Automatic switching metering method suitable for multiple ranges of pressure gauge

Abstract

The invention discloses an automatic switching metering method suitable for multiple ranges of a pressure gauge, and the method comprises the steps: S1, obtaining the information of a current pressure measurement environment, inputting a pressure range prediction model, and outputting an initial recommended range; s2, selecting a measurement parameter of a corresponding measuring range, and performing preliminary measurement on the pressure at the initial sampling frequency to obtain a preliminary measurement pressure value; s3, comparing the preliminarily measured pressure value with the upper range limit and the lower range limit of the selected range, and judging that a higher / lower range needs to be switched; s4, the calibration coefficient of the range to be switched is obtained and corrected, and a corrected pressure measurement value is obtained; s5, dynamically adjusting the sampling frequency; and S6, the pressure change rate is monitored in real time, and the measuring range capable of covering the current pressure change range is preferentially selected for measurement. According to the invention, multi-range intelligent switching of the pressure gauge is realized, and the measurement precision and response speed in the full-range range are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of pressure measurement, and particularly to an automatic switching measurement method applicable to multiple ranges of pressure gauges. Background Art

[0002] In industrial production and scientific research and testing scenarios, the accurate measurement of pressure parameters is crucial for stability, product quality control, and safe operation. In actual applications, the pressure environment is complex and variable, experiencing wide-range fluctuations from micro-pressure to high-pressure under different working conditions. Traditional single-range pressure gauges cannot meet the full-range high-precision measurement requirements, and the operation modes of manually replacing multi-range pressure gauges or manually switching ranges have significant limitations, resulting in low measurement efficiency, interrupted data continuity, and even measurement distortion caused by operation delays.

[0003] Traditional pressure measurement techniques mainly use single-range pressure gauges or manually switched multi-range pressure gauges. Single-range pressure gauges have a simple structure and low cost, but can only cover a fixed pressure range. When the measured pressure exceeds the range, the equipment needs to be replaced, and they cannot adapt to the dynamically changing pressure environment. Especially in scenarios such as chemical reactions and experimental tests, key data is missing due to range limitations. Manually switched multi-range pressure gauges can cover a wider range, but rely on the operator's experience to judge the switching timing, resulting in cumbersome operations and response lags. For example, during a rapid pressure rise, manual switching causes over-range measurement due to delays, damaging the equipment or generating incorrect data. At the same time, frequent manual operations easily cause mechanical wear, affecting the equipment life, and it is difficult to effectively eliminate the calibration errors during different range switches, resulting in poor consistency of measurement data.

[0004] Although there have been some attempts at automatic range switching in the prior art, such as mechanical switching devices or primary electronic switching systems based on simple threshold comparisons, there are still significant defects. For example, the switching method based on fixed thresholds does not consider the influence of pressure change trends and environmental factors. When environmental parameters such as temperature and humidity fluctuate, it is easy to cause incorrect switching or untimely switching due to sensor drift. Some automatic switching systems using a single algorithm lack the ability to comprehensively analyze multi-source data and only determine the range based on the current pressure value, unable to predict short-term pressure changes, resulting in a lagging switching strategy. Especially in scenarios of sudden pressure changes, measurement interruptions occur due to insufficient range coverage.

[0005] Therefore, the present invention proposes an automatic switching measurement method applicable to multiple ranges of pressure gauges, realizing the intelligent switching of multiple ranges of pressure gauges, improving the measurement accuracy and response speed within the full range, and providing an innovative solution for accurate measurement in complex pressure environments. Summary of the Invention

[0006] Based on the above technical problems, the present application discloses an automatic switching measurement method applicable to multiple ranges of pressure gauges, including:

[0007] S1. Obtain the current pressure measurement environment information, input it into a preset pressure range prediction model, and output the initial recommended range for pressure measurement in the current environment;

[0008] S2. According to the initial recommended range, select the measurement parameters corresponding to the range, and preliminarily measure the pressure at the initial sampling frequency to obtain the preliminary measured pressure value;

[0009] S3. Compare the preliminary measured pressure value with the upper and lower limits of the selected range. If the preliminary measured pressure value exceeds the preset maximum threshold of the selected range, it is determined that a higher range needs to be switched. If the preliminary measured pressure value is lower than the preset minimum threshold of the selected range, it is determined that a lower range needs to be switched;

[0010] S4. Obtain the calibration coefficient of the range to be switched, and correct the measurement data after switching the range according to the calibration coefficient to obtain the corrected pressure measurement value;

[0011] S5. Dynamically adjust the sampling frequency according to the relationship between the corrected pressure measurement value and the current range. If the measured value is close to the range boundary, increase the sampling frequency. If the measured value is in the middle area of the range, decrease the sampling frequency;

[0012] S6. Real-time monitor the pressure change rate. When the pressure change rate exceeds the preset threshold, trigger the emergency range switching mechanism, and preferentially select the range that can cover the current pressure change range for measurement.

[0013] Preferably, the presetting of the pressure range prediction model in S1 is specifically as follows: Obtain historical pressure measurement data, divide the data into a training set, a validation set, and a test set according to the time series, and label the actually applicable pressure range for each data sample; Construct a time series feature-aware network as a deep learning architecture, capture the temporal dependence of pressure data and environmental information through a dynamic memory gating module, and assign adaptive weights to the features of different time steps according to the pressure change trend through a trend weight allocation module; Use the pressure measurement environment information and historical pressure fluctuation data as inputs, and use the cross-entropy loss function to train the model, where N is the number of samples, M is the number of preset pressure range categories, y ij is the true label that sample i belongs to range category j, is the probability that the model predicts that sample i belongs to range category j. Minimize the loss function value by adjusting the model parameters on the validation set, and verify the generalization ability of the model on the test set to construct a pressure range prediction model.

[0014] Preferably, in S1, the temperature T, humidity H in the current pressure measurement environment information, and historical pressure fluctuation data {P t-n ,Pt-n+1 , …, P t} are normalized to construct an input vector Input X into a pre-set pressure range prediction model, and after being processed by the dynamic memory gating module of the time series feature perception network, obtain a hidden state sequence H = [h1, h2, …, h T , and calculate the attention weights of each hidden state through the trend weight allocation module where w and W are trainable weight matrices respectively, and b is a bias vector; perform weighted summation on the hidden states according to the attention weights to obtain a pressure continuous vector T r is the total number of time steps of the pressure fluctuation historical data; finally, input the pressure continuous vector c into a fully connected layer, and through the Softmax function calculate the prediction probabilities of each pressure range category, and take the range corresponding to the category with the largest probability as the initial recommended range for pressure measurement in the current environment, where v j is the weight vector corresponding to the j-th range category of the fully connected layer, and M is the number of preset pressure range categories

[0015] Preferably, the method for selecting the measurement parameters corresponding to the range in S2 is: establish a parameter database including the inherent parameters, environmental compensation parameters and dynamic adjustment parameters of each range pressure gauge. The inherent parameters include the sensor sensitivity coefficient S i , and the nonlinear error correction matrix E i ; the environmental compensation parameters construct a compensation function according to the temperature T and humidity H where a mn is a compensation coefficient, and the dynamic adjustment parameters set a sampling frequency adjustment factor according to the pressure change gradient γ of the initial recommended range λ is an adjustment coefficient. Index the parameter database according to the initial recommended range, and perform weighted fusion on the inherent parameters, environmental compensation parameters and dynamic adjustment parameters to obtain a measurement parameter set

[0016] Preferably, the method for obtaining the preliminary measured pressure value in S2 is: use the sensor sensitivity coefficient in the selected measurement parameter set to perform linear conversion on the original output signal V of the sensor raw to obtain a preliminary pressure value Perform polynomial fitting correction through the nonlinear error correction matrix , and the correction formula is where is the coefficient in the i-th row and k-th column of, and N ris the fitting order. The corrected pressure value is compensated by combining with the environmental compensation function C(T, H) to obtain the preliminary measured pressure value P pre = P pre2 ×(1 + C(T, H)).

[0017] Preferably, when comparing the preliminary measured pressure value with the upper and lower limits of the selected range in S3, a pressure change trend prediction model is constructed through a dynamic threshold adjustment mechanism to model the sequence of the most recent n preliminary measured pressure values and predict the pressure value at the next moment According to the predicted pressure value and the current upper limit of the range P max calculate the dynamic maximum threshold, and the formula is: where β s is the upper limit threshold adjustment coefficient. If the preliminary measured pressure value P pre is greater than the dynamic maximum threshold P dmax , it is determined that a switch to a higher range is required.

[0018] Preferably, based on the pressure change trend prediction model in S3, the sequence of the most recent n preliminary measured pressure values is predicted to obtain the pressure value at the next moment According to the predicted pressure value and the current lower limit of the range P min calculate the dynamic minimum threshold where γ s is the lower limit threshold adjustment coefficient. If the preliminary measured pressure value P pre is less than the dynamic minimum threshold Pd min , it is determined that a switch to a lower range is required.

[0019] Preferably, in S4, the calibration coefficient of the range to be switched is obtained for correction. Specifically: a dynamic calibration parameter table corresponding to each range is established, including the calibration offset ΔP cal and the scale factor K cal in different temperature intervals and pressure segments; when switching to the target range, the corresponding temperature compensation interval is queried according to the current ambient temperature T, and the calibration offset ΔP cal and the scale factor K cal corresponding to the preliminary measured pressure value P cal in this interval are extracted; for the original measured value P raw after switching the range, the piecewise linear correction formula is adopted: where P mid is the intermediate pressure value of the target range, K cal and K′ cal , ΔP cal and ΔP′ calCalibration coefficients for the low-pressure section and the high-pressure section respectively.

[0020] Preferably, in S5, the sampling frequency is dynamically adjusted, specifically: setting the effective measurement interval coefficient λ of the current range e , and the formula is: where P max and P min are the upper and lower limits of the current range, and the sampling frequency f is dynamically adjusted through an exponential function model: where f base is the basic sampling frequency, η is the boundary sensitivity coefficient, and μ is the smoothing factor; when P corr is close to the range boundary, the sampling frequency f is increased exponentially to f s , and the formula is: f s = f base ×(1 + η); when P corr is in the middle of the range, the sampling frequency remains f base .

[0021] Preferably, the emergency range switching mechanism in S6 is specifically: establishing a pressure coverage characteristic matrix for each range, and the matrix records the pressure response delay time, the minimum measurable pressure change amount, and the range overlap range of the corresponding range. When the pressure change rate exceeds the preset threshold, the current pressure value and the change trend data are extracted, and the ranges that meet the condition that the lower limit of the range is lower than the current pressure value and the upper limit of the range is higher than the predicted extreme value of the pressure change are screened in sequence from the low range to the high range as candidates. If there are multiple candidate ranges, they are prioritized according to the pressure response delay time and the minimum measurable pressure change amount recorded in the matrix, and the range with the shortest response delay time and meeting the current pressure change resolution requirement is preferably selected; if all ranges cannot fully cover the predicted pressure change range, by enabling combined range measurement, parallel measurement is performed through two adjacent ranges, and the measurement data of the two ranges are fused to generate a pressure measurement value.

[0022] Compared with the prior art, the technical solution of the present application has the following technical effects:

[0023] By constructing a time-series feature perception network, the present invention deeply analyzes the historical data of temperature, humidity, and pressure fluctuations, realizes the intelligent prediction of the initial recommended range, uses a deep learning model to capture the time-series dependence relationship and environmental parameter association of pressure data, breaks through the mechanicalness of traditional fixed-threshold switching, makes the range selection more suitable for real-time working conditions, the model can predict the pressure change trend through historical data, and recommends a suitable range in advance to avoid the risk of over-range caused by the lag of manual judgment; the mechanism of dynamically allocating time-step feature weights can focus on the data features at the moment of pressure mutation, improve the sensitivity of range prediction to transient changes, and ensure that the measurement maintains the optimal range matching state throughout the process.

[0024] By establishing a database containing inherent parameters, environmental compensation parameters, and dynamic adjustment parameters, the present invention realizes the intelligent fusion of multi-dimensional parameters, the dynamic invocation of the sensor sensitivity coefficient and the non-linear error correction matrix, and can perform precise linearization correction for the characteristics of sensors with different ranges to eliminate the inherent errors of the hardware. Based on the polynomial compensation function of temperature-humidity, it can calibrate in real time the influence of environmental drift on the measurement results, solve the non-linear deviation problem of traditional single coefficient calibration in a wide range, and meet the strict requirements of precision measurement scenarios for the consistency of the full range.

[0025] The combination of the sampling frequency dynamic adjustment mechanism and the dynamic threshold adjustment strategy in the present invention improves the real-time response ability to pressure changes. By calculating the relative position of the pressure value within the range and using the exponential function model to achieve non-linear adjustment of the sampling frequency: when the measured value is close to the range boundary, the sampling frequency automatically increases to the highest gear to capture the high-frequency details of pressure mutations, while when the measured value is in the middle of the range, the sampling frequency is reduced to save computing resources and balance efficiency and accuracy; the dynamic threshold adjustment predicts the pressure trend through the grey prediction model and adjusts the switching threshold in real time to avoid the lag problem of traditional fixed ratio thresholds during rapid pressure rise, preventing over-range damage and data distortion.

[0026] The emergency range switching mechanism designed in the present invention ensures the reliability of the measurement through the pressure coverage characteristic matrix and the combined measurement mode. The characteristic matrix integrates key indicators such as response delay and resolution of each range. When the pressure change rate exceeds the standard, it can quickly screen and preferentially activate the range with the fastest response and the most matching coverage; when a single range cannot cover the predicted pressure range, the combined measurement mode breaks through the range limit by activating adjacent ranges in parallel and fusing data, and enables parallel measurement of the medium pressure and high pressure ranges at the same time, solving the measurement blind area problem of traditional single ranges when the pressure span exceeds multiple orders of magnitude, and ensuring the integrity and credibility of full-condition data.

[0027] The above description is only an overview of the technical solution of this application. In order to be able to understand the technical means of this application more clearly, so as to be implemented in accordance with the content of the specification, and in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the following takes the preferred embodiments of this application and combines the drawings to describe in detail as follows.

[0028] Those skilled in the art will understand more clearly the above and other purposes, advantages and features of this application according to the following detailed description of the specific embodiments of this application in combination with the drawings. Brief Description of the Drawings

[0029] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts do not necessarily draw according to the actual proportion.

[0030] Figure 1 Flow chart of an automatic switching measurement method applicable to multiple ranges of pressure gauges according to the present invention;

[0031] Figure 2 Comparison chart of error curves in the low-speed pressure increase stage in the embodiment;

[0032] Figure 3 Comparison chart of pressure curves in the high-speed impact stage in the embodiment;

[0033] Figure 4 Comparison chart of periodic sampling frequencies in the embodiment;

[0034] Figure 5 Comparison chart of periodic pressure errors in the embodiment;

[0035] Figure 6 Comparison chart of data integrity under mixed working conditions in the embodiment. Specific implementation manners

[0036] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. In the following description, specific details such as specific configurations and components are provided only to help a comprehensive understanding of the embodiments of the present application. Therefore, those skilled in the art should clearly understand that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. In addition, descriptions of known functions and structures are omitted in the embodiments for clarity and conciseness.

[0037] It should be understood that the "one embodiment" or "the present embodiment" mentioned throughout the specification means that a specific feature, structure or characteristic related to the embodiment is included in at least one embodiment of the present application. Therefore, the "one embodiment" or "the present embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner.

[0038] In addition, the present application may repeat reference numerals and / or letters in different instances. Such repetition is for the purpose of simplicity and clarity and does not in itself indicate a relationship between the various embodiments and / or configurations discussed.

[0039] As used herein, the term "and / or" is merely a description of the relationship between associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, B exists alone, and A and B exist simultaneously. The term " / and" as used herein describes another relationship between associated objects, indicating that two relationships may exist. For example, A / and B may represent: A exists alone, and A and B exist alone. Additionally, the character " / " as used herein generally indicates that the associated objects before and after are in an "or" relationship.

[0040] As used herein, the term "at least one" is merely a description of the relationship between associated objects, indicating that three relationships may exist. For example, at least one of A and B may represent: A exists alone, A and B exist simultaneously, and B exists alone.

[0041] It should also be noted that, as used herein, relational terms such as first and second are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprising", "including" or any other variation thereof are intended to cover non-exclusive inclusion.

[0042] Embodiment 1

[0043] This embodiment mainly describes an automatic switching metering method applicable to multiple ranges of pressure gauges, as Figure 1 shown, specifically including:

[0044] S1. Obtain the current pressure measurement environment information, input it into a preset pressure range prediction model, and output the initial recommended range for pressure measurement in the current environment;

[0045] S2. According to the initial recommended range, select the measurement parameters corresponding to the range, and preliminarily measure the pressure at the initial sampling frequency to obtain a preliminary measured pressure value;

[0046] S3. Compare the preliminary measured pressure value with the upper and lower limits of the selected range. If the preliminary measured pressure value exceeds the preset maximum threshold of the selected range, it is determined that a switch to a higher range is required. If the preliminary measured pressure value is lower than the preset minimum threshold of the selected range, it is determined that a switch to a lower range is required;

[0047] S4. Obtain the calibration coefficient of the range to be switched, and correct the measurement data after switching the range according to the calibration coefficient to obtain a corrected pressure measurement value;

[0048] S5. Dynamically adjust the sampling frequency according to the relationship between the corrected pressure measurement value and the current range. If the measurement value is close to the range boundary, increase the sampling frequency; if the measurement value is in the middle area of the range, decrease the sampling frequency.

[0049] S6. Real-time monitor the pressure change rate. When the pressure change rate exceeds the preset threshold, trigger the emergency range switching mechanism, and preferentially select the range that can cover the current pressure change range for measurement.

[0050] Furthermore, the presetting of the pressure range prediction model in S1 is specifically as follows: Obtain historical pressure measurement data, divide the data into a training set, a validation set, and a test set according to the time series, and label the actual applicable pressure range for each data sample; Construct a time series feature-aware network as the deep learning architecture, capture the temporal dependencies of pressure data and environmental information through the dynamic memory gating module, and assign adaptive weights to the features of different time steps according to the pressure change trend through the trend weight allocation module; Use the pressure measurement environment information and the historical pressure fluctuation data as inputs, and use the cross-entropy loss function to train the model, where N is the number of samples, M is the number of preset pressure range categories, y ij is the true label that sample i belongs to range category j, is the probability that the model predicts that sample i belongs to range category j. Minimize the loss function value by adjusting the model parameters on the validation set, and verify the generalization ability of the model on the test set to construct the pressure range prediction model.

[0051] Furthermore, in S1, normalize the temperature T, humidity H in the current pressure measurement environment information and the historical pressure fluctuation data {P t-n , P t-n+1 , …, P t} to construct the input vector Input X into the preset pressure range prediction model, and after being processed by the dynamic memory gating module of the time series feature-aware network, obtain the hidden state sequence H = [h1, h2, …, h T , calculate the attention weights of each hidden state through the trend weight allocation module where w and W are trainable weight matrices respectively, and b is the bias vector; Perform weighted summation on the hidden states according to the attention weights to obtain the pressure continuous vector T r is the total number of time steps of the historical pressure fluctuation data; Finally, input the pressure continuous vector c into the fully connected layer, and calculate the prediction probabilities of each pressure range category through the Softmax function Take the range corresponding to the category with the largest probability as the initial recommended range for pressure measurement in the current environment, where vj is the weight vector corresponding to the j-th range category of the fully connected layer, and M is the number of preset pressure range categories.

[0052] Furthermore, the method for selecting the measurement parameters corresponding to the range in S2 is as follows: establish a parameter database including the inherent parameters, environmental compensation parameters, and dynamic adjustment parameters of each range pressure gauge. The inherent parameters include the sensor sensitivity coefficient S i , the non-linear error correction matrix E i ; the environmental compensation parameters construct a compensation function according to the temperature T and humidity H where a mn is the compensation coefficient, and the dynamic adjustment parameters set the sampling frequency adjustment factor according to the pressure change gradient γ of the initial recommended range λ is the adjustment coefficient. According to the initial recommended range index, the parameter database is used to perform weighted fusion of the inherent parameters, environmental compensation parameters, and dynamic adjustment parameters to obtain the measurement parameter set

[0053] Furthermore, the method for obtaining the preliminary measured pressure value in S2 is as follows: use the sensor sensitivity coefficient in the selected measurement parameter set to perform linear conversion on the original output signal V raw of the sensor to obtain the preliminary pressure value Perform polynomial fitting correction through the non-linear error correction matrix , and the correction formula is where is the coefficient of the i-th row and k-th column in , N r is the fitting order. Combine the environmental compensation function C(T, H) to compensate the corrected pressure value to obtain the preliminary measured pressure value P pre = P pre2 ×(1 + C(T, H)).

[0054] Furthermore, when comparing the preliminary measured pressure value with the upper and lower limits of the selected range in S3, a pressure change trend prediction model is constructed through a dynamic threshold adjustment mechanism to model the sequence of the most recent n preliminary measured pressure values to predict the pressure value at the next moment According to the predicted pressure value and the current upper range limit P max calculate the dynamic maximum threshold, and the formula is: where β s is the upper limit threshold adjustment coefficient. If the preliminary measured pressure value P pre is greater than the dynamic maximum threshold P dmax , it is determined that a switch to a higher range is required.

[0055] Further, in S3, based on the pressure change trend prediction model, for the sequence of the most recent n preliminary measurement pressure values a prediction is made to obtain the pressure value at the next moment According to the predicted pressure value and the current range lower limit P min the dynamic minimum threshold is calculated where γ s is the lower threshold adjustment coefficient. If the preliminary measurement pressure value P pre is less than the dynamic minimum threshold Pd min , it is determined that a switch to a lower range is required.

[0056] Further, in S4, the calibration coefficient of the range to be switched is obtained for correction. Specifically: A dynamic calibration parameter table corresponding to each range is established, including the calibration offset ΔP cal and the scale factor K cal in different temperature intervals and pressure segments; when switching to the target range, according to the current ambient temperature T, the corresponding temperature compensation interval is queried, and the calibration offset ΔP cal corresponding to the preliminary measurement pressure value P cal and the scale factor K cal in this interval are extracted; for the original measurement value P raw after switching the range, the piecewise linear correction formula is adopted: where P mid is the intermediate pressure value of the target range, K cal and K′ cal , ΔP cal and ΔP′ cal are the calibration coefficients of the low pressure segment and the high pressure segment respectively.

[0057] Further, in S5, the sampling frequency is dynamically adjusted. Specifically: The effective measurement interval coefficient λ e of the current range is set, and the formula is: where P max and P min are the upper and lower limits of the current range. The sampling frequency f is dynamically adjusted through the exponential function model: where f base is the base sampling frequency, η is the boundary sensitivity coefficient, and μ is the smoothing factor; when P corr is close to the range boundary, the sampling frequency f is exponentially increased to f s , and the formula is: f s = f base ×(1 + η); when P corr is in the middle of the range, the sampling frequency remains f base .

[0058] Furthermore, the emergency range switching mechanism in S6 is specifically as follows: a pressure coverage characteristic matrix is established for each range, and the matrix records the pressure response delay time, the minimum measurable pressure change and the range overlap range of the corresponding range. When the pressure change rate exceeds the preset threshold, the current pressure value and change trend data are extracted, and the ranges with a range lower limit lower than the current pressure value and a range upper limit higher than the predicted pressure change extreme value are screened from low range to high range as candidates. If there are multiple candidate ranges, they are prioritized according to the pressure response delay time and the minimum measurable pressure change recorded in the matrix, and the range with the shortest response delay time and that can meet the current pressure change resolution requirements is selected first; if all ranges cannot fully cover the predicted pressure change range, the combined range measurement is enabled, and parallel measurements are performed through two adjacent ranges, and the measurement data of the two ranges are integrated to generate a pressure measurement value.

[0059] This embodiment describes in detail the multi-range automatic switching measurement method for the pressure gauge of the present application. It realizes accurate measurement of the full range by intelligently predicting the initial range through the time series feature perception network, combining dynamic threshold adjustment, segmented calibration and adaptive sampling technology. The dynamic calibration parameter table and the combined range emergency mechanism solve the problems of cross-range error compensation and sudden pressure coverage. The trend prediction model enhances the switching prediction capability, improves measurement efficiency and accuracy, reduces the risk of manual operation, is suitable for complex pressure scenarios, and provides an intelligent and highly reliable solution for wide-range pressure measurement.

[0060] Based on Example 1, this example describes in detail how to construct a pressure change trend prediction model through a dynamic threshold adjustment mechanism, specifically:

[0061] Extract the latest n preliminary measured pressure value sequences from the real-time measurement data of the pressure gauge A sliding time window is formed and the sequence is preprocessed: the high-frequency noise is removed by the Savitzky-Golay filter, and the low-frequency trend component of the pressure change is retained; the pressure change rate at adjacent moments is calculated using the difference method. Construct a two-dimensional feature vector containing pressure values and change rates. Perform a Fourier transform on the data within the time window to extract the dominant frequency components and determine the periodicity of pressure fluctuations. Fit the envelope of the pressure series through local weighted regression to separate trend terms from random fluctuation terms, ensuring that the model focuses on deterministic trend prediction. Input the preprocessed feature vector into the main prediction model.

[0062] Through the grey prediction GM (1,1) algorithm, randomness is eliminated through data generation, and a first-order linear differential equation is established to describe the trend: Perform one cumulative generation (1-AGO) to obtain a new sequence S k ,in Convert the random fluctuations of the original data into a monotonically increasing or decreasing trend sequence; construct a grey differential equation where a is the development coefficient (reflecting the rate of pressure change), and b is the grey action quantity (reflecting the comprehensive influence of external factors), and solve the parameter vector by the least squares method Obtain the whitenized differential equation Its general solution is the time response function Perform inverse accumulation on the predicted values of the accumulated generation to obtain the predicted values of the original sequence where is the predicted pressure value at the next moment. Fit the pressure trend through an exponential function, which is applicable to short-term prediction in the initial stage of slow or sudden pressure changes in industrial scenarios.

[0063] The dynamic characteristics of the pressure sequence can be captured in real time through the pressure change trend prediction model, providing reliable predicted values for dynamic threshold calculation Achieve accurate judgment of the range switching timing.

[0064] This embodiment details the dynamic threshold adjustment mechanism. By constructing a pressure change trend prediction model based on grey prediction GM(1,1), combined with data preprocessing, feature enhancement, and parameter dynamic optimization, accurate prediction of the pressure value at the next moment is achieved, enabling the range switching threshold to be dynamically adjusted with the trend, improving the system's ability to predict sudden pressure changes, reducing mis-switching and lagging switching, and ensuring measurement continuity and accuracy.

[0065] Based on Embodiment 1, this embodiment details the specific implementation technical means, specifically:

[0066] Select the pressure monitoring scenario of a chemical reactor in a certain chemical enterprise. The pressure range of the equipment during the production cycle is 0.1 - 10 MPa, including working conditions of slow pressure increase (0.05 MPa / min), high-speed impact (5 MPa / s), and periodic fluctuations (period 10 minutes, amplitude ±1.5 MPa); the traditional method uses manual switching of 3 fixed-range pressure gauges (0 - 2 MPa, 0 - 6 MPa, 0 - 12 MPa), and the existing automatic technology uses a fixed 80% upper limit switching threshold;

[0067] Experimental environment and data acquisition. The sensor is a high-precision piezoresistive pressure sensor (range 0 - 12 MPa, accuracy ±0.1% FS), and the sampling frequency is 100 Hz; the data processing unit is an industrial-grade embedded controller (equipped with the algorithm of this application and the existing technology algorithm); standard pressure source: Fluke729 high-pressure calibrator (accuracy ±0.02% FS), used to generate standard pressure signals.

[0068] Collect pressure data for 72 consecutive hours, including:

[0069] Low-speed pressure boosting stage: It rises uniformly from 0.5 MPa to 8 MPa and lasts for 10 hours to simulate the temperature and pressure boosting process of the reactor.

[0070] High-speed impact stage: An instantaneous impact pressure (peak value 11 MPa, duration 200 ms) is applied on the basis of 5 MPa to simulate the feeding impact condition and is repeated 100 times.

[0071] Periodic fluctuation stage: It fluctuates in the range of 3 - 6 MPa according to a sine curve, with a frequency of 0.017 Hz (period 10 minutes) and lasts for 48 hours.

[0072] Mixed condition stage: Randomly combine the above three conditions, including 3 pressure mutations (rate > 2 MPa / s) to simulate abnormal fluctuations in actual production.

[0073] The evaluation indexes include the accuracy rate of range switching, the average measurement error, and the response time;

[0074] Experiment 1: Comparison of low-speed pressure boosting conditions, as shown in Table 1 below;

[0075] The technology of this application analyzes the historical data of the first 3 hours (temperature 25 ± 2 °C, humidity 50 ± 5% RH) through a time series feature perception network, and the initial recommended range is 0 - 6 Mpa; when the pressure rises to 4.8 MPa (80% of the range upper limit), the trend prediction model shows that the pressure will exceed 6 MPa in the next 30 minutes (predicted value 6.2 MPa), and the range switch to 0 - 12 MPa is triggered in advance, and the switching moment is when the pressure reaches 5.5 MPa (10 minutes in advance); after switching, the calibration coefficient K is called through the dynamic calibration parameter table (temperature 28 °C, pressure section 5 - 8 MPa) cal = 1.0032, ΔP cal = -0.015, and the corrected measurement error is stabilized at ±0.03 Mpa;

[0076] For the existing technology, a fixed 80% upper limit switching threshold is adopted. When the pressure reaches 4.8 MPa (80% of the 0 - 6 MPa range), the switch is triggered, and the switching moment is when the pressure reaches 4.8 MPa, without advance prediction; after switching, the fixed calibration coefficient K is used cal = 1.005, the temperature change is not compensated, and the measurement error gradually increases to ±0.08 MPa in the range of 5 - 8 MPa.

[0077] Table 1 Range switching and error data in the low-speed pressure boosting stage

[0078] Index The technology of this application Existing technology Switching moment (MPa) 5.5 (10 minutes in advance) 4.8 (threshold trigger) Average measurement error (MPa) ±0.032(n＝60000) ±0.078(n＝60000) Response time (ms) 120 210 Data integrity (%) 99.9％ 98.5% (230 points lost during switching)

[0079] Such as Figure 2As shown, the technical error of this application is always controlled within ±0.05 MPa, and the data before and after switching is continuous without interruption. For the prior art, the error increases significantly after switching, and in the range of 4.8 - 5.2 MPa, 150 ms of data is lost due to switching delay.

[0080] Experiment 2: Comparison of high-speed impact conditions, as shown in Table 2:

[0081] For the technology of this application, the normal measurement range of pressure is 0 - 6 MPa. When the detected pressure change rate reaches 5 MPa / s (exceeding the preset threshold of 2 MPa / s), the emergency switching mechanism is triggered; candidate ranges are screened from low range to high range, and the 0 - 12 MPa range with the shortest response delay (response time 80 ms) is preferentially selected. At the same time, the combined measurement mode of 0 - 6 MPa and 0 - 12 MPa is enabled; the impact peak value of 11.23 MPa is captured by the 0 - 12 MPa range, and the data of the two ranges is fused through cross-validation, with an error of +0.045 MPa.

[0082] The fixed-threshold switching mechanism of the prior art cannot identify sudden pressure changes. When the pressure exceeds 4.8 MPa (80% of the 0 - 6 MPa range), the switching is triggered, and the response time is 320 ms, resulting in the impact peak value (11.23 MPa) exceeding the range, the sensor overloading and alarming, and the data loss rate being 100% (the impact data of this time is invalid).

[0083] Index The technology of this application Existing technology Peak capture rate 100% (11.23MPa) 0% (not captured due to overrange) Error (MPa) +0.045 -(data loss) Response time (ms) 80 320 (switching not completed) Sensor protection No overload Overload alarm (need to restart calibration)

[0084] As Figure 3 shown, the technology of this application completely captures the impact waveform, and the combined measurement ensures data continuity; for the prior art, the data is truncated after the pressure exceeds 6 MPa, and an "out-of-range" alarm is displayed.

[0085] Experiment 3: Comparison of periodic fluctuation conditions, as shown in Table 3;

[0086] For the technology of this application, the initial recommended range is 0 - 6 MPa, and the sampling frequency is dynamically adjusted: when the pressure is at 3 - 4.5 MPa (in the middle of the range), the sampling frequency is reduced to 20 Hz; when the pressure approaches the 6 MPa boundary, the sampling frequency is increased to 100 Hz; the pressure fluctuation period is predicted through a trend prediction model, and the threshold is adjusted 5 minutes in advance (the dynamic maximum threshold is set to 5.8 MPa) to avoid unnecessary switching triggered by the fluctuation peak value (5.9 MPa), and there is no switching action throughout the process; segmented calibration is adopted, with an error of ±0.02 MPa in the low-pressure section (3 - 4.5 MPa) and an error of ±0.03 MPa in the high-pressure section (4.5 - 6 MPa).

[0087] In the prior art, the fixed sampling frequency is 50Hz. When the pressure exceeds 4.8MPa (80% threshold), it frequently switches to 0 - 12MPa, switching 2 times per cycle (1 time for pressure increase and 1 time for pressure decrease each), resulting in data interruption (about 50 data points are lost each time of switching); the fixed calibration coefficient leads to an error of ±0.06MPa in the low - pressure section and an error of ±0.09MPa in the high - pressure section.

[0088] Index The technology of this application Existing technology Range switching times 0 times 96 times (12 times per hour) Average measurement error (MPa) ±0.028(n＝288000) ±0.075(n＝288000) Number of data loss points 0 points 4800 points (50 points lost each time) Sampling frequency efficiency Average 35Hz (saving 65% resources) Fixed 50Hz

[0089] In the technology of this application, the sampling frequency is dynamically adjusted according to the pressure position, with stable error and no switching loss. In the prior art: frequent switching leads to data fragmentation and large error fluctuations. As Figure 4 shown, through the comparison of sampling frequencies, the blue curve represents the adaptive sampling frequency of the technology of this application (20Hz when the pressure is between 3 - 4.5MPa, rising to 100Hz when approaching 6MPa); the red dashed line represents the fixed 50Hz sampling frequency of the prior art; a semi - transparent pressure curve is added to the right Y - axis for reference; as Figure 5 shown, the blue curve represents the error of the technology of this application (±0.02MPa in the low - pressure section, ±0.03MPa in the high - pressure section); the red curve represents the error of the prior art (±0.06MPa in the low - pressure section, ±0.09MPa in the high - pressure section, including switching spikes); the shaded area marks different pressure sections.

[0090] Experiment 4: Comprehensive comparison under mixed working conditions, as shown in Table 4;

[0091] Simulate a complex scenario including low - speed pressure increase, high - speed impact, and periodic fluctuations, lasting for 24 hours, including 3 pressure mutations (rates are 2.5MPa / s, 3MPa / s, and 4MPa / s respectively);

[0092] Index The technology of this application Existing technology Total switching times 5 times (all correct) 18 times (7 times of mis-switching among them) Average measurement error (MPa) ±0.035 ±0.089 Data integrity (%) 99.9％ 92.3％ Emergency switching success rate 100% (all 3 mutations captured) 33.3% (only 1 successful capture)

[0093] As Figure 6 shown, in the technology of this application, only 0.1% of the data is lost due to sensor self - inspection, and the rest is completely recorded; in the prior art: 7.7% of the data is invalid due to mis - switching, over - range, response delay, etc.; the main body is a horizontal line of 99.9%, simulating 3 times of sensor self - inspections (each lasting 0.1 hour, and the integrity briefly drops to 99.5%), reflecting the characteristics of occasional but rapid recovery. The prior art starts from 98%, simulates 8 abnormal events through cumulative random decrease (each decrease is 0.5% - 1.0%), and finally drops to 92.3%, and adds normal distribution fluctuations to simulate the uncertainty in the actual working conditions.

[0094] Experimental results show that this technology, through a time-series feature-aware network and trend prediction model, achieves predictive switching. In low-speed boost scenarios, the range is switched 10 minutes in advance, avoiding the lag problem of existing technologies. It also suppresses unnecessary frequent switching in periodic fluctuation scenarios, reducing the number of switching times by over 95%. Existing technologies rely on fixed thresholds and are unable to adapt to pressure trend changes, resulting in a false switching rate as high as 38.9% (in mixed operating conditions).

[0095] The dynamic calibration parameter table and segmented correction mechanism keep the full-scale average error within ±0.035MPa, an improvement of over 60% compared to the existing technology (±0.089MPa). Especially in high-speed impact scenarios, the combined range measurement avoids over-range data loss, whereas the existing technology cannot switch in time, resulting in complete loss of peak data.

[0096] The emergency switching mechanism has a response time of only 80ms, a 75% improvement over the existing technology (320ms), and achieves 100% capture of sudden pressure changes. Adaptive sampling technology ensures data integrity while saving 65% of computing resources. Existing technologies suffer from frequent switching, resulting in a data loss rate of up to 7.7% and a significantly increased risk of sensor overload.

[0097] Through temperature-humidity compensation function and dynamic parameter optimization, the error fluctuation of the technology in this application is only ±0.005MPa in an environment with temperature fluctuation of ±5°C and humidity of ±10%RH, while the error of the existing technology fluctuates by more than ±0.03MPa due to environmental influences, showing stronger environmental robustness.

[0098] This embodiment describes in detail that this experiment is designed based on real industrial scenarios, and the data covers a variety of typical working conditions. It verifies the advanced nature of the technology of this application in the core links of wide-range automatic switching, dynamic error compensation, and sudden pressure response. Compared with the existing technology, it shows significant advantages in measurement accuracy, switching reliability, data integrity and other dimensions. It is particularly suitable for fields such as chemical industry, energy, aerospace, etc. that have strict requirements on pressure measurement accuracy and real-time performance, and can effectively reduce labor costs, improve production safety and data credibility.

[0099] The above are only preferred embodiments of the present invention, which do not limit the scope of protection of the present invention. For those skilled in the art, the present invention can be modified and varied in various ways. Any changes, modifications, replacements, integrations and parameter changes to these embodiments through conventional substitutions or that can achieve the same functions without departing from the principles and spirit of the present invention fall within the scope of protection of the present invention.

Claims

1. An automatic switching measurement method applicable to multiple ranges of pressure gauges, characterized in that, Including: S1. Obtain the current pressure measurement environment information, input it into a preset pressure range prediction model, and output the initial recommended range for pressure measurement in the current environment; S2. According to the initial recommended range, select the measurement parameters corresponding to the range, and perform a preliminary measurement of the pressure at the initial sampling frequency to obtain a preliminary measured pressure value; S3. Compare the preliminary measured pressure value with the upper and lower limits of the selected range. If the preliminary measured pressure value exceeds the preset maximum threshold of the selected range, it is determined that a higher range needs to be switched. If the preliminary measured pressure value is lower than the preset minimum threshold of the selected range, it is determined that a lower range needs to be switched; S4. Obtain the calibration coefficient of the range to be switched, and correct the measurement data after switching the range according to the calibration coefficient to obtain a corrected pressure measurement value; S5. Dynamically adjust the sampling frequency according to the relationship between the corrected pressure measurement value and the current range. If the measured value is close to the range boundary, increase the sampling frequency. If the measured value is in the middle area of the range, decrease the sampling frequency; S6. Monitor the pressure change rate in real time. When the pressure change rate exceeds the preset threshold, trigger an emergency range switching mechanism, and preferentially select a range that can cover the current pressure change range for measurement.

2. The automatic switching metering method applicable to multiple ranges of pressure gauges according to claim 1, wherein, The preset of the pressure range prediction model in S1 is specifically as follows: Obtain historical pressure measurement data, divide the data into a training set, a validation set, and a test set according to the time series, and label the actually applicable pressure range for each data sample; Construct a time series feature-aware network as a deep learning architecture, capture the temporal dependencies of pressure data and environmental information through a dynamic memory gating module, and assign adaptive weights to the features of different time steps according to the pressure change trend through a trend weight allocation module; Taking the pressure measurement environment information and the historical data of pressure fluctuations as inputs, with the cross-entropy loss function to perform model training, where N is the number of samples, M is the number of preset pressure range categories, and y ij is the true label that sample i belongs to range category j, is the probability that the model predicts sample i belongs to range category j. By adjusting the model parameters on the validation set to minimize the loss function value and verifying the generalization ability of the model on the test set, a pressure range prediction model is constructed.

3. The automatic switching metering method applicable to multiple ranges of pressure gauges according to claim 2, characterized in that, In S1, the temperature T, humidity H, and pressure fluctuation history data {P t-n , P t-n+1 , …, P t} in the current pressure measurement environment information are normalized to construct an input vector The X is input into a pre-set pressure range prediction model and processed by the dynamic memory gating module of the time series feature perception network to obtain a hidden state sequence H = [h1, h2, …, h T . The attention weights α i of each hidden state are calculated through the trend weight allocation module where w and W are trainable weight matrices respectively, and b is a bias vector; the hidden states are weighted and summed according to the attention weights to obtain a pressure continuous vector T r is the total number of time steps of the pressure fluctuation history data; finally, the pressure continuous vector c is input into a fully connected layer, and the prediction probabilities of each pressure range category are calculated through the Softmax function and the range corresponding to the category with the largest probability is taken as the initial recommended range for pressure measurement in the current environment, where v j is the weight vector of the fully connected layer corresponding to the j-th range category, and M is the number of pre-set pressure range categories.

4. The automatic switching metering method applicable to multiple ranges of pressure gauges according to claim 1, characterized in that, The method for selecting the measurement parameters corresponding to the range in S2 is as follows: establish a parameter database including the inherent parameters of the pressure gauges of each range, the environmental compensation parameters, and the dynamic adjustment parameters. The inherent parameters include the sensor sensitivity coefficient S i , the non-linear error correction matrix E i ; the environmental compensation parameters construct a compensation function according to the temperature T and the humidity H where a mn is the compensation coefficient, and the dynamic adjustment parameters set the sampling frequency adjustment factor according to the pressure change gradient γ of the initial recommended range λ is the adjustment coefficient. Index the parameter database according to the initial recommended range, and perform weighted fusion on the inherent parameters, environmental compensation parameters, and dynamic adjustment parameters to obtain the measurement parameter set 5. An automatic switching metering method applicable to multiple ranges of pressure gauges according to claim 1 or 4, characterized in that, The method for obtaining the preliminary measured pressure value in S2 is as follows: Using the sensor sensitivity coefficient in the selected measurement parameter set to perform a linear conversion on the original output signal V of the sensor raw to obtain the preliminary pressure value Perform polynomial fitting correction through the non-linear error correction matrix The correction formula is where is the coefficient of the i-th row and k-th column in r N is the fitting order, and combine the environmental compensation function C(T, H) to compensate the corrected pressure value to obtain the preliminary measured pressure value P pre = P pre2 ×(1 + C(T, H)).

6. The automatic switching metering method applicable to multiple ranges of pressure gauges according to claim 1, characterized in that, When comparing the preliminary measured pressure value with the upper and lower limits of the selected range in S3, a pressure change trend prediction model is constructed through a dynamic threshold adjustment mechanism to model the sequence of the most recent n preliminary measured pressure values and predict the pressure value at the next moment According to the predicted pressure value and the current upper range limit P max calculate the dynamic maximum threshold, and the formula is: where β s is the upper threshold adjustment coefficient. If the preliminary measured pressure value P pre is greater than the dynamic maximum threshold P dmax , it is determined that a switch to a higher range is required.

7. An automatic switching metering method applicable to multiple ranges of pressure gauges according to claim 1 or 6, characterized in that, In step S3, the pressure value at the next moment is predicted based on the pressure change trend prediction model for the sequence of the most recent n preliminary measurement pressure values to obtain the pressure value at the next moment According to the predicted pressure value and the current lower range limit P min calculate the dynamic minimum threshold where γ s is the lower threshold adjustment coefficient. If the preliminary measurement pressure value P pre is less than the dynamic minimum threshold P dmin , it is determined that a switch to a lower range is required.

8. An automatic switching metering method applicable to multiple ranges of pressure gauges according to claim 1, characterized in that, In S4, the calibration coefficient of the range to be switched is obtained for correction. Specifically, a dynamic calibration parameter table corresponding to each range is established, including the calibration offset ΔP for different temperature ranges and pressure segments cal and the scale factor K cal ; when switching to the target range, the corresponding temperature compensation range is queried according to the current ambient temperature T, and the calibration offset ΔP cal corresponding to the preliminary measured pressure value P cal and the scale factor K cal within this range are extracted; for the original measured value P raw after switching the range, the piecewise linear correction formula is adopted: where P mid is the intermediate pressure value of the target range, K cal and K′ cal , ΔP cal and ΔP′ cal are the calibration coefficients for the low-pressure segment and the high-pressure segment respectively.

9. The automatic switching metering method applicable to multiple ranges of pressure gauges according to claim 1 or 8, characterized in that, In S5, the sampling frequency is dynamically adjusted, specifically as follows: Set the effective measurement interval coefficient λ of the current range e , and the formula is: where P max and P min are the upper and lower limits of the current range. The sampling frequency f is dynamically adjusted through an exponential function model: where f base is the base sampling frequency, η is the boundary sensitivity coefficient, and μ is the smoothing factor; when P corr is close to the range boundary, the sampling frequency f is increased exponentially to f s , and the formula is: f s = f base ×(1 + η); when P corr is in the middle of the range, the sampling frequency remains f base .

10. The automatic switching metering method applicable to multiple ranges of pressure gauges according to claim 1, wherein, The emergency range switching mechanism in S6 is specifically as follows: Establish a pressure coverage feature matrix for each range. The matrix records the pressure response delay time, the minimum measurable pressure change amount, and the range overlap range of the corresponding range. When the pressure change rate exceeds the preset threshold, extract the current pressure value and the change trend data, and sequentially screen the ranges that meet the condition that the lower limit of the range is lower than the current pressure value and the upper limit of the range is higher than the predicted pressure change extreme value from the low range to the high range as candidates. If there are multiple candidate ranges, perform a priority ranking according to the pressure response delay time and the minimum measurable pressure change amount recorded in the matrix, and preferentially select the range with the shortest response delay time and that can meet the current pressure change resolution requirement; If all ranges cannot fully cover the predicted pressure change range, enable combined range measurement, perform parallel measurement through two adjacent ranges, and fuse the measurement data of the two ranges to generate a pressure measurement value.

Images