Method for parameterizing a process-automation field device

Abstract

The invention relates to a method for parameterizing a field device (1) which generates measured values in a process system (2) using specified parameter values. The measured values are stored and compared with associated target data. In the event of a deviation of the measured values, a parameter optimization process is started. Modified parameter values relating to the process of generating measured values by means of the field device (1) are determined by means of artificial intelligence (4) using the specified parameter values, the stored measured values, and target data.

Description

[0001] Procedure for parameterizing a field device in process automation

[0002] The invention relates to a method for parameterizing a field device in process automation.

[0003] It is well known in the art to use sensors and actuators in process plants to automate the processes taking place there. These sensors are also called field devices and are used to measure or monitor process variables such as level, flow, temperature, pH, viscosity, pressure, density, or humidity. Many field devices require adaptation to the specific application. This is achieved through parameterization or configuration of the field devices.

[0004] According to DE 102018 128 254 A1, parameterization is carried out depending on environmental conditions with a view to improving measurement performance. This is achieved through the application of artificial intelligence (KI). In DE 102018 133 164 B4, the user stores a criterion as a type of limit value for the application, and if exceeded, verification is carried out. A KI algorithm can also be used for this. DE 10 2020 102 863 A1 describes the verification of the basic parameterization of a component of a process plant. This process discovers patterns in the plant that indicate potential for optimization. EP 4 345 545 A1 relates to the control of a plant containing sensors. The sensor data is uploaded to a cloud by an evaluation module and compared there with assigned data. If deviations are found, a suggestion for an operating parameter can be made.DE 10 2019 208 865 A1 describes the monitoring of a rail vehicle. Measured values ​​are compared with a pattern, and deviations are signaled. DE 10 2020 109 858 A1 discloses that patterns are generated from measured data in a system using K1 and compared with reference patterns. DE 10 2020 111 934 A1 is dedicated to the monitoring of a turbomachine, with sensor values ​​being evaluated via K1. It remains unclear how the parameters of individual field devices can be configured during operation. Currently, a service technician must change the parameters based on their experience, perform a measurement with the field device, and make further changes if necessary based on the measured value obtained. This is therefore an iterative process.

[0005] The invention is based on the object of proposing a method for parameterizing a field device which allows optimization of the parameters as easily as possible even during use of the field device in a process plant.

[0006] The object is achieved by a method for parameterizing a field device of process automation, wherein the field device generates measured values ​​with predetermined parameter values ​​in a process plant, wherein the method comprises at least the following steps: that the measured values ​​generated by the field device are stored - preferably in a data memory - that at least some of the measured values ​​generated by the field device are compared with associated target data, that in the event that the comparison of the measured values ​​with the target data results in a deviation of the measured values ​​outside a predeterminable tolerance range, a parameter optimization is started, and that the parameter optimization comprises at least the following step: that on the basis of the predetermined parameter values,Based on the stored measured values ​​and target data, modified parameter values ​​are determined using artificial intelligence with a view to generating measured values ​​by the affected field device.

[0007] In the method according to the invention, the measured values ​​generated by the field device are saved so that they are available later for parameter optimization. The measurements are carried out using predetermined parameter values, which were set, for example, during installation of the field device. The parameter values ​​are also used in parameter optimization because they represent a starting point for optimization. At least some of the measured values ​​are compared with the associated target data to determine whether the measured values ​​agree with the target data within a tolerance range or whether there are deviations. The target data are therefore actual values ​​for the respective process variable for which the measured values ​​are determined. If there is at least one deviation between the target data and the measured values ​​as actual data, parameter optimization is started.In one embodiment, parameter optimization is only triggered if a predefined number of deviations occur over a predefined period of time. During parameter optimization, reference is made not only to the aforementioned data: specified parameter values ​​and stored measured values, but also to the target data. Artificial intelligence, e.g. using machine learning, is used for the optimization. The invention therefore detects whether the measured values ​​are no longer appropriate. This can be due, for example, to wear, deposit formation, or other aging processes. Another cause can be, for example, a special tank geometry or process-specific properties. If there is a deviation between the measured value and the target data, the parameters of the field device are optimized. The optimization is carried out towards stable measured value determination by the affected, individual field device.

[0008] One embodiment involves at least the parameter optimization being performed by artificial intelligence in a cloud. In this embodiment, a cloud infrastructure is maintained, via which – or rather the artificial intelligence implemented therein – the optimization of the field device's parameters is carried out. In addition, in one embodiment, the comparison between the target data and the measured values ​​also takes place in the cloud.

[0009] One embodiment provides that the comparison of the measured values ​​with the corresponding target data is carried out by a user on the process plant side. In this embodiment, the measured values ​​are compared with the target data known from the process in the area of ​​the process plant, e.g. in a factory. This is done by a user, who can be a person or, for example, a computer program. This embodiment has the advantage that the target data remains in the process plant and does not have to be transmitted externally. Therefore, in one embodiment, for example, only information about the deviations between the target data and the measured values ​​is used to optimize the parameters. One embodiment involves the target data associated with the measured values ​​being saved by a user on the process plant side - preferably in a data storage device.In this embodiment, the target data is stored and thus made available by a user—i.e., a person or a computer program—of the process plant. This occurs, for example, in a data storage device that also stores the stored measured values ​​and is available to a program representing the artificial intelligence. The data storage device in which the target data is stored can, for example, be implemented in a cloud, in which the optimization of the parameters is preferably also carried out via artificial intelligence. The data storage device is preferably also the one that stores the measured values ​​of the field device.

[0010] One embodiment provides that the modified parameter values ​​are determined based on target data that relate to a predefined period around the occurrence of the deviation in the measured values. In this embodiment, additional target data is available for optimizing the parameter values, which originate from a time window surrounding the at least one erroneous measured value. In one embodiment, this is the target data that follows the measured value in which the deviation was discovered.

[0011] One design involves training the artificial intelligence using data from the process plant.

[0012] A supplementary or alternative embodiment provides that the artificial intelligence is trained using data originating from at least one process plant that is comparable to the process plant.

[0013] According to one embodiment, it is provided that the (determined) modified parameter values ​​are written into the field device - preferably after release on the process plant side by a user.

[0014] Furthermore, the invention relates to a process plant in which the method is implemented. Furthermore, the invention relates to a cloud infrastructure that executes the method, particularly with regard to the parameters.

[0015] The invention is explained in more detail with reference to the following figure.

[0016] Fig. 1 shows a schematic of a process plant with a field device to be parameterized.

[0017] Fig. 1 schematically shows a process plant 2 in which the fill level of a medium 21 in a container 20 is to be measured. For this purpose, a field device 1 is installed, which may be, for example, a measuring device that determines the fill level via the propagation time of radar waves.

[0018] To monitor the measurements of field device 1, a second field device T in the form of a point level switch is installed here, for example. The measured values ​​of field device 1 are stored in a data storage device 3, which is implemented in a cloud.

[0019] A user 5 on the process plant 2 side compares the measured values ​​of the field device 1 with the target data, which, for example, result from the measurements of the second field device T or which, for example, result from the operation of an actuator (e.g., the performance of a pump (not shown here). If the user detects a deviation between the measured values ​​and the target data that lies outside a predefined tolerance range, they initiate an optimization of the parameters of the field device 1, which is carried out by an artificial intelligence 4. To trigger the optimization, the arrow from the user 5 to the cloud, which here, for example, hosts the artificial intelligence 4, is symbolically represented here.

[0020] In addition to activation, the user 5 transmits target data to the data storage 3. Depending on the design or application, the target data that follows the detection of the deviation is saved. However, this can also be target data for measured values ​​that preceded the time at which the measured value was detected as incorrect. The artificial intelligence 4 accesses the measured values ​​and target data stored in the data storage 3. Other data used relates to further measured data from the process plant 2 or to the process in which the field device 1 with the measuring function is involved. Furthermore, in the embodiment shown, relevant data from a comparable process plant 2' is also used. Based on the data and the parameter values ​​with which the measurements of the field device 4 were taken, the artificial intelligence 4 determines new parameter values ​​that are optimized for the application.

[0021] As indicated here by the double arrow, field device 1 sends its measured values ​​to the cloud and receives the optimized parameter values ​​from there—specifically from artificial intelligence 4. In one embodiment, overwriting the parameter values ​​in field device 1 is authorized by user 5.

[0022] List of reference symbols

[0023] 1 field device

[0024] T Reference field device 2 Process plant

[0025] 2' process plant

[0026] 3 data storage

[0027] 4 Artificial Intelligence

[0028] 5 users

[0029] 20 containers

[0030] 21 Medium

Claims

Patent claims 1. A method for parameterizing a field device (1) of process automation, wherein the field device (1) generates measured values ​​with predetermined parameter values ​​in a process plant (2), wherein the method comprises at least the following steps: that the measured values ​​generated by the field device (1) are stored - preferably in a data memory (3), that at least some of the measured values ​​generated by the field device (1) are compared with associated target data, that in the event that the comparison of the measured values ​​with the target data results in a deviation of the measured values ​​outside a predeterminable tolerance range, a parameter optimization is started, and that the parameter optimization comprises at least the following step: that on the basis of the predetermined parameter values,Based on the stored measured values ​​and on the basis of target data, modified parameter values ​​are determined using artificial intelligence (4) with a view to generating measured values ​​by the field device (1).

2. The method according to claim 1, wherein at least the parameter optimization is carried out by the artificial intelligence (4) in a cloud.

3. Method according to claim 1 or 2, wherein the comparison of the measured values ​​with the associated target data is carried out on the side of the process plant (2) by a user (5).

4. Method according to one of claims 1 to 3, wherein the target data associated with the measured values ​​are stored on the side of the process plant (2) by a user (5) - preferably in a data memory (3).

5. Method according to one of claims 1 to 4, wherein the modified parameter values ​​are determined on the basis of target data which refer to a predefined period of time around the occurrence of the deviation in the measured values.

6. The method according to any one of claims 1 to 5, wherein the artificial intelligence (4) is trained using data originating from the process plant (2).

7. The method according to any one of claims 1 to 6, wherein the artificial intelligence (4) is trained using data originating from at least one process plant (2') which is comparable to the process plant (2).

8. Method according to one of claims 1 to 7, wherein the modified parameter values ​​are written into the field device (1).

9. The method according to claim 8, wherein the modified parameter values ​​are written into the field device (1) after release on the side of the process plant (2) by a user (5).

Images