Intelligent calibration method and system for measuring range of magnetostrictive sensing pressure gauge

Abstract

The invention discloses a magnetostrictive sensing pressure gauge measuring range intelligent calibration method and system, and relates to the technical field of measuring range calibration, and the method comprises the steps: collecting a plurality of first elastic element sample sets of a first service life of a target pressure gauge model when the monitoring environment characteristics of the target pressure gauge model are not consistent with the standard, executing range boundary analysis through a magnetostrictive sensor to obtain a plurality of first range boundaries; configuring a first training data set by taking the plurality of first range boundaries as supervision and the plurality of first service lives as input; collecting a second elastic element sample set, and performing range boundary analysis through a magnetostrictive sensor to obtain a plurality of second range boundaries; training a first range boundary predictor by using machine learning; and performing neuron progressive increment training on the first range boundary predictor to generate a second range boundary predictor, and performing range intelligent calibration on the target pressure gauge. According to the invention, the technical problem of poor range calibration effect in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of range calibration technology, specifically to a method and system for intelligent range calibration of a pressure gauge using magnetostrictive sensing. Background Technology

[0002] Pressure measuring instruments play a crucial role in industrial process control, equipment monitoring, and safety assurance. The performance of their core components, such as elastic elements like Bourdon tubes, directly determines the instrument's measurement accuracy and range reliability. Magnetostrictive sensing technology, due to its advantages of non-contact operation, high precision, and strong anti-interference capabilities, has been widely applied in recent years to monitor the deformation of elastic elements and analyze the range of pressure gauges, providing a more advanced means for range calibration. Existing calibration methods typically rely on establishing range boundaries within a preset standard environment. However, pressure gauges are often exposed to complex and variable environmental conditions during actual service, such as temperature fluctuations, which can alter the performance of elastic elements. This change in range boundaries deviates from the preset range boundaries under the standard environment, resulting in inaccurate range calibration results. Summary of the Invention

[0003] This application provides a method and system for intelligent calibration of the range of a pressure gauge using a magnetostrictive sensor, which addresses the technical problem of inaccurate range boundary calibration in the prior art.

[0004] In view of the above problems, this application provides a method and system for intelligent calibration of pressure gauge range using magnetostrictive sensing.

[0005] In a first aspect, this application provides a method for intelligent calibration of the range of a pressure gauge using a magnetostrictive sensor, the method comprising:

[0006] When the monitoring environment characteristics of the target pressure gauge model are inconsistent with the standard environment characteristics, a sample set of first elastic elements of several first service life of the target pressure gauge model that meets the monitoring environment characteristics is collected. The range boundary analysis of the target pressure gauge model is performed through a magnetostrictive sensor to obtain several first range boundaries. The standard environment characteristics are the pre-stored ideal service environment of the elastic element, and the monitoring environment characteristics are the actual service environment of the elastic element.

[0007] The first training dataset is configured with several first range boundaries as supervision and several first service lives as input;

[0008] When the amount of data in the first training dataset is less than or equal to the training quantity threshold, a sample set of second elastic elements with a second service life that meets the standard environmental characteristics of the target pressure gauge model is collected. The range boundary analysis of the target pressure gauge model is performed through a magnetostrictive sensor to obtain a number of second range boundaries.

[0009] Using the second range boundary set as supervision and several second service lives as input, a second training dataset is configured, and machine learning is used to train the first range boundary predictor.

[0010] The first training dataset is retrieved to perform progressive incremental training of neurons on the first range boundary predictor, generating a second range boundary predictor, and performing range intelligent calibration on the target pressure gauge based on the service life monitoring of the elastic element.

[0011] Secondly, this application provides a magnetostrictive sensing-based intelligent calibration system for pressure gauge range, comprising:

[0012] The first range boundary acquisition module is used to collect a sample set of first elastic elements of the target pressure gauge model with a first service life that meet the monitoring environment characteristics of the target pressure gauge model when the monitoring environment characteristics of the target pressure gauge model are inconsistent with the standard environment characteristics. The module performs range boundary analysis on the target pressure gauge model through a magnetostrictive sensor to obtain a number of first range boundaries. The standard environment characteristics are the pre-stored ideal service environment of the elastic element, and the monitoring environment characteristics are the actual service environment of the elastic element.

[0013] The training data configuration module is used to configure the first training dataset with several first range boundaries as supervision and several first service lives as input.

[0014] The second range boundary acquisition module is used to collect a sample set of second elastic elements of a target pressure gauge model that meets the standard environmental characteristics when the amount of data in the first training dataset is less than or equal to the training quantity threshold. The module then performs range boundary analysis on the target pressure gauge model through a magnetostrictive sensor to obtain a number of second range boundaries.

[0015] The first range boundary prediction module is used to configure a second training dataset with a second range boundary set as supervision, several second service lives as input, and machine learning to train the first range boundary predictor.

[0016] The range intelligent calibration module is used to retrieve the first training dataset to perform progressive incremental training of neurons on the first range boundary predictor, generate a second range boundary predictor, and perform range intelligent calibration on the target pressure gauge based on the service life monitoring of the elastic element.

[0017] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0018] This application proposes an intelligent calibration method and system for pressure gauge range using magnetostrictive sensing. By integrating service environment feature identification, data-driven modeling, and progressive model transfer learning, it significantly improves the accuracy and reliability of pressure gauge range boundary calibration in complex and variable real-world service environments. Compared to traditional calibration methods that rely on preset standard environmental data, the technical solution provided in this application significantly overcomes the limitations of environmental differences on model applicability. It achieves high-precision, adaptive prediction, and reliable calibration of pressure gauge range boundaries even when environmental characteristics differ from standard environments, thereby improving the measurement accuracy of pressure gauges under complex operating conditions. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating an intelligent calibration method for the range of a pressure gauge using a magnetostrictive sensor, provided in an embodiment of this application.

[0021] Figure 2 This is a schematic diagram of a magnetostrictive sensing pressure gauge range intelligent calibration system provided in an embodiment of this application.

[0022] The components represented by each number in the attached diagram are explained below:

[0023] The system includes a first range boundary acquisition module 100, a training data configuration module 200, a second range boundary acquisition module 300, a first range boundary prediction module 400, and a range intelligent calibration module 500. Detailed Implementation

[0024] This application provides a method and system for intelligent calibration of the range of a pressure gauge using magnetostrictive sensing, which addresses the technical problem of inaccurate range boundary calibration in existing technologies.

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0026] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0027] Example 1, as Figure 1 As shown, this application provides a method for intelligent calibration of the range of a pressure gauge using a magnetostrictive sensor, wherein the method includes:

[0028] S10: When the monitoring environment characteristics of the target pressure gauge model are inconsistent with the standard environment characteristics, collect a sample set of first elastic elements of the target pressure gauge model that meet the monitoring environment characteristics and have a first service life. Perform range boundary analysis on the target pressure gauge model through a magnetostrictive sensor to obtain a number of first range boundaries. The standard environment characteristics are the pre-stored ideal service environment of the elastic element, and the monitoring environment characteristics are the actual service environment of the elastic element.

[0029] When the actual service environment of the target pressure gauge differs from the preset ideal standard environment, the elastic element is affected by environmental factors such as temperature, causing changes in its material properties and resulting in inaccuracies in the range boundary established based on the standard environment. Traditional methods cannot capture the true deformation pattern of the elastic element and its correlation with service life under specific non-standard environments.

[0030] Step S10 in the method provided in this application embodiment includes:

[0031] Based on the target pressure gauge model, a matching parts list is used to determine the material of the elastic element;

[0032] Using the material of the elastic element as a constraint, a dataset of elastic element deformation detection records is collected. Each set of elastic element deformation detection record data in the dataset includes storage environment record data, detection pressure record data, element size record data, pressure detection position record data, and elastic type variable record data.

[0033] Based on the pressure detection record data, the component size record data, and the pressure detection position record data, cluster analysis is performed on the elastic component deformation detection record dataset to obtain multi-cluster elastic component deformation detection record data.

[0034] Based on the deformation detection record data of the multi-cluster elastic element, the environmental attribute set is sorted by deformation correlation in combination with the storage environment record data and the elastic variable record data to obtain the selected environmental attribute set;

[0035] Specifically, from the deformation detection record data of the multiple clusters of elastic elements, the first storage environment record dataset and the first elastic type variable record dataset corresponding one-to-one in the deformation detection record data of the first cluster of elastic elements are extracted, including:

[0036] Based on the first storage environment record dataset and the first elastic variable record dataset, pairwise same attribute deviation calculations are performed on the deformation detection record data of the first cluster of elastic elements to obtain one-to-one corresponding first attribute storage environment deviation modulus calculation data up to the Nth attribute storage environment deviation modulus calculation data, as well as elastic variable deviation calculation data.

[0037] Based on the elastic variable deviation calculation data, grey relational analysis is performed on the first attribute storage environment deviation modulus calculation data up to the Nth attribute storage environment deviation modulus calculation data to obtain the first cluster first attribute correlation degree up to the first cluster Nth attribute correlation degree.

[0038] Until the correlation degree of the first attribute of the Mth cluster is obtained, up to the correlation degree of the Nth attribute of the Mth cluster;

[0039] Based on the correlation degree of the first attribute of the first cluster up to the correlation degree of the Nth attribute of the first cluster, up to the correlation degree of the first attribute of the Mth cluster up to the correlation degree of the Nth attribute of the Mth cluster, the mean of the same attribute is calculated to obtain the comprehensive correlation degree of the first attribute up to the comprehensive correlation degree of the Nth attribute.

[0040] Based on the comprehensive correlation degree of the first attribute up to the comprehensive correlation degree of the Nth attribute, environmental attributes with a comprehensive correlation degree greater than or equal to the correlation degree threshold are extracted and added to the selected environmental attribute set.

[0041] Based on the selected set of environmental attributes, configure the standard environmental features and collect the monitoring environmental features;

[0042] If any attribute of the monitored environmental feature differs from any attribute of the standard environmental feature, the monitored environmental feature is considered inconsistent with the standard environmental feature; otherwise, the monitored environmental feature is considered consistent with the standard environmental feature.

[0043] Extract the first service life from several first service lifespans;

[0044] Based on the first service life, extract the first service life elastic element sample set from the first elastic element sample set;

[0045] Traverse the first service life elastic element sample set, and perform range boundary analysis on the target pressure gauge model using a magnetostrictive sensor to obtain the range boundary set;

[0046] Specifically, by traversing the first service life elastic element sample set, and using a magnetostrictive sensor to perform range boundary analysis on the target pressure gauge model, a range boundary set is obtained, including:

[0047] Extract the first service life elastic element sample from the first service life elastic element sample set;

[0048] Obtain the rated range of the first service life elastic element sample, wherein the rated range is the factory range of the target pressure gauge;

[0049] Specifically, based on the rated range of the first service life elastic element sample, an upper range boundary analysis is performed on the first service life elastic element sample using a magnetostrictive sensor to obtain the upper range boundary of the first service life elastic element sample. Prior to this, the process includes:

[0050] For the first service life elastic element sample, multiple probes are made at the working pressure position using the rated range median pressure, rated range quarter pressure, and rated range third pressure. The magnetostrictive sensor is used to monitor the deformation of multiple median pressure probes, multiple quarter pressure probes, and multiple third pressure probes, wherein the number of probes is greater than or equal to 5.

[0051] Using the rated range midpoint pressure, the rated range quarter-point pressure, and the rated range third-quarter pressure as constraints, respectively, qualified quality test data of the first elastic element, qualified quality test data of the second elastic element, and qualified quality test data of the third elastic element of the target pressure gauge model with a service life less than or equal to the service life threshold are collected.

[0052] The qualified quality test data of the first elastic element, the qualified quality test data of the second elastic element, and the qualified quality test data of the third elastic element are respectively evaluated by the central value of elastic element type variables to obtain the median pressure detection standard shape variable, the quarter-position pressure detection standard shape variable, and the third-quarter pressure detection standard shape variable.

[0053] The percentage of first abnormal deformation variables whose deviation from the first deformation variable of the median pressure detection standard deformation variable is greater than or equal to the deformation deviation threshold is calculated among the multiple median pressure detection deformation variables.

[0054] The percentage of second abnormal deformation variables whose second deformation deviation from the standard deformation variable of the quarter-position pressure detection is greater than or equal to the deformation deviation threshold is calculated among the multiple quarter-position pressure detection deformation variables.

[0055] The percentage of the third abnormal deformation variables that deviate from the standard third-quarter pressure detection deformation variables by a factor greater than or equal to the deformation deviation threshold is calculated.

[0056] When any one of the proportions of the first abnormal deformation, the second abnormal deformation, and the third abnormal deformation is greater than or equal to the abnormal deformation proportion threshold, a range failure prompt is generated and sent to the user terminal, wherein 0.15 ≤ abnormal deformation proportion threshold ≤ 0.3;

[0057] When the proportions of the first abnormal deformation, the second abnormal deformation, and the third abnormal deformation are all less than the threshold of the abnormal deformation, the upper boundary of the range of the first service life elastic element sample is analyzed by a magnetostrictive sensor based on the rated range of the first service life elastic element sample to obtain the upper boundary of the range of the first service life elastic element sample.

[0058] Wherein, when the proportions of the first abnormal deformation, the second abnormal deformation, and the third abnormal deformation are all less than the threshold for the proportion of abnormal deformation, based on the rated range of the first service life elastic element sample, an upper range boundary analysis is performed on the first service life elastic element sample using a magnetostrictive sensor to obtain the upper range boundary of the first service life elastic element sample, including:

[0059] The magnetostrictive sensor is used to detect the percentage of the fourth abnormal deformation at the upper limit pressure of the rated range of the first service life elastic element sample.

[0060] When the proportion of the fourth abnormal deformation is less than the threshold of the abnormal deformation, the upper limit pressure of the rated range is set as the upper boundary of the sample range of the first service life elastic element.

[0061] When the proportion of the fourth abnormal deformation is greater than or equal to the threshold of the abnormal deformation, the pressure consistency deviation threshold is subtracted from the upper limit pressure of the rated range to obtain the updated upper limit pressure of the rated range, and cyclic analysis is performed.

[0062] Based on the rated range of the first service life elastic element sample, the upper boundary of the range of the first service life elastic element sample is analyzed by a magnetostrictive sensor to obtain the upper boundary of the range of the first service life elastic element sample.

[0063] Based on the rated range of the first service life elastic element sample, the lower boundary of the range of the first service life elastic element sample is analyzed by using a magnetostrictive sensor to obtain the lower boundary of the range of the first service life elastic element sample.

[0064] Based on the upper boundary and lower boundary of the first service life elastic element sample range, the first service life elastic element sample range boundary is constructed and added to the range boundary set.

[0065] The upper boundary concentration value is evaluated on the set of range boundaries to obtain the upper boundary of the range, and the lower boundary concentration value is evaluated on the set of range boundaries to obtain the lower boundary of the range.

[0066] Based on the upper and lower range boundaries, a first range boundary is constructed, and the plurality of first range boundaries are added.

[0067] In this embodiment of the application, based on the model of the target pressure gauge, a matching parts list is used to obtain the material of the elastic element, such as a Bourdon tube made of copper-based alloy or a Bourdon tube made of iron-based alloy.

[0068] Using the material of the elastic element as a constraint, a dataset of elastic element deformation detection records is collected. Each set of elastic element deformation detection records in this dataset includes: storage environment data such as temperature (°C); detection pressure data (MPa); element size data such as Bourdon tube wall thickness (mm); pressure detection position data such as detection position data obtained using GPS positioning; and elastic variable data such as the displacement distance of the moving end of the Bourdon tube relative to its original position (mm).

[0069] Based on the detection pressure record data, component size record data, and pressure detection position record data, the K-means clustering algorithm is used to perform cluster analysis on the elastic component deformation detection record dataset to obtain multiple clusters of elastic component deformation detection record data. Each cluster of elastic component deformation detection record data has similar mechanical characteristics, and the pressure detection position, shape, and detection pressure magnitude and direction of the elastic components in any cluster are consistent.

[0070] Based on deformation detection records of multiple clusters of elastic elements, combined with storage environment records and elastic variable records, deformation correlation sorting is performed on the environmental attribute set to obtain the selected environmental attribute set.

[0071] Specifically, from the deformation detection record data of multiple clusters of elastic elements, the deformation detection record data of the first cluster of elastic elements is randomly extracted as the current processing object, and the first storage environment record dataset and the first elastic variable record dataset corresponding to the deformation detection record data of the first cluster of elastic elements are extracted one by one.

[0072] Based on the first storage environment record dataset and the first elastic variable record dataset, pairwise same attribute deviation calculations are performed on the deformation detection record data of the first cluster of elastic elements. For example, the absolute value of the difference between the two detection pressures in the deformation detection record data of the first cluster of elastic elements is calculated to obtain the first attribute storage environment deviation modulus data corresponding to each other. The calculation is repeated until the Nth attribute storage environment deviation modulus calculation data and the elastic variable deviation calculation data are obtained.

[0073] Based on the calculated data of the elastic variable deviation, grey relational analysis is performed on the calculated data of the storage environment deviation of the first attribute up to the calculated data of the storage environment deviation of the Nth attribute to obtain the correlation degree of the first cluster of the first attribute up to the correlation degree of the Nth attribute. For example, the calculated data of the elastic variable deviation is dimensionless: Dimensionless elastic variable deviation data = Dimensionless variable deviation data ÷ Mean of elastic variable deviation data. Using the same calculation method, the calculated data of the storage environment deviation of the first attribute up to the calculated data of the storage environment deviation is also dimensionless: Dimensionless storage environment deviation data = Dimensionless storage environment deviation data ÷ Mean of storage environment deviation data. The dimensionless first attribute storage environment deviation data is integrated to obtain the first alignment sequence. Using the same calculation approach, the first alignment sequence to the Nth alignment sequence are obtained.

[0074] The grey relational analysis algorithm is used to calculate the correlation between the deviation of the variable and the deviation of each environmental attribute, and to obtain the correlation between the first attribute of the first cluster and the Nth attribute of the first cluster.

[0075] The same method is used to calculate until the correlation degree of the first attribute of the Mth cluster is obtained up to the correlation degree of the Nth attribute of the Mth cluster.

[0076] Based on the correlation degree of the first attribute in the first cluster up to the correlation degree of the Nth attribute in the first cluster, up to the correlation degree of the first attribute in the Mth cluster up to the correlation degree of the Nth attribute in the Mth cluster, the mean of the correlation degree of the same attribute is calculated, and the mean of the correlation degree of the first attribute in each cluster is obtained to obtain the comprehensive correlation degree of the first attribute up to the comprehensive correlation degree of the Nth attribute. For example, the comprehensive correlation degree of the first attribute = ∑(correlation degree of the first attribute in the first cluster to the correlation degree of the first attribute in the Mth cluster) ÷ M.

[0077] Based on the comprehensive correlation degree of the first attribute up to the comprehensive correlation degree of the Nth attribute, environmental attributes with a comprehensive correlation degree greater than or equal to the correlation degree threshold are extracted and added to the selected environmental attribute set. The correlation degree threshold is a pre-set threshold representing the magnitude of the influence of an attribute on the deviation of the elastic element deformation, which can be set to 0.6 for example. Environmental attributes with a comprehensive correlation degree greater than or equal to 0.6 are added to the selected environmental attribute set. The selected environmental attribute set consists of environmental attributes that have a significant impact on the elastic element deformation. Considering these attributes helps to reduce environmental influences and make the measurement more accurate.

[0078] Based on the selected set of environmental attributes, standard environmental features are configured, and environmental features are collected and monitored. The standard environmental features are the standard values ​​of all attributes in the selected set of environmental attributes. The standard values ​​are derived from the standard environmental attribute values ​​that are preset during the design process of the target pressure gauge. For example, if an environmental attribute is temperature, the standard value is 30℃.

[0079] If any attribute of the monitored environmental characteristics differs from that of the standard environmental characteristics, it is considered that the monitored environmental characteristics are inconsistent with the standard environmental characteristics; otherwise, it is considered that the monitored environmental characteristics are consistent with the standard environmental characteristics.

[0080] One service life is randomly selected from the service life of several pressure gauges as the first service life, with the unit being days.

[0081] Based on the first service life, extract the first service life elastic element sample set from the first elastic element sample set.

[0082] Extract the first service life elastic element sample from the first service life elastic element sample set.

[0083] Obtain the rated range of the first service life elastic element sample, where the rated range is the factory range of the target pressure gauge.

[0084] For the first service life elastic element sample, the rated range median pressure, rated range quarter-position pressure, and rated range third-quarter pressure were used to conduct multiple probes at the working pressure position, i.e. the actual working pressure position. Magnetostrictive sensors were used to monitor the probe deformation of multiple median pressure probes, multiple quarter-position pressure probes, and multiple third-quarter pressure probes. The number of probes was greater than or equal to 5 to ensure the accuracy of the probes.

[0085] Using the midpoint of the rated range, one-quarter of the rated range, and three-quarters of the rated range as constraints, qualified quality test data for the first, second, and third elastic elements of the target pressure gauge model with a service life less than or equal to a service life threshold were collected. Specifically, the qualified quality test data for the first elastic element was measured using the midpoint of the rated range and the service life of the elastic element of the target pressure gauge was less than or equal to the service life threshold; the qualified quality test data for the second elastic element was measured using one-quarter of the rated range and the service life of the elastic element of the target pressure gauge was less than or equal to the service life threshold; and the qualified quality test data for the third elastic element was measured using three-quarters of the rated range and the service life of the elastic element of the target pressure gauge was less than or equal to the service life threshold.

[0086] The central tendency values ​​of elastic element-type variables were evaluated based on the qualified quality test data of the first, second, and third elastic elements, respectively, to obtain the median pressure detection standard shape variable, the quarter-th percentile pressure detection standard shape variable, and the third-quarter-th percentile pressure detection standard shape variable. Specifically, the mean of the qualified quality test data of the first elastic element was calculated as the median pressure detection standard shape variable; the mean of the qualified quality test data of the second elastic element was calculated as the quarter-th percentile pressure detection standard shape variable; and the mean of the qualified quality test data of the third elastic element was calculated as the third-quarter-th percentile pressure detection standard shape variable.

[0087] The percentage of first abnormal deformation variables among multiple median pressure detection deformation variables whose first deformation variable deviation from the median pressure detection standard deformation variable is greater than or equal to the deformation deviation threshold is calculated. The first deformation variable deviation is calculated as |median pressure detection deformation variable - median pressure detection standard deformation variable|, and the percentage of first abnormal deformation variables is calculated as the median pressure detection deformation variables whose first deformation variable deviation is greater than or equal to the deformation deviation threshold ÷ the total median pressure detection deformation variables.

[0088] Among multiple quarter-position pressure detection deformation variables, the proportion of second abnormal deformation variables whose deviation from the standard quarter-position pressure detection deformation variable is greater than or equal to the deformation deviation threshold is statistically analyzed. The proportion of second abnormal deformation variables is calculated using the same method as that used for the proportion of first abnormal deformation variables.

[0089] The proportion of third-abnormal deformation variables among the three-quarters pressure detection deformation variables, where the deviation from the third-quarters pressure detection standard deformation variable is greater than or equal to the deformation deviation threshold, is calculated using the same method as for the proportion of first-abnormal deformation variables.

[0090] When any one of the percentages of the first, second, and third abnormal variables is greater than or equal to the abnormal variable percentage threshold, a range failure warning is generated and sent to the user terminal, wherein 0.15 ≤ abnormal variable percentage threshold ≤ 0.3.

[0091] When the proportions of the first, second, and third abnormal deformation variables are all less than the threshold for the proportion of abnormal deformation variables, the upper boundary of the range of the elastic element sample with the first service life is analyzed by using a magnetostrictive sensor based on the rated range of the elastic element sample with the first service life, and the upper boundary of the range of the elastic element sample with the first service life is obtained.

[0092] Specifically, the percentage of the fourth abnormal deformation of the rated range upper limit pressure of the elastic element sample in the first service life is detected by a magnetostrictive sensor.

[0093] When the proportion of the fourth abnormal deformation is less than the threshold for the proportion of abnormal deformation, the upper limit pressure of the rated range is set as the upper boundary of the sample range of the elastic element in the first service life. Where 0.15 ≤ threshold for the proportion of abnormal deformation ≤ 0.3.

[0094] When the proportion of the fourth abnormal deformation is greater than or equal to the abnormal deformation proportion threshold, the updated rated range upper limit pressure is obtained by subtracting the pressure consistency deviation threshold from the rated range upper limit pressure. The updated rated range upper limit pressure = rated range upper limit pressure - pressure consistency deviation threshold. This iterative analysis calculation is performed to finally obtain the updated rated range upper limit pressure, which serves as the upper boundary of the sample range for the elastic element in the first service life.

[0095] Using the same approach as obtaining the upper boundary of the range of the first service life elastic element sample, the lower boundary of the range of the first service life elastic element sample is obtained. Specifically, a magnetostrictive sensor is used to detect the proportion of the fifth abnormal deformation at the lower limit pressure of the rated range of the first service life elastic element sample. When the proportion of the fifth abnormal deformation is less than the abnormal deformation proportion threshold, the lower limit of the rated range is set as the lower boundary of the range of the first service life elastic element sample. Where 0.15 ≤ abnormal deformation proportion threshold ≤ 0.3.

[0096] When the proportion of the fifth abnormal deformation is greater than or equal to the abnormal deformation proportion threshold, the updated rated range lower limit pressure is obtained by adding the pressure consistency deviation threshold to the rated range lower limit pressure. The updated rated range lower limit pressure = rated range lower limit pressure + pressure consistency deviation threshold. This iterative analysis calculation is performed to finally obtain the updated rated range lower limit pressure, which serves as the lower boundary of the sample range for the elastic element in the first service life.

[0097] Based on the upper boundary and lower boundary of the first service life elastic element sample range, construct the first service life elastic element sample range boundary and add it to the range boundary set.

[0098] The upper boundary of the range is obtained by evaluating the upper boundary set of the range boundary values, and the lower boundary of the range is obtained by evaluating the lower boundary set of the range boundary values. For example, the arithmetic mean of the upper boundary of the range of the elastic element sample in the first service life is obtained as the upper boundary of the range, and the arithmetic mean of the lower boundary of the range of the elastic element sample in the first service life is obtained as the lower boundary of the range.

[0099] Based on the upper and lower boundaries of the measurement range, a first range boundary is constructed, and several other first range boundaries are added. These other first range boundaries are obtained using the same method as the first range boundary.

[0100] By collecting service samples of elastic elements that perfectly match the actual monitoring environment and employing magnetostrictive sensors to quantify the range boundary degradation trajectory within their lifespan, this application constructs a realistic range boundary degradation dataset. This overcomes the model inaccuracy problem caused by environmental differences. The obtained first range boundary accurately reflects the performance degradation law of the target pressure gauge's elastic element under actual complex working conditions, providing a high-fidelity environmentally adapted data source for subsequent model training. Compared to the traditional approach that relies solely on standard environmental data, this significantly improves the environmental representativeness of the data, fundamentally ensuring the environmental adaptability of the subsequent prediction model.

[0101] S20: Configure the first training dataset with several first range boundaries as supervision and several first service lives as input.

[0102] In this embodiment of the application, a first training dataset is obtained by integrating several first range boundaries as supervision and several first service lives as input, providing a learnable and explicit mapping target for the machine learning model.

[0103] S30: When the amount of data in the first training dataset is less than or equal to the training quantity threshold, collect a sample set of second elastic elements of the target pressure gauge model that meets the standard environmental characteristics for the second service life. Perform range boundary analysis on the target pressure gauge model through a magnetostrictive sensor to obtain a number of second range boundaries.

[0104] In real-world, complex environments, the availability of qualified service samples can be scarce, directly leading to insufficient accuracy in models trained on real-world data due to sample limitations. Traditional methods often employ standard environment models in such cases, exacerbating prediction bias.

[0105] In this embodiment, when the amount of data in the first training dataset is less than or equal to the training quantity threshold, a sample set of second elastic elements with a second service life that meets the standard environmental characteristics of the target pressure gauge model is collected. Preferably, a more diverse range of service lives, such as those closer to the service life limit or close to 0, can be selected as the second service life to cover the entire life cycle of the elastic element as much as possible. Using a magnetostrictive sensor and employing the same method as described above, range boundary analysis is performed on the target pressure gauge model to obtain a number of second range boundaries.

[0106] By introducing high-abundance sample data from standard environments, a cross-environmental data supplement was constructed. Although the obtained second-range boundary cannot directly reflect the degradation patterns in actual environments, it carries the baseline of material performance and basic degradation modes of elastic elements under ideal conditions. While retaining the core value of actual environmental data, it significantly expands the information sources that can be used for modeling, providing a general knowledge base for subsequent fusion training and effectively mitigating the risk of model overfitting caused by small sample sizes.

[0107] S40: Using the second range boundary set as supervision and several second service lives as input, configure the second training dataset and use machine learning to train the first range boundary predictor.

[0108] Standard environmental data and real-world environmental data have different distributions, and directly mixing them for training will cause the model to learn contradictory patterns. Traditional single models cannot distinguish between environmental specificity and common characteristics.

[0109] In this embodiment, machine learning is employed to construct a first range boundary predictor. For example, a random forest regression model is used to construct the first range boundary predictor, with 100 trees and a maximum depth of 10. A second range boundary set is used as supervision, and several second service lives are used as inputs, configured with a second training dataset. The constructed first range boundary predictor is trained using the second training dataset until convergence. For example, if the second service life is input and the output range boundary error is within ±0.1 MPa, the training of the first range boundary predictor is considered complete.

[0110] By isolating and training standard environment data, a benchmark predictor with generalization capabilities was constructed. The first range boundary predictor focuses on learning the basic degradation characteristics of elastic elements that are unaffected by environmental disturbances, forming a knowledge kernel that is universal across environments, providing a highly robust initialization model for subsequent adaptation and optimization.

[0111] S50: Retrieve the first training dataset to perform progressive incremental training of neurons on the first range boundary predictor, generate a second range boundary predictor, and perform range intelligent calibration on the target pressure gauge based on the service life monitoring of the elastic element.

[0112] When real-world samples are scarce, directly training a new model or fine-tuning all parameters can lead to overfitting. Traditional transfer learning methods, due to the one-time injection of small sample data, are prone to model oscillation.

[0113] Step S50 in the method provided in this application embodiment includes:

[0114] Freeze the network parameters of the first range boundary predictor, configure the first fully connected neural network, use the output of the first range boundary predictor as the input of the first fully connected neural network, use the plurality of first range boundaries as the supervision of the first fully connected neural network, and train the first asymptotic range boundary predictor.

[0115] When the first verification accuracy of the first asymptotic range boundary predictor is greater than or equal to the convergence accuracy, the first asymptotic range boundary predictor is set as the second range boundary predictor.

[0116] Otherwise, freeze the network parameters of the first asymptotic range boundary predictor, configure the second fully connected neural network, use the output of the first asymptotic range boundary predictor as the input of the second fully connected neural network, use the plurality of first range boundaries as the supervision of the second fully connected neural network, and train the second asymptotic range boundary predictor to perform loop analysis.

[0117] The Qth asymptotic range boundary predictor is set as the second range boundary predictor until the Qth verification accuracy of the Qth asymptotic range boundary predictor is greater than or equal to the convergence accuracy.

[0118] In this embodiment, the network parameters of the first range boundary predictor are frozen, and a first fully connected neural network is configured. For example, a three-layer structure is adopted: the input layer receives the range boundary output by the first range boundary predictor, the fully connected layer uses 32 nodes and is activated using the ReLU function, and the output layer outputs the first range boundary. The output of the first range boundary predictor is used as the input of the constructed first fully connected neural network, and several first range boundaries are used as supervision for the constructed first fully connected neural network to train the first asymptotic range boundary predictor.

[0119] When the first verification accuracy of the first asymptotic range boundary predictor is greater than or equal to the convergence accuracy, the first asymptotic range boundary predictor is set as the second range boundary predictor. For example, if the convergence accuracy is set to 90%, the first verification accuracy = 1 - |range boundary output by the first asymptotic range boundary predictor - first range boundary| ÷ first range boundary. When the first verification accuracy is greater than or equal to 90%, the first asymptotic range boundary predictor is set as the second range boundary predictor.

[0120] Otherwise, freeze the network parameters of the first asymptotic range boundary predictor, configure the second fully connected neural network with the same structure as the first fully connected neural network, use the output of the first asymptotic range boundary predictor as the input of the second fully connected neural network, use several first range boundaries as supervision of the second fully connected neural network, train the second asymptotic range boundary predictor, and perform loop analysis.

[0121] Until the Qth verification accuracy of the Qth asymptotic range boundary predictor is greater than or equal to the convergence accuracy, the Qth asymptotic range boundary predictor is set as the second range boundary predictor, and intelligent range calibration is performed on the target pressure gauge based on the service monitoring life of the elastic element. Specifically, the service monitoring life of the elastic element of the target pressure gauge is input into the second range boundary predictor, and the output range boundary is used as the range boundary of the target pressure gauge for calibration.

[0122] Through incremental training of neurons, accurate knowledge fusion under small sample sizes was achieved. With most parameters of the base model frozen, it was gradually fine-tuned using real-world environmental samples. This ensured that the generated second range boundary predictor retained general knowledge under standard conditions while accurately absorbing the specific patterns of real-world environments. While minimizing the risk of overfitting, the model adaptively adjusted to environmental differences. Ultimately, by inputting the service life, it outputs a range boundary adapted to the current environment, completing intelligent calibration.

[0123] Example 2, as Figure 2 As shown, based on the same inventive concept as the intelligent calibration method for the range of a magnetostrictive pressure gauge provided in Embodiment 1, this embodiment of the invention also provides an intelligent calibration system for the range of a magnetostrictive pressure gauge, comprising:

[0124] The first range boundary acquisition module 100 is used to collect a sample set of first elastic elements of the target pressure gauge model with a first service life that meets the monitoring environment characteristics when the monitoring environment characteristics of the target pressure gauge model are inconsistent with the standard environment characteristics. The module performs range boundary analysis on the target pressure gauge model through a magnetostrictive sensor to obtain a number of first range boundaries. The standard environment characteristics are the pre-stored ideal service environment of the elastic element, and the monitoring environment characteristics are the actual service environment of the elastic element.

[0125] The training data configuration module 200 is used to configure the first training dataset with several first range boundaries as supervision and several first service lives as input.

[0126] The second range boundary acquisition module 300 is used to collect a sample set of second elastic elements of a target pressure gauge model that meets the standard environmental characteristics when the amount of data in the first training dataset is less than or equal to the training quantity threshold. The module then performs range boundary analysis on the target pressure gauge model through a magnetostrictive sensor to obtain a number of second range boundaries.

[0127] The first range boundary prediction module 400 is used to train the first range boundary predictor by using the second range boundary set as supervision, several second service lives as inputs, configuring the second training dataset, and using machine learning.

[0128] The range intelligent calibration module 500 is used to retrieve the first training dataset to perform progressive incremental training of neurons on the first range boundary predictor, generate a second range boundary predictor, and perform range intelligent calibration on the target pressure gauge based on the service life monitoring of the elastic element.

[0129] In one embodiment, the first range boundary acquisition module 100 is further configured to:

[0130] Based on the target pressure gauge model, a matching parts list is used to determine the material of the elastic element;

[0131] Using the material of the elastic element as a constraint, a dataset of elastic element deformation detection records is collected. Each set of elastic element deformation detection record data in the dataset includes storage environment record data, detection pressure record data, element size record data, pressure detection position record data, and elastic type variable record data.

[0132] Based on the pressure detection record data, the component size record data, and the pressure detection position record data, cluster analysis is performed on the elastic component deformation detection record dataset to obtain multi-cluster elastic component deformation detection record data.

[0133] Based on the deformation detection record data of the multi-cluster elastic element, the environmental attribute set is sorted by deformation correlation in combination with the storage environment record data and the elastic variable record data to obtain the selected environmental attribute set;

[0134] Specifically, from the deformation detection record data of the multiple clusters of elastic elements, the first storage environment record dataset and the first elastic type variable record dataset corresponding one-to-one in the deformation detection record data of the first cluster of elastic elements are extracted, including:

[0135] Based on the first storage environment record dataset and the first elastic variable record dataset, pairwise same attribute deviation calculations are performed on the deformation detection record data of the first cluster of elastic elements to obtain one-to-one corresponding first attribute storage environment deviation modulus calculation data up to the Nth attribute storage environment deviation modulus calculation data, as well as elastic variable deviation calculation data.

[0136] Based on the elastic variable deviation calculation data, grey relational analysis is performed on the first attribute storage environment deviation modulus calculation data up to the Nth attribute storage environment deviation modulus calculation data to obtain the first cluster first attribute correlation degree up to the first cluster Nth attribute correlation degree.

[0137] Until the correlation degree of the first attribute of the Mth cluster is obtained, up to the correlation degree of the Nth attribute of the Mth cluster;

[0138] Based on the correlation degree of the first attribute of the first cluster up to the correlation degree of the Nth attribute of the first cluster, up to the correlation degree of the first attribute of the Mth cluster up to the correlation degree of the Nth attribute of the Mth cluster, the mean of the same attribute is calculated to obtain the comprehensive correlation degree of the first attribute up to the comprehensive correlation degree of the Nth attribute.

[0139] Based on the comprehensive correlation degree of the first attribute up to the comprehensive correlation degree of the Nth attribute, environmental attributes with a comprehensive correlation degree greater than or equal to the correlation degree threshold are extracted and added to the selected environmental attribute set.

[0140] Based on the selected set of environmental attributes, configure the standard environmental features and collect the monitoring environmental features;

[0141] If any attribute of the monitored environmental feature differs from any attribute of the standard environmental feature, the monitored environmental feature is considered inconsistent with the standard environmental feature; otherwise, the monitored environmental feature is considered consistent with the standard environmental feature.

[0142] Extract the first service life from several first service lifespans;

[0143] Based on the first service life, extract the first service life elastic element sample set from the first elastic element sample set;

[0144] Traverse the first service life elastic element sample set, and perform range boundary analysis on the target pressure gauge model using a magnetostrictive sensor to obtain the range boundary set;

[0145] Specifically, by traversing the first service life elastic element sample set, and using a magnetostrictive sensor to perform range boundary analysis on the target pressure gauge model, a range boundary set is obtained, including:

[0146] Extract the first service life elastic element sample from the first service life elastic element sample set;

[0147] Obtain the rated range of the first service life elastic element sample, wherein the rated range is the factory range of the target pressure gauge;

[0148] Specifically, based on the rated range of the first service life elastic element sample, an upper range boundary analysis is performed on the first service life elastic element sample using a magnetostrictive sensor to obtain the upper range boundary of the first service life elastic element sample. Prior to this, the process includes:

[0149] For the first service life elastic element sample, multiple probes are made at the working pressure position using the rated range median pressure, rated range quarter pressure, and rated range third pressure. The magnetostrictive sensor is used to monitor the deformation of multiple median pressure probes, multiple quarter pressure probes, and multiple third pressure probes, wherein the number of probes is greater than or equal to 5.

[0150] Using the rated range midpoint pressure, the rated range quarter-point pressure, and the rated range third-quarter pressure as constraints, respectively, qualified quality test data of the first elastic element, qualified quality test data of the second elastic element, and qualified quality test data of the third elastic element of the target pressure gauge model with a service life less than or equal to the service life threshold are collected.

[0151] The qualified quality test data of the first elastic element, the qualified quality test data of the second elastic element, and the qualified quality test data of the third elastic element are respectively evaluated by the central value of elastic element type variables to obtain the median pressure detection standard shape variable, the quarter-position pressure detection standard shape variable, and the third-quarter pressure detection standard shape variable.

[0152] The percentage of first abnormal deformation variables whose deviation from the first deformation variable of the median pressure detection standard deformation variable is greater than or equal to the deformation deviation threshold is calculated among the multiple median pressure detection deformation variables.

[0153] The percentage of second abnormal deformation variables whose second deformation deviation from the standard deformation variable of the quarter-position pressure detection is greater than or equal to the deformation deviation threshold is calculated among the multiple quarter-position pressure detection deformation variables.

[0154] The percentage of the third abnormal deformation variables that deviate from the standard third-quarter pressure detection deformation variables by a factor greater than or equal to the deformation deviation threshold is calculated.

[0155] When any one of the proportions of the first abnormal deformation, the second abnormal deformation, and the third abnormal deformation is greater than or equal to the abnormal deformation proportion threshold, a range failure prompt is generated and sent to the user terminal, wherein 0.15 ≤ abnormal deformation proportion threshold ≤ 0.3;

[0156] When the proportions of the first abnormal deformation, the second abnormal deformation, and the third abnormal deformation are all less than the threshold of the abnormal deformation, the upper boundary of the range of the first service life elastic element sample is analyzed by a magnetostrictive sensor based on the rated range of the first service life elastic element sample to obtain the upper boundary of the range of the first service life elastic element sample.

[0157] Wherein, when the proportions of the first abnormal deformation, the second abnormal deformation, and the third abnormal deformation are all less than the threshold for the proportion of abnormal deformation, based on the rated range of the first service life elastic element sample, an upper range boundary analysis is performed on the first service life elastic element sample using a magnetostrictive sensor to obtain the upper range boundary of the first service life elastic element sample, including:

[0158] The magnetostrictive sensor is used to detect the percentage of the fourth abnormal deformation at the upper limit pressure of the rated range of the first service life elastic element sample.

[0159] When the proportion of the fourth abnormal deformation is less than the threshold of the abnormal deformation, the upper limit pressure of the rated range is set as the upper boundary of the sample range of the first service life elastic element.

[0160] When the proportion of the fourth abnormal deformation is greater than or equal to the threshold of the abnormal deformation, the pressure consistency deviation threshold is subtracted from the upper limit pressure of the rated range to obtain the updated upper limit pressure of the rated range, and cyclic analysis is performed.

[0161] Based on the rated range of the first service life elastic element sample, the upper boundary of the range of the first service life elastic element sample is analyzed by a magnetostrictive sensor to obtain the upper boundary of the range of the first service life elastic element sample.

[0162] Based on the rated range of the first service life elastic element sample, the lower boundary of the range of the first service life elastic element sample is analyzed by using a magnetostrictive sensor to obtain the lower boundary of the range of the first service life elastic element sample.

[0163] Based on the upper boundary and lower boundary of the first service life elastic element sample range, the first service life elastic element sample range boundary is constructed and added to the range boundary set.

[0164] The upper boundary concentration value is evaluated on the set of range boundaries to obtain the upper boundary of the range, and the lower boundary concentration value is evaluated on the set of range boundaries to obtain the lower boundary of the range.

[0165] Based on the upper and lower range boundaries, a first range boundary is constructed, and the plurality of first range boundaries are added.

[0166] In one embodiment, the range intelligent calibration module 500 is further used for:

[0167] Freeze the network parameters of the first range boundary predictor, configure the first fully connected neural network, use the output of the first range boundary predictor as the input of the first fully connected neural network, use the plurality of first range boundaries as the supervision of the first fully connected neural network, and train the first asymptotic range boundary predictor.

[0168] When the first verification accuracy of the first asymptotic range boundary predictor is greater than or equal to the convergence accuracy, the first asymptotic range boundary predictor is set as the second range boundary predictor.

[0169] Otherwise, freeze the network parameters of the first asymptotic range boundary predictor, configure the second fully connected neural network, use the output of the first asymptotic range boundary predictor as the input of the second fully connected neural network, use the plurality of first range boundaries as the supervision of the second fully connected neural network, and train the second asymptotic range boundary predictor to perform loop analysis.

[0170] The Qth asymptotic range boundary predictor is set as the second range boundary predictor until the Qth verification accuracy of the Qth asymptotic range boundary predictor is greater than or equal to the convergence accuracy.

[0171] In summary, the embodiments of this application have at least the following technical effects:

[0172] This application proposes a method and system for intelligent calibration of pressure gauge range using magnetostrictive sensing. By integrating service environment feature identification, data-driven modeling, and progressive model transfer learning, it significantly improves the accuracy and reliability of pressure gauge range boundary calibration in complex and variable real-world service environments. Specifically, it first identifies the differences between the actual monitoring environment and the standard environment of the target pressure gauge, and collects service life sample data of the elastic element under the same environment accordingly. When the amount of actual sample data is insufficient, sample data from the standard environment is introduced to train the basic prediction model. Most importantly, through a progressive incremental training strategy for neurons, the deformation-life-range boundary mapping relationship contained in the limited but highly correlated actual service sample data under specific environments is gradually integrated into the basic model, achieving accurate adaptation and continuous optimization of the model. This effectively overcomes the problem that small sample data cannot support high-precision modeling, while maximizing the use of the general knowledge base of standard environment data and avoiding the performance degradation caused by directly mixing heterogeneous data. Finally, based on the trained predictor, the actual range boundary under the current environment can be accurately calculated according to the real-time service monitoring life of the elastic element, achieving intelligent calibration. Compared with traditional calibration methods that rely on preset standard environmental data, the technical solution provided in this application significantly breaks through the limitations of environmental differences on the applicability of the model. It achieves high-precision, adaptive prediction and reliable calibration of the pressure gauge range boundary under conditions where environmental characteristics are inconsistent with standard environments, thereby improving the measurement accuracy of pressure gauges under complex working conditions.

[0173] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0174] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0175] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A magnetostrictive sensor based pressure gauge range intelligent calibration method, characterized in that, The method comprises the following steps: When the monitoring environment characteristics of the target pressure gauge model are inconsistent with the standard environment characteristics, a first elastic element sample set satisfying the monitoring environment characteristics of the target pressure gauge model is collected for a plurality of first service lives, a range boundary analysis is performed on the target pressure gauge model by using a magnetostrictive sensor, and a plurality of first range boundaries are obtained, wherein the standard environment characteristics are ideal service environments of the elastic element, and the monitoring environment characteristics are actual service environments of the elastic element; A first training data set is configured by taking the plurality of first range boundaries as supervision and the plurality of first service lives as input; When the data amount of the first training data set is less than or equal to a training quantity threshold, a second elastic element sample set satisfying the standard environment characteristics of the target pressure gauge model is collected for a plurality of second service lives, a range boundary analysis is performed on the target pressure gauge model by using a magnetostrictive sensor, and a plurality of second range boundaries are obtained; A second training data set is configured by taking the second range boundary set as supervision and the plurality of second service lives as input, and a first range boundary predictor is trained by using machine learning; The first training data set is called to perform neuron progressive incremental training on the first range boundary predictor, and a second range boundary predictor is generated, and the target pressure gauge is subjected to intelligent calibration of a range based on the service monitoring life of the elastic element.

2. The method of claim 1, wherein, When the monitoring environment characteristics of the target pressure gauge model are inconsistent with the standard environment characteristics, the method comprises the following steps: Based on the target pressure gauge model, a matching accessory table is matched to obtain an elastic element material; The elastic element deformation detection record data set is collected by taking the elastic element material as a constraint, wherein any set of elastic element deformation detection record data of the elastic element deformation detection record data set comprises storage environment record data, probe pressure record data, element size record data, pressure probe position record data and elastic deformation variable record data; Based on the probe pressure record data, the element size record data and the pressure probe position record data, the elastic element deformation detection record data set is subjected to cluster analysis to obtain a plurality of clusters of elastic element deformation detection record data; Based on the plurality of clusters of elastic element deformation detection record data, the storage environment record data and the elastic deformation variable record data are combined to perform deformation correlation sorting on an environment attribute set to obtain a selected environment attribute set; Based on the selected environment attribute set, the standard environment characteristics are configured, and the monitoring environment characteristics are collected; When any one of the attribute environment characteristics of the monitoring environment characteristics and the standard environment characteristics is different, the monitoring environment characteristics and the standard environment characteristics are regarded as inconsistent, otherwise, the monitoring environment characteristics and the standard environment characteristics are regarded as consistent.

3. The method of claim 2, wherein, Based on the plurality of clusters of elastic element deformation detection record data, the storage environment record data and the elastic deformation variable record data are combined to perform deformation correlation sorting on an environment attribute set to obtain a selected environment attribute set, which comprises the following steps: From the plurality of clusters of elastic element deformation detection record data, a one-to-one corresponding first storage environment record data set and a first elastic deformation variable record data set are extracted from a first cluster of elastic element deformation detection record data; Based on the first storage environment record data set and the first elastic variable record data set, the first cluster elastic element deformation detection record data is calculated for each same attribute deviation, and one-to-one corresponding first attribute storage environment deviation module calculation data is obtained until the Nth attribute storage environment deviation module calculation data and elastic variable deviation calculation data; Based on the elastic variable deviation calculation data, the first attribute storage environment deviation module calculation data until the Nth attribute storage environment deviation module calculation data is subjected to gray correlation degree analysis, and first cluster first attribute correlation degree until first cluster Nth attribute correlation degree is obtained; Until the Mth cluster first attribute correlation degree until the Mth cluster Nth attribute correlation degree is obtained; According to the first cluster first attribute correlation degree until the first cluster Nth attribute correlation degree, and the Mth cluster first attribute correlation degree until the Mth cluster Nth attribute correlation degree, same attribute mean value calculation is performed, and first attribute comprehensive correlation degree until Nth attribute comprehensive correlation degree is obtained; According to the first attribute comprehensive correlation degree until the Nth attribute comprehensive correlation degree, the environment attribute with a comprehensive correlation degree greater than or equal to a correlation degree threshold is extracted and added to the selected environment attribute set.

4. The method of claim 1, wherein, Collect a first elastic element sample set of a plurality of first service lives of a target pressure gauge type that meets the monitoring environment characteristics, and perform range boundary analysis on the target pressure gauge type through the magnetostrictive sensor to obtain a plurality of first range boundaries, including: Extract the first service life from the plurality of first service lives; Based on the first service life, extract a first service life elastic element sample set from the first elastic element sample set; Traverse the first service life elastic element sample set, perform range boundary analysis on the target pressure gauge type through the magnetostrictive sensor to obtain a range boundary set; Perform upper boundary central value evaluation on the range boundary set to obtain a range upper boundary, and perform lower boundary central value evaluation on the range boundary set to obtain a range lower boundary; Based on the range upper boundary and the range lower boundary, a first range boundary is constructed and added to the plurality of first range boundaries.

5. The method of claim 4, wherein, Traverse the first service life elastic element sample set, perform range boundary analysis on the target pressure gauge type through the magnetostrictive sensor to obtain a range boundary set, including: Extract the first service life elastic element sample from the first service life elastic element sample set; Obtain the rated range of the first service life elastic element sample, wherein the rated range is the factory range of the target pressure gauge; Based on the rated range of the first service life elastic element sample, perform upper boundary analysis on the first service life elastic element sample through the magnetostrictive sensor to obtain the upper boundary of the range of the first service life elastic element sample; Based on the rated range of the first service life elastic element sample, perform lower boundary analysis on the first service life elastic element sample through the magnetostrictive sensor to obtain the lower boundary of the range of the first service life elastic element sample; According to the first service life elastic element sample range upper boundary and the first service life elastic element sample range lower boundary, a first service life elastic element sample range boundary is added to the range boundary set.

6. The method of claim 5, wherein, Based on the first service life elastic element sample rated range, the first service life elastic element sample is analyzed by the magnetostrictive sensor to obtain the first service life elastic element sample range upper boundary. The first service life elastic element sample is detected multiple times at the working pressure position by the rated range median pressure, the rated range quarter-bit pressure, and the rated range three-quarter-bit pressure, respectively, and the magnetostrictive sensor is used to monitor multiple median pressure detection deformation variables, multiple quarter-bit pressure detection deformation variables, and multiple three-quarter-bit pressure detection deformation variables, wherein the number of detections is greater than or equal to 5 times. The first elastic element qualified quality test data, the second elastic element qualified quality test data, and the third elastic element qualified quality test data of the target pressure type number with a service life less than or equal to a service life threshold are collected with the rated range median pressure, the rated range quarter-bit pressure, and the rated range three-quarter-bit pressure as constraints, respectively. The first elastic element qualified quality test data, the second elastic element qualified quality test data, and the third elastic element qualified quality test data are evaluated for elastic element deformation variable central values to obtain a median pressure detection standard deformation variable, a quarter-bit pressure detection standard deformation variable, and a three-quarter-bit pressure detection standard deformation variable. The first abnormal deformation variable proportion of the multiple median pressure detection deformation variables with a first deformation variable deviation greater than or equal to a deformation deviation threshold from the median pressure detection standard deformation variable is calculated. The second abnormal deformation variable proportion of the multiple quarter-bit pressure detection deformation variables with a second deformation variable deviation greater than or equal to a deformation deviation threshold from the quarter-bit pressure detection standard deformation variable is calculated. The third abnormal deformation variable proportion of the multiple three-quarter-bit pressure detection deformation variables with a third deformation variable deviation greater than or equal to a deformation deviation threshold from the three-quarter-bit pressure detection standard deformation variable is calculated. When any one of the first abnormal deformation variable proportion, the second abnormal deformation variable proportion, and the third abnormal deformation variable proportion is greater than or equal to an abnormal deformation variable proportion threshold, a range failure prompt is generated and sent to the user end, wherein 0.15 ≤ abnormal deformation variable proportion threshold ≤ 0.

3. When the first abnormal deformation variable proportion, the second abnormal deformation variable proportion, and the third abnormal deformation variable proportion are all less than the abnormal deformation variable proportion threshold, the first service life elastic element sample is analyzed by the magnetostrictive sensor based on the first service life elastic element sample rated range to obtain the first service life elastic element sample range upper boundary.

7. The method of claim 6, wherein, When the first abnormal deformation variable proportion, the second abnormal deformation variable proportion and the third abnormal deformation variable proportion are all less than the abnormal deformation variable proportion threshold, based on the first service life elastic element sample rated range, the range upper boundary of the first service life elastic element sample is analyzed by the magnetostrictive sensor to obtain the first service life elastic element sample range upper boundary, including: The fourth abnormal deformation variable proportion of the upper limit pressure of the rated range of the first service life elastic element sample is detected by the magnetostrictive sensor; When the fourth abnormal deformation variable proportion is less than the abnormal deformation variable proportion threshold, the upper limit pressure of the rated range is set as the first service life elastic element sample range upper boundary; When the fourth abnormal deformation variable proportion is greater than or equal to the abnormal deformation variable proportion threshold, the updated upper limit pressure of the rated range is obtained by subtracting the pressure consistency deviation threshold from the upper limit pressure of the rated range, and the loop analysis is executed.

8. The method of claim 1, wherein, The first training data set is called to perform neuron progressive incremental training on the first range boundary predictor to generate a second range boundary predictor, including: The network parameters of the first range boundary predictor are frozen, the first full connection neural network is configured, the output of the first range boundary predictor is taken as the input of the first full connection neural network, and the first range boundary is taken as the supervision of the first full connection neural network to train the first progressive range boundary predictor; When the first verification accuracy of the first progressive range boundary predictor is greater than or equal to the convergence accuracy, the first progressive range boundary predictor is set as the second range boundary predictor; Otherwise, the network parameters of the first progressive range boundary predictor are frozen, the second full connection neural network is configured, the output of the first progressive range boundary predictor is taken as the input of the second full connection neural network, and the first range boundary is taken as the supervision of the second full connection neural network to train the second progressive range boundary predictor to execute the loop analysis; Until the Qth verification accuracy of the Qth progressive range boundary predictor is greater than or equal to the convergence accuracy, the Qth progressive range boundary predictor is set as the second range boundary predictor.

9. A magnetostrictive sensor based pressure gauge range intelligent calibration system, characterized by, A magnetostrictive sensor pressure gauge range intelligent calibration method for implementing any one of claims 1-8, the system comprising: A first range boundary acquisition module is configured to collect a first elastic element sample set of a target pressure gauge type when the monitoring environment characteristics of the target pressure gauge type are inconsistent with the standard environment characteristics, and to perform range boundary analysis on the target pressure gauge type by a magnetostrictive sensor to obtain a plurality of first range boundaries, wherein the standard environment characteristics are pre-stored ideal service environment of the elastic element, and the monitoring environment characteristics are the actual service environment of the elastic element. A training data configuration module is configured to configure a first training data set with the first range boundary as the supervision and the first service life as the input. The second range boundary acquisition module is configured to collect a second elastic element sample set of a plurality of second service lives of a target pressure gauge model satisfying a standard environmental feature when a first training data set is less than or equal to a training quantity threshold, and perform range boundary analysis on the target pressure gauge model by using a magnetostrictive sensor to obtain a plurality of second range boundaries. The first range boundary prediction module is configured to configure a second training data set by taking the second range boundary set as supervision and taking the plurality of second service lives as input, and train a first range boundary predictor by using machine learning. The range intelligent calibration module is configured to perform neuron progressive incremental training on the first range boundary predictor by using the first training data set, generate a second range boundary predictor, and perform range intelligent calibration on the target pressure gauge based on the service monitoring life of the elastic element.

Images