METHOD AND ARRANGEMENT FOR THE ANALYSIS OF A FIELD DEVICE

Description

[0001] The present invention lies in the field of computer-implemented inventions and relates to a method and an arrangement for analyzing a field device with a machine learning system.

[0002] Artificial intelligence is increasingly being used in industrial settings, particularly in the manufacturing and process industries. Artificial intelligence (abbreviated "AI") refers to a technology that enables computers—that is, electronic computing units—to mimic intelligence to a certain degree, using appropriate algorithms and infrastructure. This is achieved, for example, through machine learning systems.

[0003] Machine learning (abbreviated "ML") is a subfield of artificial intelligence in which algorithms, during their execution, learn from large amounts of information (data) using electronic processing units, enabling them to adapt and act autonomously to a corresponding degree. In other words, machine learning is about generating knowledge in the form of statistical models from information.

[0004] The so-called Industrial Internet of Things (IIoT) places particular demands on the management and administration of field devices in the manufacturing and process industries due to increasing complexity. The task of analysis—that is, the detection and, in particular, the prediction of the future operating state of such devices—is becoming increasingly important in the area of ​​plant asset management (PAM) of these facilities.

[0005] Machine learning methods are now used to analyze field devices in industrial environments, for example, due to increasingly powerful and cost-effective electronic computing units (computers). This allows for sufficiently probable predictions of the (future) functionality and thus the future operating state of the field device. In particular, this involves identifying potential, expected errors or failures of field devices.

[0006] Setting up, administering, and / or extending such analysis methods requires, for example, in-depth knowledge of the programming language (e.g., "Python") or program library (e.g., "PyTorch") used, as well as the structure of the algorithms, or at least components of algorithms, that underlie the machine learning system and are intended for execution in order to process information. This has the disadvantage, for example, that often only the respective application developers with corresponding knowledge of the algorithms and the underlying programming language have access to the machine learning system, leading to a limited scope of application.

[0007] The publication DE 10 2018 133 316 A1 teaches a method for operating a field device of automation technology and field device of automation technology.

[0008] It is therefore an object of the present invention to provide an improved method for analyzing a field device with a machine learning system, which is characterized in particular by simpler configuration, i.e., administration with regard to adaptation, maintenance, and extension of the machine learning system. Furthermore, it is an object of the present invention to provide an arrangement comprising a field device, an electronic processing unit, and preferably a user input unit, wherein the electronic processing unit is configured to carry out the improved method for analyzing a field device.

[0009] The problem is solved by the features of claims 1 and 9. Further embodiments and applications of the present invention will become apparent from the dependent claims and are explained in more detail in the following description with partial reference to the figures.

[0010] According to a first general point of view, the present invention relates to a method for analyzing a field device according to claim 1. CHANGED SHEET

[0011] The present invention provides, for example, a method for analyzing a field device with a machine learning system that is easy to configure and therefore optimized for administration.

[0012] The field device can be configured as an actuator or a sensor, or comprise an actuator or a sensor. Alternatively, the field device can be configured as an arrangement of multiple actuators or multiple sensors, or comprise an arrangement of multiple actuators or multiple sensors. For example, the field device could be an electrically / electronically controlled mechanical valve for controlling a medium in a circuit within a production process, or, for example, an electrically / electronically controlled optical sensor for detecting an object.

[0013] The method according to the present invention can preferably be implemented, at least partially, as a computer-implemented method and executed in a runtime environment of an electronic computing unit. The electronic computing unit can, for example, be configured according to the von Neumann architecture and comprise at least one arithmetic logic unit (ALU, processor), a control unit, a bus system, an input / output unit, and other components and elements. The ALU can be configured to execute at least one algorithm and / or preferably comprise several processing cores. In other words, the electronic computing unit can be designed as a digital computer and thus as a computer. It is possible for certain sections of the method to run or be executed, at least partially, in parallel and / or simultaneously.

[0014] The method according to the present invention is characterized, for example, in particular by the fact that criteria or settings for the execution of the method for analyzing the field device can be easily determined, i.e., defined or determined, by generating descriptive information and processing information, which influence the further execution of the method and at least partially control the at least one first algorithm and / or the at least one second algorithm.

[0015] The descriptive information can include at least one function based on template information. The template information can, for example, contain a predefined number of configurable analysis functions for analyzing the field device to a specific extent. For instance, the template information can also include details or provide functions specifying which parameters and / or properties of the field device are to be analyzed, how they are to be analyzed, and preferably, to what extent. Thus, the template information can contain a function or a function specification that, for example, relates to the procedure for analyzing the field device to determine a future wear state.

[0016] The descriptive information can therefore, for example, specify what type of analysis of the field device is to be performed by the procedure, thus defining an expected result. For instance, the field device can be analyzed in its entirety, as specified by the corresponding descriptive information. Alternatively, the descriptive information can specify that only certain or individual components and / or elements of the field device are to be analyzed, such as individual actuators or sensors. The analysis options therefore vary depending on the field device.

[0017] The descriptive information can, for example, include details for analyzing the field device, such as a function like "TASK: Determining future functional state" or "TASK: Determining future wear state." Furthermore, by generating the descriptive information, it is possible to specify, before executing further sections or processes of the procedure, which algorithm of the machine learning system should be selected, called, and / or executed during the field device analysis. However, no program syntax, i.e., no source code, needs to be included in the descriptive information. This allows the field device analysis procedure to be administered relatively easily, as described above.

[0018] Analogous to the descriptive information, the processing information can also be configured. The processing information can preferably contain further details and, in particular, define at least one first information source with information about the field device. Preferably, this first information source can contain information in the form of process data from the field device, which was determined by measurement and / or simulation, i.e., by calculation, and which preferably characterizes the field device in a normal or non-critical operating state. The information in this first information source can include, for example, temperature values, pressure values, acceleration values, switching times, etc., that occur on and / or in the field device.

[0019] The processing information may contain further details that, for example, further specify and / or supplement the information, i.e., at least one function, of the descriptive information. Details in the processing information may include, for example, "JOB: Analysis interval 30 seconds" and / or "JOB: Query parameters = temperature, pressure, switching time".

[0020] At least one initial parameter for the field device is a measurable property and / or at least a calculable property of the field device or a physical quantity relating to the field device in a specific operating state, such as an operating temperature, an operating pressure, an operating switching time, an acceleration behavior, etc.

[0021] The descriptive information and / or the template information and / or the processing information can be generated at least partially by means of or based on a query at a user input unit, preferably with a graphical user interface.

[0022] The description information, template information, and / or processing information can, for example, be at least partially syntax-free text information, preferably in UTF-8 format. The description information, template information, and / or processing information can each be stored as a separate file or saved together in a single file during generation. Preferably, the description information and / or template information can be represented, generated, and / or saved at least partially in JSON (JavaScript Object Notation) format.

[0023] The machine learning system can preferably be computer-implemented and comprise at least one first algorithm and / or at least one second algorithm. The first algorithm and / or the second algorithm can be configured to process information in a runtime environment of an electronic computing unit and generate experiences from this in the form of information, i.e., data.

[0024] In other words, the machine learning system can be configured to analyze, categorize, and / or process provided information in such a way that it is possible to predict future values—that is, parameters and / or properties—based on the analyzed, categorized, and / or processed information. Such values ​​can, in turn, represent information that, with a certain probability, allows or at least suggests a prediction about the future behavior of the field device.

[0025] The machine learning system is trained using information from at least one initial information source, enabling the at least one initial algorithm and / or at least one secondary algorithm to recognize relationships based on this training information, preferably with a corresponding probability. The at least initial information source contains measured parameters and / or properties of the field device in a specific operating state (e.g., temperature values, pressure values, etc.).

[0026] Based on the information from at least one initial information source, the machine learning system is trained to evaluate previously unknown or new information about the field device. This evaluation provides information about the future operational behavior of the field device, such as a future functional state and / or a future wear state, and will be included in the generated initial protocol information for the field device.

[0027] The method for analyzing a field device is thus characterized, for example, by significantly simplified administration and / or configuration. Important criteria for execution and operation can therefore preferably be defined by generating the descriptive information and / or the processing information prior to the further execution of the method, preferably of at least one first algorithm and / or at least one second algorithm.

[0028] The execution of the at least one first algorithm and / or the at least one second algorithm further comprises: processing the information of the at least one first information source and the at least one first parameter with information about the field device of at least one second information source, preferably with respect to the at least one first parameter, by the at least one first algorithm and / or the at least one second algorithm.

[0029] The second information source, as described above based on the first information source, contains measured parameters and / or properties of the field device, or parameters and / or properties of the field device yet to be measured. Information to be measured, meaning parameters and / or properties of the field device, can be queried and / or read from the field device using this method.

[0030] Processing can include categorizing the information from at least one first information source and the information from at least one second information source. For example, the first algorithm and / or the second algorithm can, based on the processed information from the first information source, identify further relationships between parameters and / or properties of the field device.

[0031] It is possible that the procedure includes: comparing at least one initial protocol information with a reference information, and generating at least one initial quality information if the at least one initial protocol information deviates from the reference information according to at least one criterion.

[0032] The reference information can, for example, contain parameters and / or properties of the field device that represent the maximum permissible operating behavior of the field device for a specific service life. At least one criterion can be a limit value, such as a temperature limit, or a limit range. At least one initial quality information can, for example, contain a further assessment of the future functional state of the field device. It is possible that at least one initial quality information is displayed on a graphical user interface of the user input unit.

[0033] Furthermore, the method according to the present invention may include: adapting the descriptive information and / or the processing information when at least one first quality information is available, based on the at least one first protocol information, in relation to the information from the at least one second information source.

[0034] For example, the processing information can specify a further, that is, at least a third, source of information containing information about the field device. This third source of information can, for example, contain information about the field device's current operating state, such as a critical operating state.

[0035] The method according to the present invention may comprise: executing the at least one first algorithm and / or the at least one second algorithm based on or depending on the adapted descriptive information and / or based on or depending on the adapted processing information; and generating at least one second protocol information for the field device by the at least one first algorithm and / or by the at least one second algorithm; and comparing the at least one second protocol information with the reference information and generating at least one second quality information if the at least one second protocol information deviates from the reference information according to the at least one criterion.

[0036] According to another aspect of the present invention, generating the descriptive information may include: defining an identifier of the descriptive information to define an execution interval; and / or defining at least one parameter for the at least one function for processing by the at least one first algorithm and / or by the at least one second algorithm.

[0037] According to a further aspect of the present invention, the descriptive information, preferably the at least one function, can be configured, at least partially, to control the at least one first algorithm and / or the at least one second algorithm and preferably comprises at least two functions, wherein the at least one function of the at least two functions comprises at least one of the following: establishing a communication link to the field device; establishing at least one information source with information about the field device; establishing querying the information from the at least one information source; establishing preprocessing the information from the at least one information source; establishing adapting the information from the at least one information source; establishing postprocessing the information from the at least one information source;and / or specifying the storage of information from at least one information source. Specifying here can be understood, as described above, as defining or determining.

[0038] According to a further aspect of the invention, the method may include: visualizing at least the generated at least one first protocol information and / or preferably at least one second protocol information and / or preferably the generated at least one first and / or at least one second quality information, preferably on a graphical user interface of the user input unit.

[0039] The method according to the present invention may include: parameterizing at least the descriptive information and / or the processing information via the user input unit, preferably via a graphical user interface of the user input unit. This allows, for example, at least the descriptive information to be changed, i.e., adapted and / or extended, in a simple and quick manner. For example, start and stop times for the method and / or at least one termination criterion for prematurely ending the method for analyzing the field device can be defined.

[0040] According to a further aspect of the invention, it can be provided that the at least one interface is designed as a REST interface and / or is designed for communication according to the HTTP protocol and / or the method is executed at least partially in a batch mode with at least one generated first description information and at least one generated second description information, which is at least partially different from the generated first description information.

[0041] According to a second general point of view, the invention relates to an arrangement comprising a field device, an electronic computing unit and preferably a user input unit, wherein the electronic computing unit is configured to at least partially execute the method as disclosed herein.

[0042] To avoid repetition, features relating solely to the process and / or disclosed in connection therewith shall also be deemed to be disclosed and claimable as belonging to the device, and vice versa.

[0043] The embodiments and features of the present invention described above can be combined in any way desired. Further details and advantageous effects of the present invention are explained in more detail below with reference to the accompanying figures.

[0044] They show: Fig. 1 a flowchart of an example of the method according to the present invention; Fig. 2 a schematic view of an example of an arrangement according to the present invention.

[0045] Identical or functionally equivalent components or elements are identified in the figures with the same reference numerals. To avoid repetition, reference is sometimes made to the descriptions of other embodiments and / or figures for their explanation.

[0046] The following detailed description of the embodiments shown in the figures serves to illustrate or clarify the invention in no way limiting its scope.

[0047] Figure 1 The diagram shows a schematic representation of a process, that is, a flowchart of an example of the method according to the present invention.

[0048] The method according to the present invention serves for analysis, that is, for the analysis of a field device 10 (see also Figure 2Preferably, the method according to the present invention serves to analyze a current operating state of the field device 10 with regard to at least one parameter and / or at least one property of the field device 10 and to predict a future operating state of the field device 10, among other things based on the analyzed current operating state of the field device 10, preferably by means of or including a user input unit 30 and an electronic computing unit 40, which is or can be connected to the field device 10 via at least one interface 20 in order to preferably form a communication link for the transmission of signals.

[0049] A field device 10 can be configured as an actuator, preferably as a control element or as a valve. Alternatively, the field device 10 can be configured as a sensor, preferably as a transmitter. A field device 10 can be used in the manufacturing and process industries, i.e., in the field of automation technology, and can be connected via a network and thus a bus system, for example Industrial Ethernet, to other devices and equipment connected to the network.

[0050] For the purpose of transmitting information, for example for control and / or regulation, the field device 10 is preferably configured to transmit and / or receive electrical signals.

[0051] To further illustrate the described example of the method according to the present invention, the field device 10 (see below) Figure 2) described, for example, as a valve for controlling and / or regulating a process medium, such as water, in a plant.

[0052] The field device 10 therefore comprises a throttling element in the form of a valve for regulating the volume flow and / or pressure of the process medium. Furthermore, the field device comprises at least one electronic component and at least one sensor for detecting at least one parameter and / or at least one property that is related to at least one parameter (not shown in detail in the figures).

[0053] The at least one parameter can, for example, include a switching time, i.e., an operating switching time [t] of the field device 10, a temperature, i.e., an operating temperature [T] of the field device 10, and / or a pressure, i.e., an operating pressure [p] in the field device 10. It is possible that the at least one parameter includes further information relating to the field device 10 (e.g., accelerations, etc.). The at least one parameter can, for example, be a parameter measured or queried at a specific time.

[0054] The method according to the present invention is preferably at least partially a computer-implemented method and can be carried out on and / or by means of an electronic computing unit 40. The electronic computing unit 40 can, for example, be configured according to the von Neumann architecture and comprise at least one arithmetic unit (central processing unit, processor), a control unit, a bus system, an input / output unit, and other components and elements. The electronic computing unit 40 is preferably configured to process information in the form of electrical signals or pulses.

[0055] The electronic computing unit 40 can be located in the field device 10, i.e., contained therein, or connected to the field device 10, preferably via a communication link, i.e., via a signal link.

[0056] The following example of the method according to the present invention serves to analyze the field device 10 in order to determine and / or predict a current operating state and / or preferably a future operating state or future operating behavior and / or future functionality of the field device 10, in particular a potentially imminent failure of the field device 10.

[0057] This method can also be used to determine future expected wear characteristics. For example, the method according to the present invention can be used to determine whether the field device 10 has already reached a wear limit or will reach one in the future.

[0058] The procedure for analyzing the field device can be found in section S10 in Figure 1 for example via a user input unit 30 (see here) Figure 2The process is initiated by a user. For this purpose, the user input unit 30 can be connected to the field device 10 to be analyzed via at least one interface 20. This interface 20 can include at least one interface according to the so-called Representational State Transfer ("REST") specification.

[0059] Alternatively, the procedure described in section S10 can be started more or less automatically. In other words, the procedure is not started by a command manually entered by a user. Rather, the procedure for analyzing the field device 10 can be started based on predefined criteria, for example, by the electronic processing unit 40, which is connected to the field device 10 via a communication link, such as a bus system, within a network.

[0060] A first information source, i.e., storage device 51, can contain and / or store a database that contains information in the form of so-called process data about the field device 10 and makes this information available for retrieval. Process data can include or relate to information that represents the operating characteristics, i.e., the operating behavior of the field device 10 and thus the field device 10 in its operating state.

[0061] In the case of the field device 10 to be analyzed being configured as a valve, the database in the first storage device 51 may, for example, contain and / or store information on operating temperatures in degrees Celsius, operating pressures in bar, and operating switching times of the valve in seconds. The operating switching times of the valve 10 may depend on its operating temperature and / or operating pressure. Such information preferably represents parameters, and particularly preferably operating parameters or properties of the valve 10, that characterize its operating behavior in a given operating state. It is possible that the first storage device 51, as a primary information source, contains further information on the valve 10 in the form of process data.

[0062] After the procedure in section S10 has been started, a description information 61 is preferably generated in section S20. The description information 61 in turn comprises or includes at least one function based on a template information (not shown in the figures).

[0063] The template information can, for example, include a predefined number of configurable analysis functions for analyzing the field device 10 to a specific extent. For instance, the template information can also contain details or provide functions specifying which parameters and / or properties of the field device 10 are to be analyzed, how they are to be analyzed, and / or preferably, to what extent. Thus, the template information can contain a function or a function specification that, for example, relates to the procedure for analyzing the field device 10 to determine a future wear state. A parameter could, for example, include the operating temperature of the valve 10.

[0064] It is also possible that the template information includes a specification or a regulation for a standardization and / or restructuring of at least one parameter to be determined and / or of at least one property to be determined of the valve 10, and thus a function for processing the at least one parameter and / or the at least one property can be derived from it.

[0065] Additionally or alternatively, the template information may contain a specification or regulation for the standardization and / or restructuring of analysis results.

[0066] Additionally or alternatively, the template information can contain further details in the form of specifications or regulations to form a function for the descriptive information. Thus, the template information can include the selection and / or specification of an analysis method in a machine learning system, which is used in the course of analyzing the field device 10.

[0067] The machine learning system can be configured as an application of artificial intelligence, run on the electronic computing unit 40, and include at least one first algorithm and / or at least one second algorithm.

[0068] The machine learning system accesses the information from the first information source, that is, the database in the first storage facility 51.

[0069] The information contained in the database of the first storage device 51 includes training data for the field device 10 for the machine learning system. In other words, this information represents the operational behavior of the field device 10 as it behaves under normal operating conditions.

[0070] The machine learning system, that is, at least a first algorithm and / or at least a second algorithm of the machine learning system, can, for example, be a regression algorithm with corresponding processing instructions. In such a configuration, the machine learning system can determine a prediction of the future operating state of the valve 10 based on or depending on process data of the field device 10, which is contained and / or stored in the database in the first storage device 51.

[0071] For example, at least one of the following regression methods can be used, which is implemented by at least one first algorithm and / or by at least one second algorithm: Ordinary least squares regression (OLSR), linear regression, logistic regression, stepwise regression, multivariate adaptive regression splines (MARS) and / or locally estimated scatterplot smoothing (LOESS).

[0072] Alternatively or additionally, it is possible that the machine learning system, i.e., at least one first algorithm and / or at least one second algorithm, is designed and / or executed as a memory-based or instance-based algorithm.

[0073] The template information may additionally or alternatively include information on at least one interface through which the analysis results of the procedure are to be stored in a storage device, i.e., in a database (in Figure 2 (not shown).

[0074] The descriptive information 61 is preferably configured to select and / or call the at least one first algorithm of the machine learning system and / or the at least one second algorithm of the machine learning system; and / or to be processed by the at least one first algorithm of the machine learning system and / or by the at least one second algorithm of the machine learning system. The machine learning system can be configured as disclosed herein.

[0075] Generating the descriptive information preferably includes essentially a selection and / or specification of communication links for information, algorithms of the machine learning system with at least one first algorithm and / or at least one second algorithm, normalization procedures and / or restructuring procedures.

[0076] The functions contained in the descriptive information can also be referred to as "tasks" and thus as a task specification for the method for analyzing the field device 10. A task therefore preferably specifies the main task of the method according to the present invention and thus relates to the analysis of the field device 10, preferably the determination of a future operating state of the valve.

[0077] Generating the descriptive information involves creating a file containing the generated descriptive information and the relevant details. This file can be created and / or edited directly by a user via a user input unit 30 to configure the procedure. Alternatively, the descriptive information can be generated more or less automatically as a file using a software program, for example, one running on the electronic processing unit 10. Preferably, however, a user can interact with the software program via the user input unit 30.

[0078] The description file 61 may further contain information specifying which information from a second information source, stored in a second storage device 52, is processed for analysis together with the information from the first information source in the first storage device 51. The information from the second information source may contain or represent measured or measurable parameters and / or properties of the valve 10.

[0079] Now that the description information has been generated, which parameters and / or properties of the field device are to be determined during the analysis of the field device 10, a processing information 62 can now be generated in section S30.

[0080] The processing information 62 is based on and / or depends on the description information 61.

[0081] In the processing information 62, further criteria for the execution of the method according to the present invention are now defined - in analogy to the description information 61 - in the course of the further analysis of the valve 10.

[0082] For example, in processing information 62 a so-called "job" can be defined to monitor the field device, i.e. the valve 10, for example in the short term, for example with regard to defined parameters and / or properties of the valve 10 such as temperature, pressure and switching time.

[0083] Section S40 describes the selection of at least one first and / or at least one second algorithm of the machine learning system. This selection is based on or dependent upon the descriptive information and the processing information, including the respective details contained therein.

[0084] The selected and called-upon at least one first algorithm and / or at least one second algorithm now analyzes, categorizes and processes the information from the first information source from the first storage device 51 with the information from the second information source from the second storage device 52.

[0085] As mentioned above, the information from the first information source in the first storage device 51 represents training data for the machine learning system with respect to valve 10. The information from the second information source in the second storage device 52 represents current data, i.e., parameters and / or properties, with respect to valve 10. By processing this data using at least one first algorithm and / or at least one second algorithm, a future operating behavior of valve 10, which can be expected with a certain probability, is determined.

[0086] The expected future operating behavior of valve 10 is recorded by generating at least an initial log entry for valve 10. This initial log entry can preferably be stored as a file in a storage device.

[0087] The initial protocol information contains at least one expected profile of switching times for valve 10 at given operating temperatures and / or operating pressures.

[0088] The switching times of valve 10 are calculated and / or predicted based on the processed information from the first and second information sources by at least one first algorithm and / or by at least one second algorithm.

[0089] The initial protocol information can then be further processed in section S50, where the data it contains, which relates to the parameters and / or properties of valve 10, is compared with reference information. This reference information can, for example, contain parameters and / or properties of valve 10 that represent the maximum permissible operating behavior of valve 10 for a specific service life. If the initial protocol information deviates from the reference information according to at least one criterion, initial quality information can be generated. This criterion can be a limit value, such as a temperature limit, or a limit range. This initial quality information can, for example, include a further assessment of the future functional state of valve 10.

[0090] Finally, at least one initial protocol information and / or at least one initial quality information can be displayed in a section S60 on a graphical user interface of the user input unit 40 to give a user feedback on the analysis of valve 10.

[0091] Depending on the result, i.e. the information in the at least one first protocol information and / or in the at least one first quality information, further sections and / or processes of the method according to the present invention, as disclosed herein, may be carried out.

[0092] Figure 2Figure 1 shows a schematic view of an example of an arrangement according to the present invention with the field device to be analyzed, i.e. the valve 10, the interface 20, the user input unit 30, the electronic computing unit 40 for at least partial execution of the method according to the present invention, the first and second storage devices 51 and 52 with the information on the valve 10 as well as the description information 61 and the processing information 62, by means of which the method according to the present invention can be administered, preferably at least partially controlled, as described above. Reference symbol list

[0093] 10 Field device 20 Interface 30 User input unit 40 Electronic processing unit 51 First information source / first storage device 52 Second information source / second storage device 61 Description information 62 Processing information

Claims

1. Method for the analysis of a field device, preferably by means of a user input unit (30) which is connected to the field device (10) via at least one interface (20), wherein the field device (10) is embodied as a valve (10) for the control and / or regulation of a process medium in a plant, comprising: • Generating a description information (61) which comprises at least one function on the basis of a template information, wherein the template information contains a predefined number of configurable analysis functions for the analysis of the field device, wherein the description information (61) is configured to call at least one first algorithm of a machine learning system and / or at least one second algorithm of a machine learning system; and / or wherein the description information (61) is configured to be processed by at least one first algorithm of a machine learning system and / or by at least one second algorithm of a machine learning system, wherein generating the description information (61) comprises the creation of a file with the generated description information (61), wherein the creation of the file is created and / or edited directly by a user via a user input unit (30), so as to thereby configure the method, • Generating a processing information (62) on the basis of the description information (61), for the specification of at least one first information source (51) with information about the field device (10) and at least one first parameter of the field device (10) for processing by the at least one first algorithm and / or by the at least one second algorithm, wherein the information about the field device (10) represents training data for the machine learning system with respect to the valve (10), and wherein the at least one first parameter of the field device (10) comprises a measurable property and / or at least one computable property of the field device (10) or physical quantity relating to the field device (10) in a particular operating state, comprising at least one indication of operating temperature, operating pressure, operating switching time, or acceleration behaviour; • Executing the at least one first algorithm of the machine learning system and / or the at least one second algorithm of the machine learning system on the basis of the description information (61) and the processing information (62), wherein executing comprises: Processing the information of the at least one first information source (51) and of the at least one first parameter with information of the field device (10) of at least one second information source (52) by the at least one first algorithm and / or the at least one second algorithm, wherein the information of the at least one first information source (51) and of the at least one first parameter comprises training data which represent an operating behaviour of the field device (10) which represents the field device (10) in a normal operation, wherein the at least one first information source (51) is stored in a first storage device, and wherein the information about the field device (10) of the at least one second information source (52) contains or represents measured or to-be-measured parameters and / or properties of the field device (10), wherein the measured or to-be-measured parameters comprise at least one indication of operating temperature, operating pressure, operating switching time, or acceleration behaviour; wherein the at least one second information source (52) is stored in a second storage device; • Generating at least one first log information about the field device (10) by calling the at least one first algorithm and / or the at least one second algorithm on the basis of the processed information of the specified first information source (51) and / or of the specified first parameter, wherein the log information comprises an expected progression of switching times of the valve (10) at given operating temperatures and / or operating pressures, and wherein the progression of the switching times of the valve (10) is a calculated and / or predicted progression on the basis of the processed information of the first and the second information source (51, 52) by the at least one first algorithm and / or by the at least one second algorithm.

2. Method according to claim 1, comprising: • Comparing the at least one first log information with a reference information, and • Generating at least one first quality information upon deviation of the at least one first log information from the reference information according to at least one criterion.

3. Method according to claim 2, further comprising: • Adapting the description information (61) and / or the processing information (62) upon the presence of at least one first quality information on the basis of the at least one first log information with respect to the information of the at least one second information source (52).

4. Method according to any one of the preceding claims, wherein generating the description information (61) comprises: • Defining an identifier of the description information (61) for specifying an execution interval.

5. Method according to any one of the preceding claims, wherein the description information (61), preferably the at least one function, is configured at least partially to control the at least one first algorithm and / or the at least one second algorithm and preferably comprises at least two functions, wherein the at least one function of the at least two functions comprises at least one of the following: • Specifying a communication link to the field device (10); • Specifying at least one information source (51, 52) with information about the field device (10); • Specifying the querying of the information of the at least one information source (51, 52); • Specifying the pre-processing of the information of the at least one information source (51, 52); • Specifying the adapting of the information of the at least one information source (51, 52); • Specifying the post-processing of the information of the at least one information source (51, 52); and / or • Specifying the storing of the information of the at least one information source (51, 52).

6. Method according to any one of the preceding claims, comprising: • Visualising at least the generated at least one first log information and / or preferably at least one second log information and / or preferably the generated at least one first and / or at least one second quality information, preferably on a graphical user interface of the user input unit (30).

7. Method according to any one of the preceding claims, comprising: • Parameterising the description information (61) via the user input unit (30).

8. Method according to any one of the preceding claims, wherein the at least one interface (20) is embodied as a REST interface and / or is embodied for communication according to the HTTP protocol and / or wherein the method is at least partially executed in a batch mode with at least one generated first description information (61) and at least one generated second description information (61) which is at least partially different from the generated first description information (61).

9. Arrangement with a field device (10), an electronic computing unit (40) and a user input unit (30), wherein the electronic computing unit (40) is configured to execute the method according to any one of the preceding claims.