indicates that the microstructure of the area has changed, which may lead to the formation of local defects. In this case, it is necessary to cross-analyze the changing trend of the grain size and the change of the dislocation density to determine whether there is an area of ​​defect accumulation. For example, during the measurement process, if it is found that the grain size of a certain area continues to decrease, while the dislocation density increases rapidly between adjacent measurement points, it indicates that the area may be in a critical state of local plastic deformation, and further analysis is required to determine whether it will affect the stability of the overall structure. Through the above cross-comparison, the defect type of the area is finally determined and classified into different quality monitoring results. If the grain size of the area changes dramatically and the dislocation density increases significantly, it can be determined that there is plastic deformation abnormality in the area, which may affect the performance of the material. If the dislocation density changes slightly, but the fluctuation of the grain size exceeds the normal range, it can be classified as microstructure abnormality, which may affect the uniformity of the material. If the grain size and dislocation density of the area are both within the normal range, it can be classified as a defect-free area, indicating that the material quality of the area meets the standard requirements. Finally, by analyzing the characteristic data of different areas, summarizing the change trend of its microstructure, and based on the cross-comparison results, generating quality monitoring results for further optimizing the material processing technology or screening abnormal areas.

[0226] Sensor-based quality monitoring system for non-ferrous metal forgings, including:

[0227] The microstructure analysis module obtains grain size data, dislocation density data, and precipitate phase distribution data, classifies the data according to the processing section number, process stage sequence, and material batch identification, establishes a section feature index, calculates the interval difference between grain size data and dislocation density data, constructs attribution rules, adjusts the data index order, forms a matching relationship, calculates the stress and morphological parameter matching ratio, and obtains the integrated value of the microstructure parameters;

[0228] The stress response matching module calls the integrated value of microstructure parameters, matches the equivalent stress data and dislocation density data, calculates the difference normalization, screens the amplification mutation point, and obtains the stress mapping response interval value;

[0229] The curvature fluctuation screening module calls the stress mapping response interval value, analyzes the curvature change rate of the displacement data and processing time series data, screens the mutation area, and obtains the curvature abnormal concentration value;

[0230] The stress growth trend positioning module calls the curvature anomaly concentration value, analyzes the growth ratio of the stress monitoring data, screens the continuous growth area, and obtains the stress growth focus interval value;

[0231] The defect area aggregation module calls the stress growth focus interval value, analyzes the change ratio of dislocation density data and structural trend data, screens the concentrated area, and obtains the defect area feature set.

[0232] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A sensor-based method for monitoring the quality of nonferrous metal forgings, characterized in that: The following steps are involved: S1: Collect grain size, dislocation density and precipitated phase data, classify them by processing section, process stage and material batch, adjust the data index order, build attribution rules, and form integrated values ​​of microstructural parameters; S2: calling the integrated value of the microstructure parameter, collecting the stress data of the spindle sensor, matching the potential difference density by section, and normalizing it, classifying it according to the grain size and stress difference, screening the amplification mutation points, establishing the morphological combination relationship, and generating the stress mapping response interval value; S3: calling the stress mapping response interval value, combining it with the edge point displacement data, calculating the curvature change rate, analyzing the inflection point spacing, screening abnormal points and superimposing the stress increase, determining the curvature abnormal area, and obtaining the curvature abnormal concentration value; S4: calling the curvature abnormal concentration value, analyzing the forging contour stress monitoring data, calculating the stress growth ratio per unit time, constructing a growth curve, combining the stress mapping response value superposition analysis, screening the stress continuous growth section and locating the change focus area, and obtaining the stress growth focus interval value; The microstructure parameter integration value includes a grain size parameter, a dislocation density parameter, a precipitated phase distribution parameter, a regional classification index, and a stress matching index; the stress mapping response interval value includes equivalent stress data, a dislocation density normalized value, a stress difference classification value, an amplification mutation point, and a morphological combination interval; the curvature anomaly concentration value includes a curvature change rate, an inflection point spacing, an anomaly point set, a stress amplification amount, and a structural response value; and the stress growth focus interval value includes a stress growth ratio, a growth curve, a baseline interval, a continuous growth section, and a change focus area; The steps for obtaining the stress mapping response interval value are as follows: S201: calling the integrated value of the microstructure parameter, collecting equivalent stress data output by the main shaft pressure sensor of the press forming machine, matching dislocation density data according to the segment number, adjusting the data format of the equivalent stress data and the dislocation density data, identifying the accuracy of the data correspondence, and obtaining a stress and dislocation matching data set; S202: Based on the stress and dislocation matching data set, the data difference is calculated and normalized, and the data is classified according to the grain size value and the stress difference. The classified data is compared and analyzed for changes within the segment based on the amplification coefficient and the coupling distribution position, using the formula: ； Calculate the stress increase mutation values ​​in different sections, screen the increase mutation points, and obtain the increase mutation point set; in, Representative section With segment The stress increase mutation value between and Separate sections and The mean equivalent stress of and Separate sections and The dislocation density value, is the normalization adjustment factor, and Separate sections and The distribution ratio of precipitated phase; S203: Based on the set of amplification mutation points, establishing morphological combination intervals by position, calculating the equivalent stress variation range within each morphological combination interval, adjusting the data index sequence, and finally obtaining stress mapping response interval values; The method further includes, S5: calling the stress growth focus interval value, collecting regional potential difference density and structural trend data, calculating a change ratio, and making a judgment based on the stress mutation amplitude to identify stress concentration points, establish a regional feature set, and combine grain size and potential difference distribution sequence cross comparison to merge defect types and generate quality monitoring results; The quality monitoring results include distribution concentration points, regional feature sets, change ratios, stress mutation amplitudes and dislocation distribution sequences; The steps for obtaining the quality monitoring results are: S501: Call the stress growth focus interval value, collect the potential difference density data and structural trend data of the corresponding area, arrange the data in time series, calculate the mutation amplitude value and morphological change density, calculate the potential difference density of each measurement point, and set a certain time point. The potential difference density at , adjacent time points The potential difference density at , then the potential density change ratio is calculated as follows: ； Calculate the change ratio of all time points to obtain the potential difference density change ratio; in, represents the ratio of potential density change, Representative time point The potential difference density, Representative time point The potential difference density, represents the measurement time point, represents adjacent time points, represents a time interval; S502: Based on the potential difference density change ratio, a combined judgment is performed according to the change ratio and the stress mutation amplitude to identify the distribution concentration points and establish a regional feature set, using the formula: ； Calculate and obtain the characteristic values ​​of the distribution points and construct the regional feature set; in, Representative measurement point The characteristic index value of Represents the stress mutation amplitude of each measuring point, Represents the ratio of potential difference density change at the corresponding measurement point, represents the balancing factor, preventing the denominator from approaching zero, represents the eigenvalues ​​of local morphological parameters, Represents the morphological data of the corresponding area, Represents the number of measurement points, Represents the number of morphological data points; S503: Combined with the regional feature set, cross-comparison is performed on the grain size and dislocation distribution sequence to determine the defect type within the region, statistically analyze the distribution of measurement points in the regional feature set, calculate the comparison deviation of abnormal points, analyze the type of structural defects, and obtain quality monitoring results.

2. The sensor-based nonferrous metal forging quality monitoring method according to claim 1, characterized in that: The steps for obtaining the integrated value of the microstructure parameter are: S101: Based on the processing section number, process stage sequence and material batch identification, grain size data, dislocation density data and precipitated phase distribution data are collected, these data are combined and classified according to the processing section number, process stage sequence and material batch identification, structural parameters are called and grouped according to section characteristics to obtain a structural grouping parameter set; S102: Based on the structure grouping parameter set, for the structure grouping parameter set, calculate the data parameter differences between different groups and construct data attribution rules using the formula: ； Calculate the stress difference coefficient between each group, adjust the data index order based on the regional classification standard, and obtain the stress matching index; in, Representative Group With grouping The stress difference coefficient between and Represent the average stress values ​​of the two groups, Represents the influencing factor of the precipitation phase within the group, where represents the distribution influence coefficient of the precipitated phase, Represents the influence weight of the precipitated phase, represents the total number of precipitated phases; S103: Based on the stress matching index, the data index order is adjusted to match the regional classification standard, the stress and morphology matching index between the structural data groups is formed, the overall microstructure parameter integrated value is calculated, and the microstructure parameter integrated value is obtained.

3. The sensor-based nonferrous metal forging quality monitoring method according to claim 1, characterized in that: The steps for obtaining the curvature anomaly concentration value are as follows: S301: calling the stress mapping response interval value, collecting edge point displacement sensor data and processing time series data, extracting displacement change values, and establishing a displacement sequence that changes with time, calculating the curvature change rate of continuous points, and obtaining a curvature change rate sequence; S302: Based on the curvature change rate sequence, the distance between adjacent inflection points is calculated and compared with the set inflection point distance range, and curvature change points that meet the characteristics are screened. The abnormal point set is selected according to the change rate trend, using the formula: ； Calculate the curvature change rate anomaly index of each point, filter out the anomaly points, and obtain the anomaly point set; in, represents the abnormal index of curvature change rate, Represents the displacement speed of the current adjacent point, Represents the displacement speed of adjacent points, Ti represents the current corresponding time point and Represents the corresponding time points of adjacent points, represents the cumulative effect of curvature mutation, where is the local curvature value, is the corresponding inflection point spacing, To calculate the number of sample points; S303: Based on the abnormal point set, combined with the stress increase and structural response value corresponding to the abnormal point, the curvature change frequency of each point is calculated, and the abnormal area is statistically analyzed to obtain the curvature abnormal concentration value.

4. The sensor-based nonferrous metal forging quality monitoring method according to claim 1, characterized in that: The steps for obtaining the stress growth focus interval value are: S401: Call the curvature abnormal concentration value, obtain the stress monitoring sensor data of the forging contour, extract the stress value at each time point, and construct a time series data set to calculate the stress growth ratio per unit time. Suppose a certain time point The corresponding stress value is , adjacent time points The corresponding stress value is , then the stress growth ratio per unit time is calculated as follows: ； Calculate the stress growth ratios at all time points and form a growth ratio sequence to obtain a stress growth ratio sequence; in, Represents the stress growth ratio per unit time, Representative time point The corresponding stress value at Represents the current time point The corresponding stress value at Representative time point , Represents the current time point , represents a time interval; S402: Based on the stress growth ratio sequence, a growth curve is constructed, and the growth ratio is compared with the segment difference baseline interval to screen abnormal growth segments. The trend change of the growth ratio is calculated according to the time series using the formula: ； Calculate the growth trend indicators for each time period, filter out the segments with continuous growth, and obtain the set of segments with continuous growth; in, Representative time period The stress growth trend value, is the stress growth ratio, is the section difference baseline, is the number of sampling points, represents the normalized calculation of the local stress variation amplitude, where is the local stress increase, is the corresponding time interval, is the number of statistical samples; S403: In combination with the stress mapping response interval value, a corresponding value superposition analysis is performed on the screened continuous growth segment set, and the change focus area is located, the stress increase trend in the area is calculated, and the abnormal area is focused and located to obtain the stress growth focus interval value.

5. Sensor-based nonferrous metal forging quality monitoring system, characterized in that: The system is used to execute the sensor-based nonferrous metal forging quality monitoring method according to any one of claims 1 to 4, comprising: The microstructure analysis module obtains grain size data, dislocation density data, and precipitate phase distribution data, classifies the data according to the processing section number, process stage sequence, and material batch identification, establishes a section feature index, calculates the interval difference between grain size data and dislocation density data, constructs attribution rules, adjusts the data index order, forms a matching relationship, calculates the stress and morphological parameter matching ratio, and obtains the integrated value of the microstructure parameters; The stress response matching module calls the integrated value of the microstructure parameters, matches the equivalent stress data and the dislocation density data, calculates the difference normalization, screens the amplification mutation point, and obtains the stress mapping response interval value; The curvature fluctuation screening module calls the stress mapping response interval value, analyzes the curvature change rate of the displacement data and the processing time series data, screens the mutation area, and obtains the curvature abnormal concentration value; The stress growth trend positioning module calls the curvature anomaly concentration value, analyzes the growth ratio of the stress monitoring data, screens the continuous growth area, and obtains the stress growth focus interval value; The defect area aggregation module calls the stress growth focus interval value, analyzes the change ratio of dislocation density data and structural trend data, screens the concentrated area, and obtains the defect area feature set.

Images