Debugging and checking method for pressure transmitter

Abstract

The invention relates to the technical field of industrial automation instrument and intelligent sensor detection, in particular to a pressure transmitter debugging and checking method. A parameter acquisition module, a thermal lag evolution module, an error calculation module and a pressure reconstruction module are included; the method comprises the following steps: acquiring temperature and pressure data in an unsteady state environment; the core of the method is to calculate the internal temperature by using a filling liquid thermal hysteresis model, solve errors caused by thermal imbalance and material hysteresis in combination with a transient pseudo pressure and viscoelastic hysteresis error model, and reconstruct real steady-state pressure; according to the method, the conversion from static waiting to dynamic prediction is realized, the heat balance waiting time is eliminated through prediction, namely verification, and the production efficiency is greatly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial automation instruments and intelligent sensor detection, in particular to a pressure transmitter debugging and verification method. BACKGROUND

[0002] In the production and measurement testing field of pressure transmitters, in order to ensure the measurement accuracy of the sensor, the equipment must be strictly debugged and verified and error corrected; the transmitter internally relies on the filling liquid and the metal diaphragm for pressure transmission, and its physical properties are extremely sensitive to environmental temperature changes and pressure loading rates; the existing verification scheme generally adopts a static verification mode based on thermal equilibrium, that is, the equipment is strictly required to be placed in a constant temperature environment until the internal temperature field is completely stable, and then data collection and calibration can be performed; although this scheme can obtain reference data in the steady state, it highly depends on the long physical standing process, resulting in a long verification cycle of a single device, which seriously restricts the production efficiency; in addition, this static mode cannot effectively identify the dynamic error characteristics in the non-steady state environment, cannot eliminate the transient pseudo pressure caused by the heat conduction lag of the filling liquid, and cannot compensate for the mechanical hysteresis error caused by the viscoelasticity of the metal material, resulting in the decline of the measurement performance of the equipment under actual fast temperature or pressure changes; Therefore, how to accurately decouple and compensate the thermal error and mechanical hysteresis error in the non-steady state environment in real time, so as to greatly shorten the verification time and improve the measurement accuracy of the equipment under dynamic working conditions, has become a technical problem to be solved. SUMMARY

[0003] To solve the above technical problems, the present application provides a pressure transmitter debugging and verification method, in particular, the technical scheme of the present application comprises: Step one, in a non-steady state verification environment, through the temperature measuring element integrated in the metal shell of the sensitive component of the pressure transmitter and the output port of the pressure transmitter, real-time collection of physical field parameters is performed, the physical field parameters include the real-time temperature of the shell and the observed pressure value at the current time, and a sampling time variable is defined; Step two, a filling liquid thermal lag evolution model is constructed, and the real-time temperature of the shell is used to calculate the average estimated temperature of the internal filling liquid through discrete iteration; Step three, a transient pseudo pressure model and a viscoelastic hysteresis error model are established, the real-time temperature of the shell, the average estimated temperature and the observed pressure value are input into the transient pseudo pressure model and the viscoelastic hysteresis error model, and the transient pseudo pressure value caused by thermal imbalance and the dynamic hysteresis pressure error value caused by material viscoelasticity are respectively calculated; Step four, a real pressure reconstruction model is constructed based on the signal superposition principle, the observed pressure value is subjected to multi-physical field error deduction by using the transient pseudo pressure value and the dynamic hysteresis pressure error value, and a predicted real steady pressure value is generated. Step five, get the standard reference true value given by the high-precision pressure controller, compare the real steady-state pressure value with the standard reference true value, construct the target loss function, and solve the model parameters by the nonlinear least squares optimization algorithm in reverse until the real steady-state pressure value converges, and output the qualified result.

[0004] Preferably, step one includes: S11, start the non-steady-state calibration environment of rapid temperature or pressure change to ensure that the pressure transmitter is in a dynamic response state; S12, collect the real-time temperature of the shell using the temperature measuring element, collect the observed pressure value directly output by the pressure transmitter using the data interface of the pressure transmitter, and record the corresponding sampling time variable; S13, input the collected real-time temperature of the shell and observed pressure value into the subsequent processing model as time series data.

[0005] Preferably, step two includes: S21, based on the lumped parameter method of heat transfer, determine that the rate of heat transfer from the shell to the internal filling liquid is proportional to the current temperature difference; S22, introduce a thermal response lag coefficient, which is determined by the specific heat capacity, density of the filling liquid, and the physical properties of the cavity geometry; S23, build a filling liquid thermal lag evolution calculation logic: the rate of change of the internal filling liquid temperature with time is equal to the product of the thermal response lag coefficient and the temperature difference, where the temperature difference is the difference between the real-time temperature of the shell and the average estimated temperature; S24, discretize the filling liquid thermal lag evolution calculation logic, and use the measurable real-time temperature of the shell to calculate the average estimated temperature which cannot be directly measured.

[0006] Preferably, the process of establishing a transient pseudo-pressure model in step three includes: S31, based on the thermodynamic state equation of the filling liquid, define the static thermal expansion difference pressure coefficient and the dynamic thermal shock factor; S32, calculate the shell temperature change rate, which is the derivative of the real-time temperature of the shell with respect to time; S33, combine the real-time temperature of the shell and the average estimated temperature to construct a transient pseudo-pressure calculation logic: the transient pseudo-pressure value is equal to the sum of the first component and the second component; Wherein, the first component represents the volume mismatch pressure caused by the temperature difference, and the value is equal to the static thermal expansion difference pressure coefficient multiplied by the difference between the real-time temperature of the shell and the average estimated temperature; The second component represents the fluid viscous damping effect and thermal shock stress, and the value is equal to the dynamic thermal shock factor multiplied by the shell temperature change rate.

[0007] Preferably, the process of establishing the viscoelastic hysteresis error model in step three comprises: S34, defining a pressure relaxation time constant corresponding to the ratio of the material viscosity coefficient to the elastic modulus based on the modified Kelvin-Voigt viscoelastic rheological mechanics model; S35, extracting the first derivative feature of the observed pressure value to calculate the real-time change rate of the observed pressure; S36, constructing viscoelastic hysteresis error calculation logic: dynamic hysteresis pressure error value equals to pressure relaxation time constant multiplied by real-time change rate of observed pressure.

[0008] Preferably, step four comprises: S41, taking the observed pressure value as the basic signal; S42, taking the transient pseudo-pressure value and the dynamic hysteresis pressure error value as error components; S43, constructing real pressure reconstruction calculation logic: predicted real steady-state pressure value equals to observed pressure value minus transient pseudo-pressure value, and then minus dynamic hysteresis pressure error value; S44, substituting the parameters in step three, and expanding to obtain comprehensive prediction logic, so as to predict the steady-state true value in advance when the physical system has not reached equilibrium.

[0009] Preferably, step five comprises: S51, defining a parameter vector to be optimized, the parameter vector consisting of thermal response lag coefficient, static thermal expansion difference pressure coefficient, dynamic thermal shock factor and pressure relaxation time constant, and setting a theoretical initial parameter vector based on physical prior knowledge; S52, obtaining the standard reference true value corresponding to each sampling time; S53, constructing an optimization objective function based on the nonlinear least squares method, the value of the optimization objective function consisting of two parts: the first part is the sum of squares of the differences between the estimated values calculated based on the current parameter vector and the standard reference true value at all sampling times, and is subjected to dimensionless normalization processing; the second part is a regularization term, the value of which is equal to the regularization coefficient multiplied by the weighted quadratic form of the difference between the current parameter vector and the theoretical initial parameter vector.

[0010] Preferably, step five further comprises: S54, using Levenberg-Marquardt algorithm to iteratively solve the optimization objective function; S55, updating the parameter vector in each iteration to calculate the current optimization objective function value; S56, presetting a convergence threshold, and judging whether the optimization objective function value is less than the convergence threshold, if so, determining that the parameter identification is completed, and locking the final individualized numerical model parameters.

[0011] Preferably, step five further comprises evaluation and output of the verification result: S57, calculating the residual error between the real steady-state pressure value at the final time and the standard reference true value; S58, when the residual error is less than the preset qualified precision limit value, determining that the verification result label is qualified, and writing the optimized parameter vector into the embedded microprocessor of the pressure transmitter; S59, when the residual error is greater than or equal to the qualified precision limit value, determining that the verification result label is unqualified, and generating a re-verification instruction or a hardware failure warning signal.

[0012] Preferably, after step S58, the pressure transmitter uses the written parameter vector and the real pressure reconstruction model to output corrected high-precision pressure data online in real time in subsequent operation according to the real-time collected shell real-time temperature and observed pressure value.

[0013] Compared with the prior art, the present application has the following beneficial effects: 1. The present method discards the traditional verification which must wait for the device to reach the static mode of thermal equilibrium, and proposes a debugging and verification method in a non-steady-state environment. By constructing a real pressure reconstruction model, the steady-state true value can be predicted in advance using an algorithm during the dynamic process when the physical system has not reached thermal equilibrium or mechanical equilibrium. This prediction and verification method does not require long-term physical static waiting, solves the problem of long verification time of a single device caused by dependence on thermal equilibrium, and greatly improves the production and manufacturing efficiency of the pressure transmitter. 2. The present method overcomes the difficulty that the internal fluid temperature cannot be directly measured by constructing a filling liquid thermal hysteresis evolution model and a transient pseudo-pressure model. The hysteresis internal fluid temperature is calculated using the measurable shell temperature, and the thermal error is further subdivided into a volume mismatch pressure component caused by temperature difference and a fluid viscous damping and thermal shock component caused by temperature change rate. This multi-physical field decoupling mechanism can effectively identify and eliminate the transient pseudo-pressure caused by filling liquid thermal conduction hysteresis, and significantly improves the measurement accuracy of the device under severe environmental temperature fluctuations. 3. For the material properties of the metal isolation diaphragm, the present method introduces a viscoelastic hysteresis error model. By extracting the first-order derivative feature of the observed pressure, the transient influence of the dynamic loading rate on the pressure reading is quantified, so as to accurately calculate the dynamic hysteresis pressure error value caused by the material viscoelasticity. This mechanism compensates for the defect that the existing static verification mode cannot identify dynamic mechanical errors, effectively solves the problem of measurement performance degradation caused by mechanical hysteresis under rapid pressure change conditions, and ensures the high fidelity of the transmitter under dynamic working conditions. 4. This method utilizes a global optimization algorithm to solve for model parameters in reverse based on standard reference truth values, completing the mapping from a general theoretical model to an individualized numerical model for each specific transmitter. By embedding the identified high-precision physical parameters into the embedded microprocessor of the pressure transmitter, the device is endowed with self-correction capabilities. In subsequent operation, the transmitter can eliminate environmental interference online in real time without relying on external calibration equipment, greatly improving the product's intelligence level and long-term operational stability. Attached Figure Description

[0014] The present invention will be further explained below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0016] Example 1: Please see Figure 1 A method for commissioning and verifying a pressure transmitter, the specific steps of which include: Step 1: Under unsteady-state calibration environment, physical field parameters are collected in real time through the temperature sensing element integrated in the metal housing of the pressure transmitter and the output port of the pressure transmitter. The physical field parameters include the real-time temperature of the housing and the observed pressure value at the current moment, and the sampling time variable is defined. Step 2: Construct a thermal hysteresis evolution model of the filling liquid. Using the real-time temperature of the shell, calculate the average estimated temperature of the internal filling liquid through discretization and iterative solution. Step 3: Establish the transient pseudo-pressure model and the viscoelastic hysteresis error model. Input the real-time shell temperature, average estimated temperature and observed pressure value into the transient pseudo-pressure model and the viscoelastic hysteresis error model, respectively, and calculate the transient pseudo-pressure value caused by thermal imbalance and the dynamic hysteresis pressure error value caused by material viscoelasticity. Step 4: Construct a real pressure reconstruction model based on the signal superposition principle, and use transient pseudo-pressure values ​​and dynamic hysteresis pressure error values ​​to perform multi-physics field error subtraction on the observed pressure values ​​to generate the predicted real steady-state pressure values; Step 5: Obtain the standard reference true value given by the high-precision pressure controller, compare the actual steady-state pressure value with the standard reference true value, construct the target loss function, and solve the model parameters in reverse using a nonlinear least squares optimization algorithm until the actual steady-state pressure value converges, and output the verification result.

[0017] The embodiment provides a pressure transmitter debugging and checking method, aims to solve the technical problems of long time consumption, low efficiency and difficulty in eliminating dynamic environment error of the existing static checking method; the method realizes high-precision checking of the pressure transmitter in a non-steady state environment by constructing a multi-physics field coupling model; Specifically, the method of the embodiment comprises the following core steps: Step one, in a non-steady state checking environment, a temperature measuring element integrated in a metal shell of the pressure transmitter is used to collect the real-time temperature of the shell, a pressure transmitter output port is used to collect an observed pressure value at the current time, and a sampling time variable is defined; Step two, a filled liquid thermal hysteresis evolution model is constructed, the real-time temperature of the shell is used as a boundary condition, and the average estimated temperature of the internal filled liquid is calculated by discrete iteration solution; Step three, a transient pseudo-pressure model and a viscoelastic hysteresis error model are established, the real-time temperature of the shell, the average estimated temperature and the observed pressure value are used as inputs, and the transient pseudo-pressure value caused by thermal imbalance and the dynamic hysteresis pressure error value caused by material viscoelasticity are respectively calculated; Step four, a real pressure reconstruction model is constructed based on the signal superposition principle, the transient pseudo-pressure value and the dynamic hysteresis pressure error value are used to deduct the multi-physics field error of the observed pressure value, and a predicted real steady-state pressure value is generated; Step five, a standard reference true value given by a high-precision pressure controller is obtained, the real steady-state pressure value is compared with the standard reference true value, a target loss function is constructed, and the model parameters are inversely solved by a nonlinear least squares optimization algorithm until the real steady-state pressure value converges, and a checking qualified result is output; The embodiment realizes accurate separation of thermal error and mechanical hysteresis error from observed data by establishing a thermal-flow-solid multi-physics field coupling model; compared with the traditional static checking which waits for system thermal equilibrium, the method allows checking in a non-steady state environment, predicts the steady-state true value by algorithm compensation, thereby greatly shortening the checking period of a single device and significantly improving the production efficiency.

[0018] Embodiment 2: Step one comprises: S11, a non-steady state checking environment of rapid temperature change or pressure change is started to ensure that the pressure transmitter is in a dynamic response state; S12, a temperature measuring element is used to collect the real-time temperature of the shell, a data interface of the pressure transmitter is used to collect the observed pressure value directly output, and a corresponding sampling time variable is recorded; S13, the collected real-time temperature of the shell and the observed pressure value are input into a subsequent processing model as time series data.

[0019] The embodiment is a specific embodiment of step one in embodiment 1. In step one, the non-steady state calibration environment refers to an environment where the pressure or temperature changes at a preset high rate, such as a rapid temperature changing oven or a rapid pressure changing platform; in this environment, the pressure transmitter is in a dynamic response state, and its output signal contains a significant dynamic error component; The data acquisition process is as follows: S11, start the non-steady state calibration environment of rapid temperature or pressure change, ensure that the pressure transmitter is in a dynamic response state; wherein the calibration environment is equipped with a high-frequency dynamic standard pressure sensor with a dynamic response frequency much higher than the measured pressure transmitter, to ensure that the standard reference true value obtained in the non-steady state environment has no lag; S12, use the temperature measuring element to collect the real-time temperature of the shell , at the same time, use the data interface of the pressure transmitter to collect the observed pressure value directly output , and record the corresponding sampling time variable ; S13, input the collected real-time temperature of the shell and the observed pressure value as time series data into the subsequent processing model to form a discretized data set ; By collecting data in a dynamic response state, the transient error characteristics that the pressure transmitter may encounter in actual working conditions can be captured, providing rich excitation signals for subsequent model parameter identification and ensuring the applicability of the calibration results in dynamic working conditions.

[0020] Embodiment 3: Step two includes: S21, based on the lumped parameter method of heat transfer, determine that the rate of heat transfer from the shell to the internal filling liquid is proportional to the current temperature difference; S22, introduce a thermal response lag coefficient, which is determined by the specific heat capacity, density and geometric structure of the filling liquid; S23, build a filling liquid thermal lag evolution calculation logic: the rate of change of the internal filling liquid temperature with time is equal to the product of the thermal response lag coefficient and the temperature difference, wherein the temperature difference is the difference between the real-time temperature of the shell and the average estimated temperature; S24, discretize the filling liquid thermal lag evolution calculation logic, and use the measurable real-time temperature of the shell to calculate the average estimated temperature which cannot be directly measured.

[0021] This embodiment is a specific implementation of step two in embodiment 1, focusing on solving the problem of the inability to directly measure the internal fluid temperature; In this embodiment, the filling liquid thermal lag evolution model is built based on the lumped parameter method of heat transfer; the core is to use the measurable shell temperature to calculate the lagged internal fluid temperature; S21, based on the heat transfer lumped parameter method, determine the rate of heat transfer from the shell to the internal filling liquid is proportional to the current temperature difference; S22, introduce the thermal response lag coefficient ; the thermal response lag coefficient refers to the physical quantity representing the rate of the internal system of the sensor reaching thermal equilibrium, and the unit is ; in this embodiment, its value is determined by the specific heat capacity, density of the filling liquid and the geometric structure of the sensor cavity, which reflects the size of thermal inertia; S23, build filling liquid thermal hysteresis evolution calculation logic, its mathematical expression is as follows: ; Among them: : the average estimated temperature of the internal filling liquid, the unit is , its value is obtained by model iteration calculation; : real-time temperature of the shell, the unit is , its value comes from the real-time acquisition of the temperature measuring element; : thermal response lag coefficient, the unit is , its value is preset in the initial stage, and is optimized by algorithm in the subsequent stage; : sampling time variable, the unit is ; S24, discretize the above differential equation, specifically using explicit Euler difference method, set the sampling period as , discretize the continuous time variable to sequence , the iteration formula is as follows: ; Among them, is the average estimated temperature of the internal filling liquid at the current time, is the estimated value at the last time, and the initial time is set to be equal to the ambient temperature; is the real-time temperature of the shell at the time; Use the measurable real-time temperature of the shell at the current time and the fluid temperature at the last time to calculate the average estimated temperature at the current time which cannot be directly measured ; This step establishes the dynamic correlation between the shell temperature and the internal fluid temperature; due to the large heat capacity of the internal insulating silicone oil and the poor thermal conductivity of the metal isolation diaphragm, the fluid temperature lags significantly behind the shell temperature; this model accurately quantifies this lag effect, providing an accurate temperature field benchmark for subsequent calculations of thermal expansion errors caused by temperature differences.

[0022] Embodiment 4: The process of establishing a transient pseudo-pressure model in step three includes: S31, based on the thermodynamic state equation of the filling liquid, define the static thermal expansion difference induced pressure coefficient and the dynamic thermal shock factor; S32, calculate the shell temperature change rate, which is the derivative of the real-time shell temperature with respect to time; S33, combine the real-time shell temperature and the average estimated temperature to construct the transient pseudo-pressure calculation logic: the transient pseudo-pressure value is equal to the sum of the first component and the second component; Wherein, the first component represents the volume mismatch pressure caused by temperature difference, the value is equal to the static thermal expansion difference induced pressure coefficient multiplied by the difference between the real-time shell temperature and the average estimated temperature; The second component represents the fluid viscous damping effect and thermal shock stress, the value is equal to the dynamic thermal shock factor multiplied by the shell temperature change rate.

[0023] This embodiment is a specific embodiment of the process of establishing a transient pseudo-pressure model in step three in embodiment 1; S31, based on the thermodynamic state equation of the fluid, define the following key parameters: Static thermal expansion difference induced pressure coefficient : refers to the physical quantity representing the pressure drift caused by the difference in expansion coefficient between the shell and the liquid under unit temperature difference, unit ; Dynamic thermal shock factor : refers to the physical quantity representing the influence of fluid viscous damping effect and thermal shock stress on pressure reading at the moment of rapid temperature change, unit ; S32, calculate the shell temperature change rate , which is the derivative of the real-time shell temperature with respect to time, in order to suppress the interference of sensor high-frequency noise on differential calculation, five-point moving average filter is used to smooth the original temperature sequence , and then the center difference method is used to calculate the temperature change rate: ; Wherein, is the smoothed shell temperature at time t, is the smoothed shell temperature at time t-1, is the smoothed shell temperature at time t+1, is the smoothed shell temperature at time t-2, For the shell temperature after smoothing processing at the moment; S33, combining the real-time temperature of the shell and the average estimated temperature , construct the transient pseudo-pressure calculation logic: ; Wherein: : the transient pseudo-pressure value caused by thermal imbalance, unit ; The first component : represents the volume mismatch pressure caused by temperature difference; when , the volume change of the metal cavity does not match the volume change of the silicone oil, resulting in additional internal pressure; The second component : represents the viscous damping effect and thermal shock stress of the filling liquid; This model successfully decouples the causes of dynamic thermal error; it not only considers the traditional static temperature drift, but also creatively introduces a dynamic thermal shock component related to the temperature change rate, so that it can restore the real physical field disturbance pressure with very high precision when the environmental temperature fluctuates dramatically.

[0024] Example 5: The process of establishing the viscoelastic hysteresis error model in step three includes: S34, based on the modified Kelvin-Voigt viscoelastic rheological mechanics model, define the pressure relaxation time constant, which corresponds to the ratio of the material viscosity coefficient to the elastic modulus; S35, extract the first derivative feature of the observed pressure value, and calculate the real-time change rate of the observed pressure; S36, construct the viscoelastic hysteresis error calculation logic: the dynamic hysteresis pressure error value is equal to the pressure relaxation time constant multiplied by the real-time change rate of the observed pressure.

[0025] This embodiment is a specific embodiment of the process of establishing the viscoelastic hysteresis error model in step three in example 1; S34, based on the modified Kelvin-Voigt viscoelastic rheological mechanics model, define the pressure relaxation time constant ; the pressure relaxation time constant refers to the characteristic time required for the internal lattice of the sensor diaphragm to recover from transient to steady state after being subjected to a step pressure shock, unit ; physically, it corresponds to the ratio of the material viscosity coefficient to the elastic modulus; S35, extract the first derivative feature of the observed pressure value, and calculate the real-time change rate of the observed pressure ; S36, construct the viscoelastic hysteresis error calculation logic, whose mathematical expression is as follows: ; wherein: : dynamic hysteresis pressure error value caused by material viscoelasticity, unit ; : real-time change rate of observed pressure, unit , derived from differential calculation of collected data; In the high-frequency calibration of rapid pressure rise and fall, the metal isolation diaphragm shows stress relaxation characteristics, resulting in return error; this model quantifies the transient influence of dynamic loading rate on pressure reading by introducing the first derivative term of pressure change rate, effectively eliminating the inherent mechanical hysteresis error of the sensor.

[0026] Embodiment 6: Step four includes: S41, taking the observed pressure value as the basic signal; S42, taking the transient pseudo-pressure value and the dynamic hysteresis pressure error value as error components; S43, constructing real pressure reconstruction calculation logic: the predicted real steady-state pressure value is equal to the observed pressure value minus the transient pseudo-pressure value, and then minus the dynamic hysteresis pressure error value; S44, substituting the parameters in step three, expanding to get the comprehensive prediction logic, so as to predict the steady-state true value in advance when the physical system has not reached equilibrium.

[0027] This embodiment is a specific implementation of step four in embodiment 1, which realizes the reconstruction of real pressure; S41-S42, taking the observed pressure value as the basic signal, and taking the calculated transient pseudo-pressure value and the dynamic hysteresis pressure error value as error components; S43, constructing real pressure reconstruction calculation logic: ; S44, substituting the parameters in step three, expanding to get the comprehensive prediction logic: ; wherein: : predicted real steady-state pressure value, unit ; This step uses signal superposition and error compensation theory to eliminate thermal error and mechanical hysteresis error through mathematical means when the physical system has not reached thermal equilibrium or mechanical equilibrium, predicts the steady-state true value of the system in advance, and realizes the efficient process of prediction and calibration.

[0028] Embodiment 7: Step five comprises: S51, defining a parameter vector to be optimized, the parameter vector being composed of a thermal response lag coefficient, a static thermal expansion difference pressure coefficient, a dynamic thermal shock factor and a pressure relaxation time constant, and setting a theoretical initial parameter vector based on physical prior knowledge; S52, obtaining a standard reference true value corresponding to each sampling time; S53, constructing an optimization objective function based on a nonlinear least squares method, the value of the optimization objective function being composed of two parts: the first part is the sum of squares of differences between estimated values calculated based on a current parameter vector and standard reference true values at all sampling times, and is subjected to dimensionless normalization processing; the second part is a regularization term, the value being equal to a regularization coefficient multiplied by a weighted quadratic form of differences between the current parameter vector and the theoretical initial parameter vector.

[0029] This embodiment is a specific implementation of the optimization objective construction in step five in embodiment 1; S51, defining a parameter vector to be optimized ; setting a theoretical initial parameter vector based on physical prior knowledge ; specifically, the theoretical initial value is calculated according to sensor design parameters: an initial thermal response coefficient ; wherein, is a natural convection heat transfer coefficient, is a contact area, is a specific heat capacity of silicone oil, is a mass of silicone oil; an initial thermal expansion coefficient is the product of a volume expansion coefficient of silicone oil and a diaphragm volume modulus; a regularization coefficient is selected by using an L-curve method, by calculating a logarithmic curve of a residual norm and a solution norm at different , a value of corresponding to a maximum curvature point of the curve is selected, so as to achieve the best balance between fitting accuracy and parameter physical constraints; S52, obtaining a standard reference true value corresponding to each sampling time ; S53, constructing an optimization objective function based on a nonlinear least squares method : Considering that the dimensions and numerical value levels of each physical quantity in the parameter vector are huge, in order to prevent large numerical value parameters from dominating the optimization process, a diagonal weighting matrix is introduced; the corrected objective function is: ; wherein, is the total number of sampling times, is a parameter normalization weight matrix, which eliminates the dimension effect by using the reciprocal square of initial value, ensuring that each parameter has the same sensitivity weight in the optimization space; is a characteristic pressure normalization constant, which is used to eliminate the physical dimension of the first part of the residual term, so that it can be added to the second part of the dimensionless parameter term; is a regularization coefficient; represents the sum of squares of the differences between the model estimated value and the standard reference true value at all sampling times, which is used to ensure the fitting accuracy; Due to the slight difference in oil filling amount of each transmitter and the processing tolerance, the physical parameters have discreteness; the objective function not only pursues the minimization of fitting error, but also constrains the optimization parameters not to deviate from the allowed range of physical meaning through the regularization term, ensuring the physical authenticity and robustness of the model parameters.

[0030] Embodiment 8: Step five further comprises: S54, using Levenberg-Marquardt algorithm to iteratively solve the optimization objective function; S55, updating the parameter vector in each iteration to calculate the current optimization objective function value; S56, presetting a convergence threshold, judging whether the optimization objective function value is less than the convergence threshold, if so, determining that the parameter identification is completed, and locking the final individualized numerical model parameters.

[0031] This embodiment is a specific embodiment of the solving algorithm in step five of embodiment 1; S54, using Levenberg-Marquardt (L-M) algorithm to iteratively solve the optimization objective function ; L-M algorithm combines the advantages of gradient descent method and Gauss-Newton method, and has the characteristics of fast convergence speed and strong robustness; S55-S56, updating the parameter vector in each iteration to calculate the current optimization objective function value; presetting a convergence threshold , judging whether it is true; if so, determining that the parameter identification is completed, and locking the final individualized numerical model parameters; Using the efficient optimization ability of L-M algorithm, the mapping from the general theoretical physical model to the individualized numerical model for each specific transmitter is completed, and the accurate compensation for individual differences is realized.

[0032] Embodiment 9: Step five further comprises evaluation and output of the verification result: S57, calculate the residual error between the real steady-state pressure value at the final time and the standard reference true value; S58, when the residual error is less than the preset qualified precision limit value, determine that the verification result label is qualified, and write the optimized parameter vector into the embedded microprocessor of the pressure transmitter; S59, when the residual error is greater than or equal to the qualified precision limit value, determine that the verification result label is unqualified, and generate a re-verification instruction or a hardware failure warning signal.

[0033] After step S58, in subsequent operation, the pressure transmitter uses the written parameter vector and the real pressure reconstruction model to output corrected high-precision pressure data online in real time only according to the real-time collected shell real-time temperature and observed pressure value.

[0034] The present embodiment describes the output and subsequent application of the verification result; S57-S59, calculate the residual error between the real steady-state pressure value at the final time and the standard reference true value; when the residual error is less than the preset qualified precision limit value, determine that the result is qualified, and write the optimized parameter vector into the EEPROM of the embedded microprocessor of the pressure transmitter; if the residual error is out of standard, generate a re-verification instruction or a hardware failure warning; In subsequent operation, the pressure transmitter no longer relies on external calibration equipment, but uses the written parameter vector and the built-in real pressure reconstruction model to output corrected high-precision pressure data online in real time only according to the real-time collected and . By solidifying the identified high-precision parameters into hardware, the pressure transmitter is given self-correction capability; no matter how the subsequent working environment changes, the transmitter can use the embedded model to offset environmental interference in real time, ensure the continuous output of high-precision pressure data, and greatly improve the intelligent level and environmental adaptability of the product.

[0035] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A method of pressure transmitter commissioning verification, the method comprising: The specific steps include: Step one, in the non-steady state verification environment, through the temperature measuring element integrated in the sensitive component metal shell of the pressure transmitter and the pressure transmitter output port, real-time acquisition of physical field parameters, physical field parameters include real-time shell temperature and current time observation pressure value, and definition of sampling time variable; Step two, build a filling liquid thermal hysteresis evolution model, use the real-time shell temperature, and calculate the average estimated temperature of the internal filling liquid by discrete iteration; Step three, establish a transient pseudo pressure model and a viscoelastic hysteresis error model, input the real-time shell temperature, average estimated temperature and observed pressure value into the transient pseudo pressure model and viscoelastic hysteresis error model, respectively, to calculate the transient pseudo pressure value caused by thermal imbalance and the dynamic hysteresis pressure error value caused by material viscoelasticity; Step four, based on the signal superposition principle, build a real pressure reconstruction model, use the transient pseudo pressure value and dynamic hysteresis pressure error value to deduct the observed pressure value from the multi-physical field error, and generate the predicted real steady pressure value; Step five, get the standard reference true value given by the high-precision pressure controller, compare the real steady pressure value with the standard reference true value, build a target loss function, and solve the model parameters by a nonlinear least squares optimization algorithm, until the real steady pressure value converges, and output the verification qualified result.

2. The method of claim 1, wherein: Step one includes: S11, start the non-steady state verification environment of rapid temperature or pressure change, and ensure that the pressure transmitter is in a dynamic response state; S12, use the temperature measuring element to collect the real-time shell temperature, use the data interface of the pressure transmitter to collect the directly output observed pressure value, and record the corresponding sampling time variable; S13, input the collected real-time shell temperature and observed pressure value into the subsequent processing model as time series data.

3. The method of claim 2, wherein: Step two includes: S21, based on the lumped parameter method of heat transfer, determine that the heat transfer rate from the shell to the internal filling liquid is proportional to the current temperature difference; S22, introduce a thermal response hysteresis coefficient, which is determined by the specific heat capacity, density and cavity geometry of the filling liquid; S23, build a filling liquid thermal hysteresis evolution calculation logic: the change rate of the internal filling liquid temperature with time is equal to the product of the thermal response hysteresis coefficient and the temperature difference, wherein the temperature difference is the difference between the real-time shell temperature and the average estimated temperature; S24, discretize the filling liquid thermal hysteresis evolution calculation logic, and use the measurable real-time shell temperature to calculate the average estimated temperature which cannot be directly measured.

4. The method of claim 3, wherein: The process of establishing a transient pseudo pressure model in step three includes: S31, based on the filling liquid thermodynamic state equation, define the static thermal expansion difference pressure coefficient and the dynamic thermal shock factor; S32, calculate the shell temperature change rate, which is the derivative of the real-time shell temperature with respect to time; S33, combine the real-time shell temperature and the average estimated temperature to build a transient pseudo pressure calculation logic: the transient pseudo pressure value is equal to the sum of the first component and the second component; Wherein, the first component represents the volume mismatch pressure caused by temperature difference, and the value is equal to the static thermal expansion difference pressure coefficient multiplied by the difference between the real-time shell temperature and the average estimated temperature; The second component represents fluid viscous damping effect and thermal shock stress, and is equal to the dynamic thermal shock factor multiplied by the rate of change of the shell temperature.

5. The method of claim 4, wherein: The process of establishing the viscoelastic hysteresis error model in step three includes: S34, based on the modified Kelvin-Voigt viscoelastic rheological mechanics model, define the pressure relaxation time constant, which corresponds to the ratio of the material viscosity coefficient to the elastic modulus; S35, extract the first derivative feature of the observed pressure value, and calculate the real-time change rate of the observed pressure; S36, construct the viscoelastic hysteresis error calculation logic: the dynamic hysteresis pressure error value is equal to the pressure relaxation time constant multiplied by the real-time change rate of the observed pressure.

6. The method of claim 5, wherein: Step four includes: S41, take the observed pressure value as the basic signal; S42, take the transient pseudo-pressure value and the dynamic hysteresis pressure error value as error components; S43, construct the real pressure reconstruction calculation logic: the predicted real steady-state pressure value is equal to the observed pressure value minus the transient pseudo-pressure value, and then minus the dynamic hysteresis pressure error value; S44, substitute the parameters in step three, and expand to obtain the comprehensive prediction logic, so as to predict the steady-state true value in advance when the physical system has not reached equilibrium.

7. The method of claim 6, wherein: Step five includes: S51, define the parameter vector to be optimized, which is composed of the thermal response lag coefficient, the static thermal expansion difference pressure coefficient, the dynamic thermal shock factor and the pressure relaxation time constant, and set the theoretical initial parameter vector based on physical prior knowledge; S52, obtain the standard reference true value corresponding to each sampling time; S53, based on the nonlinear least squares method, construct the optimization objective function, the value of which is composed of two parts: the first part is the sum of the squares of the differences between the estimated values calculated based on the current parameter vector and the standard reference true value at all sampling times, and is subjected to dimensionless normalization processing; the second part is the regularization term, the value of which is equal to the regularization coefficient multiplied by the weighted quadratic form of the difference between the current parameter vector and the theoretical initial parameter vector.

8. The method of claim 7, wherein: Step five also includes: S54, use the Levenberg-Marquardt algorithm to iteratively solve the optimization objective function; S55, update the parameter vector in each iteration to calculate the current optimization objective function value; S56, preset the convergence threshold, and determine whether the optimization objective function value is less than the convergence threshold, if so, determine that the parameter identification is completed, and lock the final individualized numerical model parameters.

9. The method of claim 8, wherein: Step five also includes evaluation and output of the verification result: S57, calculate the residual error between the real steady-state pressure value at the final time and the standard reference true value; S58, when the residual error is less than the preset qualified precision limit value, determine that the verification result label is qualified, and write the optimized parameter vector into the embedded microprocessor of the pressure transmitter; S59, when the residual error is greater than or equal to the qualified precision limit value, determine that the verification result label is unqualified, and generate a re-verification instruction or a hardware fault warning signal.

10. The pressure transmitter debugging verification method according to claim 9, characterized in that: After step S58, the pressure transmitter uses the written parameter vector and the real pressure reconstruction model in subsequent operation, and only according to the real-time collected shell real-time temperature and observed pressure value, online real-time output corrected high-precision pressure data.

Images