Method for detecting a signal calming of an output signal of a sensor

Abstract

The invention relates to a method for detecting a signal calming of an output signal (I) of a sensor (1), comprising the following steps: - applying the at least one chemical and / or physical process variable or a signal representing the chemical and / or physical process variable to the sensor (1) (S100); - sampling the output signal (I) of the sensor (1) n times, said output signal being output by the sensor (1) in response to the application, in order to obtain correspondingly sampled measurement values (S200); - determining gradients or difference values for the sampled measurement values (S300); - ascertaining a minimum or minima for the determined gradients or differential values (S500); - generating a regression model for the determined gradients or differential values and deriving at least one regression coefficient (S600); - checking and deciding on the basis of the determined minima and the at least one regression coefficient from the regression model whether the output signal (I) of the sensor (1) has settled around a setpoint value within a predefinable tolerance range and the output signal (I) has calmed (S700).

Description

[0001] Method for detecting signal stabilization of a sensor's output signal

[0002] The invention relates to a method for detecting a signal stabilization of an output signal of a sensor, in particular during a calibration and / or compensation process of the sensor, as well as a manufacturing device for calibrating and / or compensating a sensor.

[0003] Sensors, especially pressure sensors, are used, for example, in process and automation technology and serve to monitor and / or determine at least one process parameter of a medium, for example chemical and / or physical.

[0004] The process variable to be determined by the sensor can be, for example, pressure, fill level, flow rate, temperature, pH value, redox potential, or conductivity of the respective medium. The various possible measurement principles underlying the determination of the process variable are known from the prior art and will not be discussed further here. Sensors for measuring pressure are primarily designed as so-called absolute, relative, or differential pressure sensors.

[0005] Such sensors typically comprise a sensor unit that comes into contact with the process at least partially and / or temporarily, and an electronics unit that serves, for example, signal acquisition, signal processing, and / or signal input. The sensor's electronics unit is typically housed in a casing and also has at least one connection element for connecting the electronics unit to the sensor unit and / or an external unit. The connection element can be any electrical connection; a wireless connection is also possible. The electronics unit and the sensor unit can be designed as separate units with separate casings or as a single unit with a common casing.

[0006] Typically, the sensor unit does not directly determine and / or monitor the at least one process variable, but instead determines and / or monitors at least one measured variable from which the at least one process variable can be calculated by the electronic unit. In the case of pressure sensors, for example, a measuring diaphragm can be used, upon which the pressure of the process medium acts and which transmits the pressure to a piezoresistive element via a transmission medium. The effect of this pressure on the measuring diaphragm can be detected by piezoresistive resistors. These resistors are usually configured as a Wheatstone bridge to generate an output voltage, in particular a diagonal voltage, which ideally is linear to the pressure exerted on the measuring diaphragm. This output voltage is then further processed by the electronic unit into an output signal representing the measured values.

[0007] With dynamic signals like sensor output, it's usually necessary to wait until the signal has settled, meaning it has stabilized within a predefined tolerance range around a target value. Only then can the signal be measured. Typically, fixed waiting times are used for this purpose, based, for example, on empirical data.

[0008] The use of stabilized signals is particularly important for sensor calibration and / or compensation processes. Such calibration and / or compensation processes typically occur during sensor manufacturing. During calibration, the stabilized sensor reading is compared to a known reference value or standard to identify any deviations. However, the sensor itself is not adjusted in any way. Compensation, on the other hand, reduces measurement deviations caused by external factors such as temperature, humidity, etc.

[0009] The disadvantage of this is that, in order to ensure that the signal has calmed down, one waits too long.

[0010] The invention is based on the objective of proposing a way to reduce signal settling times.

[0011] The problem is solved according to the invention by the method according to claim 1 and the manufacturing equipment according to claim 11.

[0012] The inventive method for detecting a signal stabilization of an output signal of a sensor, in particular during a calibration and / or compensation process of the sensor, wherein the sensor is configured to determine measured values ​​of at least one chemical and / or physical process variable, provides the following steps:

[0013] Applying the sensor with at least one chemical and / or physical process variable or a signal representing the chemical and / or physical process variable; sampling the sensor's output signal, which the sensor outputs in response to the application, n times in order to obtain correspondingly sampled measured values;

[0014] Determining gradients or difference values ​​for the sampled measurements; identifying a minimum or minima for the determined gradients or difference values;

[0015] Generating a regression model for the determined gradients or difference values ​​and deriving at least one regression coefficient; checking and deciding, based on the determined minima or minimum and the at least one regression coefficient from the regression model, whether the sensor's output signal has settled within a predefinable tolerance range around a target value and whether the output signal has calmed down.

[0016] According to the invention, a method is proposed which dynamically detects the settling time of a sensor's output signal, thus eliminating the need to wait for fixed signal settling times (the time until the signal has settled). This offers the advantage of reducing signal settling times. This, in turn, offers the advantage of shortening calibration and / or compensation times during sensor manufacturing.

[0017] An advantageous embodiment of the method according to the invention can provide that a simulated input signal representing the chemical and / or physical process variable is applied to the sensor, in particular to an electronic unit of the sensor, so that the sensor outputs the output signal for the simulated input signal.

[0018] An alternative embodiment of the method according to the invention can provide that the sensor, in particular a sensor unit of the sensor, is exposed to the process variable, so that the sensor outputs the output signal for the process variable.

[0019] A further advantageous embodiment of the method according to the invention may provide that, in order to check and decide whether the sensor's output signal has stabilized within the predefinable tolerance range around the setpoint, the determined minima or minimum are compared with a threshold value for the minima or minimum, and the determined at least one regression coefficient is compared with a threshold value for the regression coefficient. In particular, the embodiment may provide that if the minima or minimum are below the threshold value for the minima or minimum, and the determined at least one regression coefficient is below the threshold value for the regression coefficient, it is assumed that the sensor's output signal has stabilized within the predefinable tolerance range around the setpoint.

[0020] Another advantageous embodiment of the method according to the invention can provide that the n-fold scanning is carried out equidistantly.

[0021] Another advantageous embodiment of the method according to the invention can provide that the n-fold sampling is carried out in such a way that the output signal of the sensor is sampled in the range of 2 to 10 times per second, preferably in the range of 4 to 6 times per second.

[0022] Another advantageous embodiment of the method according to the invention can provide that a linear regression model is generated for the determined gradients or difference values ​​and that a slope of a regression line is derived as the at least one regression coefficient.

[0023] Another advantageous embodiment of the method according to the invention may provide that, after determining the gradients or difference values ​​for the sampled measured values, a low-pass filtering of the determined gradients or difference values ​​is carried out. In particular, the embodiment may provide that the low-pass filtering is carried out using a Butterworth filter, in particular a first-order Butterworth filter.

[0024] The invention further relates to a manufacturing device for calibrating and / or compensating a sensor, in particular a pressure sensor, comprising at least one sensor unit configured to detect at least one chemical and / or physical process variable and an electronic unit configured to output a signal depending on the process variable detected by the sensor unit, comprising: a voltage source connected to the sensor and configured to supply voltage to the sensor; a measuring unit, in particular a voltage measuring unit, configured to detect the output signal of the sensor; and a manufacturing computer configured to execute the method according to one of the preceding claims.

[0025] An advantageous embodiment of the manufacturing device according to the invention can provide that the measuring unit comprises a voltage measuring device and a shunt resistor which is inserted into an output circuit of the sensor, wherein the voltage measuring device measures the voltage across the shunt resistor in order to derive the measured values.

[0026] A further advantageous embodiment of the manufacturing device according to the invention can further comprise the sensor with the at least one sensor unit which is configured to detect the at least one chemical and / or physical process variable and the electronic unit which is configured to output a signal depending on the process variable detected by the sensor unit.

[0027] The invention is explained in more detail with reference to the following drawings. They show:

[0028] Fig. 1: A manufacturing device for calibrating and / or compensating a sensor used to determine measured values ​​of at least one chemical and / or physical process parameter, and

[0029] Fig. 2: a flowchart of the inventive method for detecting a signal stabilization of an output signal of a sensor, in particular during a calibration and / or compensation process of the sensor.

[0030] Fig. 1 shows an example of a sensor 1 to be calibrated and / or compensated, which serves to determine measured values ​​of at least one chemical and / or physical process variable, as well as a manufacturing device for calibrating and / or compensating the sensor 1. The sensor 1 can, for example, be a pressure sensor. However, the invention is not limited to pressure sensors, but can also be transferred or applied accordingly to sensors that serve to determine measured values ​​of another chemical and / or physical process variable.

[0031] As mentioned above, the sensor 1 comprises a sensor unit 1.1 that comes into contact with the process at least partially and / or at least temporarily, and an electronic unit 1.2, which can serve for signal acquisition, signal evaluation, and / or signal input. The electronic unit 1.2 of the sensor can be arranged in a housing 1.3 and can additionally have at least one connection element for connecting the electronic unit to the sensor unit and / or an external unit. For example, the sensor can be configured to output the measured values ​​in the form of an analog current output signal I, in particular in the form of a 4-20 mA signal. However, the invention is not limited to sensors that output the measured values ​​in the form of an analog current signal, but can also be applied accordingly to sensors that output the measured values ​​in the form of a digital signal. The manufacturing device 2 comprises a voltage source 2.1, which serves to supply voltage to sensor 1, a multimeter 2.2 for determining a voltage, a shunt resistor 2.22 inserted into the output circuit 2.7 of sensor 1, across which the voltage is measured by means of the multimeter, and a manufacturing computer 2.4, which is set up to carry out procedures for calibrating and / or compensating sensor 1 which will be described in more detail below.

[0032] The multimeter 2.2 and the production computer 2.4 are connected via a data link, so that the voltage values ​​measured by the multimeter 2.2 are available via the production computer 2.4. The data interface or connection 2.5 can be, for example, a USB or Ethernet connection. The production computer 2.4 can also be connected to the sensor 1 to be calibrated and / or compensated via a data link 2.6. This connection can be, for example, a proprietary interface or connection.

[0033] The manufacturing computer 2 is configured to execute the method described below and shown in Fig. 2 for detecting a signal stabilization of an output signal I of a sensor 1, especially during a calibration and / or compensation process of the sensor 1.

[0034] In a first step, S100 stipulates that sensor 1, specifically sensor unit 1.1, is exposed to the chemical and / or physical process variable. For this purpose, sensor unit 1.1 is exposed to the chemical and / or physical process variable, so that it delivers a sensor signal to electronic unit 1.2 for further processing. The chemical and / or physical process variable can be, in particular, one of the aforementioned process variables: pressure, level, flow rate, temperature, pH value, redox potential, and / or conductivity of the respective medium. Alternatively, a signal representing the chemical and / or physical process variable can be applied directly to electronic unit 1.2 of sensor 1.

[0035] In a second step S200, following the first step, the output signal I of sensor 1, generated by the application of pressure to sensor 1, is sampled n times per second to obtain correspondingly sampled measured values. Sampling is preferably equidistant, so that the sampling occurs at regular time intervals. The sampling interval can range from a few tens to a few hundred milliseconds. For example, the sampling frequency can be 2 to 10, preferably 4 to 6, per second, so that the output signal is sampled 2 to 10, preferably 4 to 6 times per second.

[0036] In a third step, S300 then determines a gradient for each of the sampled measurements of the obtained measurement vector. Alternatively, a difference value can be determined for each pair of consecutive sampled measurements of the obtained measurement vector.

[0037] The gradients or difference values ​​determined in this way can be low-pass filtered in an optional fourth step, S400, following the third step, S300. For example, the gradients or difference values ​​can be low-pass filtered using a Butterworth filter. Low-pass filtering is particularly effective at removing superimposed noise. A first-order low-pass filter is preferably used to avoid excessive settling time.

[0038] In the fifth step S500, which either directly follows the step of gradient determination or difference value calculation S300 or alternatively the step of low-pass filtering S400, a mathematical minimum or mathematical minima of the determined and possibly filtered gradients or difference values ​​is determined, whereby in particular the minimum or minima of the absolute gradients or absolute difference values, i.e. the magnitude gradients or difference values, are determined.

[0039] In a sixth step (S600), a regression analysis is performed on the determined and, if necessary, filtered gradients or difference values, and a regression model is generated. At least one regression coefficient is then derived from the regression model. For example, a linear regression model can be created to approximate a straight line and thus obtain the slope of a regression line as the regression coefficient. Subsequently, the absolute regression coefficient (S601) is preferably determined.

[0040] The process steps five and six can, in principle, be interchanged in the execution order, so that, for example, the sixth process step S600 can also be executed before the fifth process step S500.

[0041] Subsequently, in a seventh process step S700, it is checked whether the output signal I of sensor 1 has stabilized within a predefinable tolerance range around a target value, based on the previously determined absolute minimum(s) of the determined and optionally filtered gradients or difference values ​​and the at least one absolute regression coefficient from the regression analysis. For this purpose, the minima or minimum, preferably determined in absolute terms, can be compared with a threshold value for the minima or minimum, and the at least one regression coefficient, preferably determined in absolute terms, can be compared with a threshold value for the regression coefficient.

[0042] If both values ​​are below their respective thresholds, the procedure assumes that the output signal I of sensor 1 has stabilized within a tolerance range around a target value, allowing the sampled measured values ​​to be output, transmitted, or processed. These output values ​​can then be used in a subsequent step for the actual calibration and / or compensation of sensor 1.

[0043] In the event that the minima or the minimum determined, preferably in absolute terms, are not below the threshold for the minima or for the minimum and / or the at least one regression coefficient determined, preferably in absolute terms, is not below the threshold for the regression coefficient, the method provides that the method is repeated from the second method step and the output signal I of the sensor is sampled n times again.

[0044] The predefinable tolerance range within which the output signal I of sensor 1 must lie in order to assume appropriate signal stabilization and to allow the sampled measured values ​​to be output, transmitted, or processed, can be selected depending on the desired application and / or the desired quality for the sensor.

[0045] The threshold values ​​for the minima or the minimum and the at least one regression coefficient can be determined based on statistical evaluations of sensors of the same type.

[0046] In addition, further statistical algorithms can be used to check and determine whether the output signal I of sensor 1 has stabilized within a predefined tolerance range around a target value and whether the output signal I has calmed down. In particular, averaging and / or calculating a standard deviation can be performed as further statistical algorithms. For this purpose, the sampled measured values, preferably the low-pass filtered sampled measured values, can be used. To check and determine whether the output signal I of sensor 1 has stabilized within a predefined tolerance range around a target value and whether the output signal I has calmed down, the mean value is additionally compared with a mean target value and / or the standard deviation with a standard deviation target value. That is, in addition to comparing the determined minima or...In addition to the minimum and at least one regression coefficient from the regression model, the mean and / or the standard deviation are also used to decide whether the output signal I of sensor 1 has settled within a predefinable tolerance range around a target value and whether the output signal I has calmed down.

[0047] Reference symbol list

[0048] 1 sensor

[0049] 1.1 Sensor unit

[0050] 1.2 Electronic unit

[0051] 1.3 Housing

[0052] 2 Manufacturing facility

[0053] 2.1 Voltage source

[0054] 2.2 Measuring unit, in particular voltage measuring unit

[0055] 2.21 Voltage measuring device, especially multimeter

[0056] 2.22 Shunt resistance

[0057] 2.4 Manufacturing computer

[0058] 2.5 Data-conducting connection between manufacturing computer and measuring unit

[0059] 2.6 Data-conducting connection between manufacturing computer and measuring unit

[0060] 2.7 Output circuit

[0061] S100 First process step

[0062] S200 Second process step

[0063] S300 Third procedural step

[0064] S400 Fourth process step

[0065] S500 Fifth process step

[0066] 5600 Sixth procedural step

[0067] 5601 Determining the absolute value of the regression coefficient

[0068] S602

[0069] I. Sensor output signal

Claims

Patent claims 1. Method for detecting a signal stabilization of an output signal (I) of a sensor (1), in particular during a calibration and / or compensation process of the sensor (1), wherein the sensor (1) is configured to determine measured values ​​of at least one chemical and / or physical process variable and the method comprises the following steps: Applying the sensor (1) with at least one chemical and / or physical process quantity or a signal representing the chemical and / or physical process quantity (S100); sampling the output signal (I) of the sensor (1) n times, which the sensor (1) outputs as a response to the application, in order to obtain correspondingly sampled measured values ​​(S200); Determining gradients or difference values ​​for the sampled measured values ​​(S300); Determining a minimum or minima for the specified gradients or difference values ​​(S500); Generating a regression model for the determined gradients or difference values ​​and deriving at least one regression coefficient (S600); Check and decide, based on the determined minima or minimum and at least one regression coefficient from the regression model, whether the output signal (I) of the sensor (1) has settled within a predefinable tolerance range around a target value and whether the output signal (I) has calmed down (S700).

2. Method according to claim 1, wherein a simulated input signal representing the chemical and / or physical process variable is applied to the sensor (1), in particular to an electronic unit (1 .2) of the sensor, such that the sensor (1) outputs the output signal (I) for the simulated input signal.

3. Method according to claim 1, wherein the sensor (1), in particular a sensor unit (1.1) of the sensor, is exposed to the process variable, so that the sensor (1) outputs the output signal (I) for the process variable.

4. Method according to one or more of the preceding claims, wherein, in order to check and decide whether the output signal (I) of the sensor (1) has settled within the predefinable tolerance range around the setpoint, the determined minima or the determined minimum are compared with a threshold value for the minima or for the minimum and the determined at least one regression coefficient is compared with a threshold value for the regression coefficient.

5. Method according to the preceding claim, wherein in the case that the minima or the minimum are below the threshold for the minima or the minimum and the determined at least one regression coefficient is below the threshold for the regression coefficient, it is assumed that the output signal (I) of the sensor (1) has settled within the predefinable tolerance range around the target value.

6. Method according to one or more of the preceding claims, wherein the n-fold scanning is carried out equidistantly.

7. Method according to one or more of the preceding claims, wherein the n-fold sampling is carried out such that the output signal (I) of the sensor (1) is sampled in the range of 2 to 10 times per second, preferably in the range of 4 to 6 times per second.

8. Method according to one or more of the preceding claims, wherein a linear regression model is generated for the determined gradients or difference values ​​and a slope of a regression line is derived as the at least one regression coefficient (S600).

9. Method according to one or more of the preceding claims, wherein after determining the gradients or difference values ​​for the sampled measured values, a low-pass filtering of the determined gradients or difference values ​​is carried out (S400).

10. Method according to the preceding claim, wherein the low-pass filtering is carried out using a Butterworth filter, in particular a first-order Butterworth filter.

11. Method according to one or more of the preceding claims, wherein further statistical algorithms (S602) are used to check and decide whether the output signal (I) of the sensor (1) has settled within a predefinable tolerance range around a setpoint value and whether the output signal (I) has calmed down.

12. Method according to one or more of the preceding claims, wherein further statistical algorithms include averaging and / or calculating a standard deviation, wherein the sampled measured values, preferably the low-pass filtered sampled measured values, are used for averaging and / or calculating the standard deviation, and wherein, to check and decide whether the output signal (I) of the sensor (1) has stabilized within a predefinable tolerance range around a setpoint value and the Once the output signal (I) has settled, the mean will be compared with a mean target value or the standard deviation with a standard deviation target value.

13. Manufacturing device (2) for calibrating and / or compensating a sensor (1), in particular a pressure sensor, comprising at least one sensor unit (1.1) configured to detect at least one chemical and / or physical process variable and an electronic unit (1.2) configured to output a signal (I) depending on the process variable detected by the sensor unit (1), comprising: a voltage source (2.1) that is connectable to or connected to the sensor (1) and configured to supply voltage to the sensor (1); a measuring unit (2.2), in particular a voltage measuring unit, configured to detect the output signal (I) of the sensor (1); a manufacturing computer configured to execute the method according to one of the preceding claims.

14. Manufacturing apparatus according to the preceding claim, wherein the measuring unit (2.2) comprises a voltage measuring device (2.21) and a shunt resistor (2.22) which is inserted into an output circuit of the sensor (1), and wherein the voltage measuring device (2.21) measures the voltage across the shunt resistor (2.22) in order to derive the measured values.

15. Manufacturing device according to one of claims 13 or 14, further comprising the sensor (1) with the at least one sensor unit (1.1) which is configured to detect the at least one chemical and / or physical process variable and the electronic unit (1.2) which is configured to output a signal (I) depending on the process variable detected by the sensor unit (1.1).

Images