Method for calibrating a measuring apparatus against a reference measuring apparatus, method for generating additional information concerning a measuring apparatus, and measuring apparatus that can be calibrated by the method

Abstract

A method for adjusting a measuring apparatus (M j ) that measures a measured variable of a medium (U) dependent on a state of a medium (U), comprising: ascertaining first measured values (xi) of the measured variable with the measuring apparatus (M j ); ascertaining second measured values (yi) of the measured variable with a reference measuring apparatus (R); generating data tuples (C j , i ) from the first measured values (xi) and the second measured values (yi) that have been acquired for the same state of the medium (U); carrying out a regression analysis (D) with the data tuples (C j , i ), and with a model function (E) with parameters (F); wherein parameter values (H j ) are calculated by means of the regression analysis (D); wherein the model function (E) maps the first measured values (xi) onto the second measured values (yi) by means of the parameter values (H j ); characterized in that the regression analysis (D) additionally includes prescribed parameter tuples (G); and / or in that only parameters (H j ) calculated by means of the data tuples (C j , i ) are compared with the prescribed parameter tuples (G) in the regression analysis (D).

Description

[0001] Method for comparing a measuring device with a reference measuring device, method for generating additional information about a measuring device, and measuring device that can be compared using the method. The invention relates to a method in process and automation technology for comparing a measuring device with a reference measuring device, including additional information about the measuring device. Furthermore, the invention relates to a method for generating additional information about a measuring device, and to a measuring device that can independently carry out the method for comparing it with a reference measuring device, including additional information. Measuring devices are used in process and automation technology, for example in the water and wastewater industry, and in particular to detect a fill level and / or to determine the properties of a medium.Such measuring devices can be actuator field devices, for example electronic engine or valve controls, or also sensor field devices, namely measuring devices such as Coriolis mass flow meters, vibronic density meters, magnetic-inductive flow meters, vortex flow meters, ultrasonic flow meters, thermal mass flow meters, pressure measuring devices, level measuring devices, vibronic level limit switches, temperature measuring devices, pH value measuring devices, etc., which are each used to determine measured values ​​representing a physical and / or chemical measured variable that varies over time within a predetermined measuring range and to generate at least one digital or analogue measured value signal that transmits the same measured values ​​externally to the respective measuring device.In the case of the above-mentioned measuring device, the measured variable to be recorded can be, for example, a mass flow, a density, a viscosity, a fill or limit level, a pressure, a pH value, an electrical conductivity, or a temperature of a liquid, powder, vapor, or gaseous medium that is conveyed or stored in a corresponding container, such as a pipeline or a tank. Furthermore, the applicant manufactures and markets a variety of measuring devices for determining dielectric properties using an electromagnetic signal in the microwave range, known from DE102017130728A1, which teaches a measuring device that determines a dielectric value of a medium by measuring a phase shift of an electromagnetic signal that occurs when electromagnetic waves with different frequencies propagate in a medium.Furthermore, the measuring device for measuring a dielectric value, known from DE102020134320A1, measures the permittivity of a fluid conveyed in a pipe, measuring a change in the amplitude and phase of a microwave signal that has passed through a container and is closely related to the dielectric value of the medium. Examples of measuring devices for determining a fill level are known from DE102012104858A1 and DE102013108490A1, which teach measuring devices that determine a distance to a surface of a medium or to another radar target in a pipe by generating frequency-modulated radar transmission signals that are reflected from the surface of a medium and received again. Properties such as the distance to the surface of the medium are determined from the properties of the received signal.Furthermore, the level measuring device known from DE102020134061A1 measures the level of filling materials in containers by emitting a radar signal in the high-frequency range and receiving the radar signal reflected by the filling material according to the time-of-flight principle. Furthermore, document DE10322083A1 teaches an ultrasonic measuring device with at least one ultrasonic transducer that transmits ultrasonic measurement signals via a membrane toward the surface of a medium in a container or receives the measurement signals reflected from the surface of the medium. A measured variable of the medium is determined based on the time-of-flight of the measurement signal. Measuring devices of all types are calibrated and compared with reference devices of known quality. This process is typically performed by a measuring device manufacturer in a controlled environment.However, it may be useful for the measuring device to be calibrated by a process operator at the process location. In this case, the process operator has a reference measuring device at their disposal which measures and displays measured variables with a known accuracy. For the calibration, at least one measurement of a medium in one state is carried out using both the measuring device and the reference measuring device, with the two measurements forming a data tuple. Often, several data tuples representing different states of the medium are created in this way. A simple linear regression is usually used to create a model function which maps the measured values ​​(influencing variable, x) measured by the measuring device to the measured values ​​(target variable, y) measured by the reference measuring device.In simple linear regression, a so-called regression line is fitted to a set of data tuples using two parameters in such a way that the linear relationship between the predictor and the target variable is described as accurately as possible. The equation of linear regression is given by ^^^^ = ^^0 + ^^1 ∗ ^^^^, where the model parameters to be determined (^^0, ^^1) = ^^ are the intercept and the slope of the regression line. A probability density suitable for fitting the regression line to the data tuples is, for example, a so-called likelihood function given as follows: The model parameters ^^ can be calculated, for example, by forming a log-likelihood function and then maximizing the following expression: The problem with this approach is that the above likelihood function, when fitting to the data tuples, includes model parameters ^^ with extreme values ​​and in particular negative values, for example due to statistical fluctuations in the measured variables, especially with small data sets. However, adjusting a measuring device with extreme and / or negative model parameters leads to incorrect measurement results and thus renders the calibration of the measuring device useless. The invention is based on the object of providing a method that avoids incorrect calibration of a measuring device with a reference measuring device and, if necessary, indicates insufficient and / or unsuitable measurement data. The invention solves this problem by a method according to independent claim 1. The inventive method for adjusting a measuring device measuring a measured variable of a medium, wherein the measured variable is dependent on a state of the medium,comprising at least the following steps: determining at least one first measured value of the measured variable with the measuring device; determining at least one second measured value of the measured variable with a reference device; generating at least one data tuple from the first measured values ​​and the second measured values ​​that were collected for the same state of the medium; performing a regression analysis using the data tuples and using a model function with parameters; wherein parameter values ​​are calculated by means of the regression analysis; wherein the model function maps the first measured values ​​to the second measured values ​​using the parameter values; characterized in thatthat in addition to the data tuples generated from the first measured values ​​and the second measured values, additional predefined parameter tuples are included in the regression analysis; and / or that in the regression analysis, parameters calculated exclusively using the data tuples are compared with the predefined parameter tuples. In one embodiment of the method according to the invention, the model function is a polynomial of the nth degree, with n<3. In one embodiment of the method according to the invention, a linear regression is performed; wherein the model function to be determined is a regression line,with an intercept and a slope. In one embodiment of the method according to the invention, one of the predetermined parameter tuples is formed from an expected value and a variance for an intercept of a regression line typical for the measuring device. In one embodiment of the method according to the invention, one of the predetermined parameter tuples is formed from an expected value and a variance for the slope of a regression line typical for the measuring device. In one embodiment of the method according to the invention, the regression analysis is carried out with only one data tuple; wherein for one of the parameter tuples predetermined for carrying out the regression analysis, preferably formed from an expected value and a variance of the intercept, the variance is multiplied by a factor of no more than 0.1, preferably no more than 0.01.In one embodiment of the method according to the invention, the measuring device is configured to determine a dielectric value of a medium by emitting a high-frequency signal; and wherein the measured variable to be determined is a solids content of the medium determined by means of the dielectric value. In one embodiment of the method according to the invention, ^^, ^^ measured values ​​​​measured with a measuring device and ^^ ^^ measured values ​​measured with a reference measuring instrument; where the model function is a regression line given by the equation^^^^ = ^^0 + ⋅ ^^^^ and with parameters given by a slope und einen Intercept ^^0; where the given parameters are the slope parameter with an expected value ^^^^^^^^^^^^ and a variance ^^^^^^^^^^^^, and the intercept with an expected value ^^ ^^^^^^^^ and a variance ^^ ^^^^^^^^Where the parameters to be calculated, including ^^0, ^^1, are calculated by maximizing the following expression: A method according to the invention for determining the predetermined data tuples for a plurality of measuring devices measuring a measurand of a medium, wherein the measurand is dependent on a state of the medium, comprising at least the following steps: performing a regression analysis for each measuring device from the plurality of measuring devices, comprising: determining first measured values ​​of the measurand with the measuring device; determining second measured values ​​of the measurand with a reference device; generating data tuples from the first measured values ​​and the second measured values ​​that were collected for the same state of the medium; calculating parameter values ​​using a model function with parameters, and using the first measured values ​​and the second measured values; determining an expected value and a variance for each type of parameter from the parameter values ​​for each measuring device.A measuring device according to the invention, which is adjusted using a method according to an inventive claim, comprises a measuring sensor; wherein the measuring sensor can be acted upon by a medium and is configured to generate at least one measurement signal dependent on the measured variable. An embodiment of the measuring device according to the invention further comprises a measuring and operating circuit; wherein the measuring and operating circuit is configured to receive the measurement signal and to determine the first measured values ​​based on the measurement signal; wherein the measuring and operating circuit is configured to receive and store the second measured values; and wherein the measuring and operating circuit is configured to perform the regression analysis based on the first measured values ​​and second measured values.An embodiment of the measuring device according to the invention is designed to determine a dielectric value of a medium by emitting a high-frequency signal; and wherein the measured variable to be determined is a solid content of the medium determined by means of the dielectric value. The invention has the advantage that, on the one hand, a process operator is guaranteed a comparison of a measuring device with plausible model parameters, and, on the other hand, a measuring device can indicate if the model parameters calculated from the available measured values ​​are not compatible with the additional information about the measuring device. A further advantage of this method is that it is easy to handle mechanically and can be calculated even on simple devices with little expenditure of computing capacity and energy. The invention is explained with reference to the following figures. It shows: Fig.Fig. 1: a flowchart of an embodiment of the inventive method for field calibration of a measuring device with additional information. Fig. 2: a flowchart of an embodiment of the inventive method for generating additional information for a measuring device. Fig. 3: a cross-section of an embodiment of the inventive measuring device. The embodiment of the inventive method shown in Fig. 1 comprises the method with a measuring device ^^. ^^ measured first values ​​^^ ^^ of a medium M, which together with the second measured values ​​measured by a reference measuring device 3 ^^ ^^ the data tuple The data tuples consisting of measured values ​​measured with a measuring device ^^ ^^ and with a reference measuring device 3 measured values ​​^^ ^^, are processed with a regression analysis D, whereby a regression line was chosen as the model function E in this embodiment to represent the relationship between the measured variables as follows: ^^^^ = ^^0 + ^^1 ⋅ ^^^^ , with associated parameters F, namely slope and intercept ^^0. Predefined parameters G included in the regression analysis D are an expected value ^^^^^^^^^^^^ for the slope parameter and a variance ^^^^^^^^^^^^ and an expected value ^^ for the intercept ^^^^^^^^ and a variance ^^ ^^^^^^^^ . The data tuples and the given parameters are combined to the following expression: Maximizing this expression yields a value for the desired model parameters ^^0and The model function E allows the two model parameters to be the measured values ​​^^ ^^ on the measured values ​​^^ ^^The embodiment of the method according to the invention for generating additional information of a measuring device shown in Fig.2 shows the multiplicity of measuring devices ^^ ^^ and a reference measuring device R generated multitude of data tuples ^^ ^^,^^ , which by regression analysis without additional information D' parameter values ​​^^ ^^ for each measuring device ^^ ^^ calculated. From the multitude of calculated parameter values ​​^^ ^^ is taken for each type of parameter value from the set of parameter values ​​^^ ^^ an expected value ^^ ^^ and a variance ^^ ^^calculated, for example by forming an average value. Fig. 3 shows a medium U in a container B, to which a measuring sensor A of an embodiment of the measuring device Mj according to the invention is applied, which has a measuring and operating circuit S with which the method according to the invention can be carried out. The reference measuring device R can be supplied with the medium U via its own measuring sensor A'. In order to ensure that the measuring device Mj and the reference measuring device R measure the medium in the same state, in this embodiment of the measuring device according to the invention a sample can be measured and taken from the measuring device Mj in order to then supply the reference measuring device R with it.

[0002] List of reference symbols U Medium ^^ ^^ Measuring device R Reference device A Sensor A' Sensor S Measuring and operating circuit B Container xi First measured values ​​yi Second measured values ​​^^ ^^,^^Data tuple D Regression analysis D' Regression analysis without additional information E Model function F Parameter G Specified parameters^^^^ Calculated parameter values^^ ^^ Expected value of a parameter ^^ ^^ Variance of a parameter

Claims

Patent claims 1. Method for adjusting a measuring device (^^) measuring a measured variable of a medium (U) ^^ ), wherein the measured variable is dependent on a state of the medium (U), comprising at least the following steps: ^ Determining at least one first measured value (xi) of the measured variable with the measuring device (^^ ^^ ); ^ Determining at least one second measured value (yi) of the measured quantity with a reference measuring device (R); ^ Generating at least one data tuple (^^ ^^,^^ ) from the first measured values ​​(xi) and the second measured values ​​(yi) collected for the same state of the medium (U); ^ Performing a regression analysis (D) using the data tuples (^^ ^^,^^ ), and using a model function (E) with parameters (F); ^ where by means of the regression analysis (D) parameter values ​​(^^ ^^ ) are calculated; ^ where the model function (E) is calculated using the parameter values ​​(^^ ^^) maps the first measured values ​​(xi) to the second measured values ​​(yi); ^ gekennzeichnet dadurch, ^ that in the regression analysis (D) in addition to the data tuples generated from the first measured values ​​(xi) and second measured values ​​(yi) (^^ ^^,^^ ) additionally predefined parameter tuples (G) are entered; ^ and / or that in the regression analysis (D) exclusively using the data tuples (^^ ^^,^^ ) calculated parameters (^^ ^^ ) are compared with the predetermined parameter tuples (G).

2. The method according to claim 1, wherein the model function (E) is a polynomial of the nth degree, with n<3.

3. The method according to one of claims 1 or 2, ^ wherein a linear regression is carried out; ^ wherein the model function (E) to be determined is a regression line with an intercept and a slope.

4. Method according to one of claims 1 to 3, ^ wherein one of the predetermined parameter tuples (G) is formed from an expected value and a variance for an intercept of a regression line typical for the measuring device.

5. Method according to one of claims 1 to 4, ^ wherein one of the predetermined parameter tuples (G) is formed from an expected value and a variance for the slope of a regression line typical for the measuring device.

6. Method according to one of claims 1 to 5, ^ wherein the regression analysis (D) is carried out with only one data tuple (^^ ^^,^^ ) is carried out; ^ wherein for one of the parameter tuples (G) specified for carrying out the regression analysis (D), preferably formed from an expected value and a variance of the intercept, the variance is multiplied by a factor of not more than 0.1, preferably not more than 0.

01.

7. Method according to one of claims 1 to 6, ^ wherein the measuring device (^^^^ ) is configured to determine a dielectric value of a medium (U) by emitting a high-frequency signal; and ^ wherein the measured variable to be determined is a solid content of the medium (U) determined by means of the dielectric value.

8. Method according to one of claims 1 to 7, ^ wherein the model function (E) is a regression line given by the G leichung ^^^^ = ^^0 + ⋅ ^^^^ und mit Parametern (F) gegeben durch eine Steigung and an intercept ^^0; ^ where the given parameters (G) are the slope parameter with an expected value ^^ ^^^^^^^^^^ and a variance ^^ ^^^^^^^^^^ , and the intercept with an expected value ^^ ^^^^^^^^ and a variance ^^ ^^^^^^^^ include; ^ Wobei die zu berechnenden Parameter (^^^^), umfassend ^^0, ^^1, berechnet werden maximizing the following expression:

9. Method for determining the predetermined data tuples (G) for a plurality of measuring devices (^^) measuring a measured variable of a medium (U). ^^), wherein the measured variable depends on a state of the medium (U), comprising at least the following steps: ^ Carrying out a regression analysis (D') for each measuring device (^^ ^^ ), comprising: ^ Determining first measured values ​​(xi) of the measured quantity with the measuring instrument (^^ ^^ ); ^ Determining second measured values ​​(yi) of the measured quantity with the reference measuring device (R); ^ Generating data tuples (^^ ^^,^^ ) from the first measured values ​​(xi) and the second measured values ​​(yi) collected for the same state of the medium (1); ^ Calculating parameter values ​​(^^ ^^ ) using a model function (E) with parameters (F), and using the first measured values ​​(xi) and the second measured values ​​(yi); determining an expected value (^^ ^^ ) and a variance (^^ ^^ ) for each type of parameter from the set of parameter values ​​(^^ ^^).

10. A measuring device which is adjusted using a method according to one of claims 1 to 8, comprising: ^ a measuring sensor (A); ^ wherein the measuring sensor (A) can be supplied with a medium (U) in a container (B) and is designed to generate at least one measuring signal which is dependent on the measured variable.

11. A measuring device according to claim 10, further comprising a measuring and operating circuit (S); ^ wherein the measuring and operating circuit (S) is designed to receive the measuring signal and to determine the first measured values ​​(xi) on the basis of the measuring signal; ^ wherein the measuring and operating circuit (S) is designed to receive and store the second measured values ​​(yi); ^ and wherein the measuring and operating circuit (S) is configured to perform the regression analysis (D) based on the first measured values ​​(xi) and second measured values ​​(yi).

12. Measuring device according to claim 10 or 11, ^ wherein the measuring device is configured to determine a dielectric value of a medium (U) by emitting a high-frequency signal; and ^ wherein the measured variable to be determined is a solids content of the medium (U) determined by means of the dielectric value.

Images