DE102022125355A1 - Procedure for testing an automation field device - Google Patents

Abstract

The invention comprises a method for checking a field device (FG1, FG2, FG3, FG4) of automation technology, wherein the field device (FG1, FG2, FG3, FG4) is connected to a communication network (KN), wherein the field device (FG1, FG2, FG3, FG4) receives and transmits data telegrams (DT) via the communication network (KN), wherein the field device (FG1, FG2, FG3, FG4) transmits the data telegrams (DT) in a predetermined communication frequency band, comprising:
- detecting the data telegrams (DT) transmitted by the field device (FG1, FG2, FG3, FG4) over at least a predetermined period of time;
- performing a Fourier transformation on the data telegrams (DT) acquired over the predetermined period of time to obtain a specific frequency spectrum (FS), wherein the Fourier transformation is performed at a sampling rate which is higher than the maximum frequency of the communication frequency band;
- comparing the specific frequency spectrum (FS) with a predetermined frequency spectrum; and
- Generating a notification in case of deviation of the specific frequency spectrum (FS) from the predetermined frequency spectrum.

Description

The invention relates to a method for checking a field device in automation technology, wherein the field device is connected to a communication network, wherein the field device receives and transmits data telegrams via the communication network, wherein the field device transmits the data telegrams in a predetermined communication frequency band.

Field devices that are used in industrial systems are already known from the prior art. Field devices are often used in process automation technology as well as in production automation technology. In principle, all devices that are used close to the process and that provide or process process-relevant information are referred to as field devices. Field devices are used to record and/or influence process variables. Measuring devices or sensors are used to record process variables. These are used, for example, for pressure and temperature measurement, conductivity measurement, flow measurement, pH measurement, level measurement, etc. and record the corresponding process variables pressure, temperature, conductivity, pH value, level, flow, etc. Actuators are used to influence process variables. These are, for example, pumps or valves that can influence the flow of a liquid in a pipe or the level in a container. In addition to the previously mentioned measuring devices and actuators, field devices also include remote I/Os, radio adapters or, in general, devices that are arranged at the field level.

In modern industrial plants, field devices are usually connected to higher-level units via communication networks such as field buses (Profibus ^® , Foundation ^® Fieldbus, HART ^® , etc.). The higher-level units are usually control systems (DCS) or control units such as a PLC (programmable logic controller). The higher-level units are used, among other things, for process control, process visualization, process monitoring and for commissioning the field devices.

The measured values recorded by the field devices, in particular by sensors, are transmitted via the respective bus system to one (or possibly several) higher-level unit(s). In addition, data transmission from the higher-level unit via the bus system to the field devices is also required, in particular for the configuration and parameterization of field devices and for controlling actuators.

• - Punkt zu Punkt-Topologie: Eine Systemkomponente verbunden mit einem Peripheriegerät (Feldgerät oder ebenfalls eine Systemkomponente)
• - Stern-Topologie: Eine Systemkomponente (Bsp.: Remote-IO, E/A Karte, Multiplexer) mit mehreren Peripheriegeräten
• - Ring-Topologie: Eine Ringanordnung mehrerer Geräte (sowohl System-, als auch Peripherie-Geräte)
• - Netz-Topologie: Hierarchische Netzstrukturen als Kombination der oben genannten Strukturen.

The field devices can be arranged differently depending on the network type. Common arrangements include:

• - Point-to-point topology: A system component connected to a peripheral device (field device or also a system component)
• - Star topology: A system component (e.g.: remote IO, I/O card, multiplexer) with several peripheral devices
• - Ring topology: A ring arrangement of several devices (both system and peripheral devices)
• - Network topology: Hierarchical network structures as a combination of the above structures.

The digital communication of the field devices must follow the relevant electrical engineering standards, which are specified by the respective fieldbus protocol. The devices are designed accordingly in electrical engineering terms and meet the respective standard within the permissible tolerances. Nowadays, there are strict requirements for the bus approval of a field device in connection with fieldbus systems. Central testing and certification bodies ensure compliance with the relevant standards.

Some fieldbus systems also define extensive protection mechanisms to protect a fieldbus system against unauthorized manipulation.

• - Manipulation der Datenleitungen zwischen den in der Anlage verbauten Feldgeräten und Netzwerkkomponenten;
• - Manipulation an der Transmitter- oder Sensor-Elektronik eines Feldgeräts;
• - Austausch des Transmitters oder Sensors eines Feldgeräts ohne Wissen des Anlagenbetreibers.

Industrial automation systems can be susceptible to manipulation and attacked in a variety of ways. Possible scenarios for such an attack include:

• - Manipulation of the data lines between the field devices and network components installed in the system;
• - Manipulation of the transmitter or sensor electronics of a field device;
• - Replacing the transmitter or sensor of a field device without the knowledge of the system operator.

The consequences can be diverse and serious. For example, denial of service (DoS) attacks on fieldbus systems can lead to longer downtimes of (partial) systems due to sporadic protocol violations. Manipulation of the data from field devices can lead to failures of (partial) systems but also lead to damage to the system equipment and danger to the personnel working in the system and the environment.

Based on this problem, the invention is based on the object of presenting a method which enables a field device to be checked for manipulation and/or malfunctions in a simple and safe manner.

• - Erfassen der von dem Feldgerät ausgesendeten Datentelegramme über zumindest einen vorbestimmten Zeitraum;
• - Durchführen einer Fourier-Transformation an den über den vorbestimmten Zeitraum erfassten Datentelegrammen zum Erlangen eines spezifischen Frequenzspektrums, wobei die Fourier-Transformation mit einer Abtastrate durchgeführt wird, welche höher ist als die maximale Frequenz des Kommunikationsfrequenzbands;
• - Vergleichen des spezifischen Frequenzspektrums mit mindestem einem vorbestimmten Frequenzspektrum; und
• - Erstellen einer Notifikation im Falle einer Abweichung des spezifischen Frequenzspektrums von dem mindestens einen vorbestimmten Frequenzspektrum.

The task is solved by a method for checking a field device of automation technology, the field device being connected to a communication network, the field device receiving and sending data telegrams via the communication network, the field device sending out the data telegrams in a predetermined communication frequency band, comprising:

• - Acquiring the data telegrams sent by the field device over at least a predetermined period of time;
• - performing a Fourier transformation on the data telegrams acquired over the predetermined period of time to obtain a specific frequency spectrum, the Fourier transformation being performed at a sampling rate which is higher than the maximum frequency of the communication frequency band;
• - Comparing the specific frequency spectrum with at least one predetermined frequency spectrum; and
• - Creating a notification in the event of a deviation of the specific frequency spectrum from the at least one predetermined frequency spectrum.

According to the invention, the field device sends out data telegrams via the communication network, which are designed in accordance with the protocol of the communication network. Such data telegrams are captured and recorded over a predetermined period of time. Specifically, the course of the voltage or current is recorded over time. Through a Fourier transformation of the recorded data telegrams, the time course of the voltage or current is converted into a specific frequency spectrum. This specific frequency spectrum contains the frequencies recorded in the predetermined period of time and their proportion.

Deviations in the field device or the network segment in which the field device is integrated can be detected via the subsequent analysis of the specific frequency spectrum with a predetermined frequency spectrum. In the frequency space, malfunctions and deviations can be detected that would not be detectable or would be difficult to detect by analyzing the course of the voltage or current in time space.

Using the method, it is not only possible to detect that there is a deviation, i.e. that an error on the field device or on the network segment is likely. It can also be concluded about the type of error present. For this purpose, the type of deviation is considered, i.e. in which frequency components there is a deviation and in what size, which can be specific for certain types of errors. For example, the specific frequency spectrum is compared with one or more other reference spectra, which are specific for certain types of errors.

• - ob die Elektronik des Feldgeräts in einem kritischen Alterungsstadium ist;
• - ob Manipulationen an der Physik des Anlagen- oder Kommunikationsnetzwerksegments vorgenommen wurden;
• - ob Feldgeräte getauscht wurden oder ein Feldgerät physikalisch verändert wurde (bspw. Elektronikmodule getauscht wurden);
• - welches Datentelegramm von welchem Feldgerät geschickt wurde (DoS-Angriffe lassen sich so direkt zuordnen) - Voraussetzung hierfür ist, dass mehrere Feldgeräte vorhanden sind, welche Datentelegramme aussenden.

The deviations suggest, for example, the following types of errors:

• - whether the electronics of the field device are in a critical stage of aging;
• - whether any manipulations have been made to the physics of the plant or communication network segment;
• - whether field devices were replaced or a field device was physically changed (e.g. electronic modules were replaced);
• - which data telegram was sent by which field device (DoS attacks can be assigned directly) - The prerequisite for this is that there are several field devices that send out data telegrams.

Field devices which are mentioned in connection with the invention have already been listed in the introductory part of the description.

One embodiment of the method provides that an AI or machine learning algorithm is used for the step of comparing the specific frequency spectrum with the predetermined frequency spectrum. Algorithms common in the state of the art can be used for this purpose, for example based on neural networks or deep learning.

One embodiment of the method provides that the AI or machine learning algorithm is taught with a large number of frequency spectrums of the field device and creates the predetermined frequency spectrum from this. For this purpose, several frequency spectra of the same field device type are used, which show a good, error-free condition of the field device. The reason for this is that every field device, even of the same device type the scatter in the components has its own characteristic. The characteristics differ minimally in internal resistance and impedance, which affects the quality of a transmitted data telegram. Transforming these data telegrams using Fourier analysis leads to slightly different frequency spectra, which are unique for each individual field device. By supplying the different frequency spectrums to the AI or machine learning algorithm, it is taught as broadly as possible so that it can recognize deviations from a good, error-free state as well as possible.

In a first variant of the method it is provided that a user provides feedback after creating a notification as to whether an undesirable deviation has occurred. For example, new, unknown deviations can be detected that trigger a notification but do not indicate an actual error in the field device and/or the network. This allows the Kl or machine learning algorithm to be further trained and improved.

In a second variant of the method, the AI or machine learning algorithm determines whether an undesirable change has occurred and feeds this back to itself as feedback. The AI or machine learning algorithm therefore learns independently.

An embodiment of the first variant or the second variant of the method provides that the AI or machine learning algorithm adjusts the predetermined frequency spectrum based on the feedback. By adjusting the reference spectrum, false-negative notifications can be reduced.

One embodiment of the method provides that the field device itself carries out the step of capturing the transmitted data telegrams. For this purpose, for example, a module is provided which is attached to the field device or connected to it. This has corresponding interfaces for receiving the transmitted data telegrams. Such a module can have its own resources, e.g. its own electronic module or its own storage unit, or it can use resources of the field device.

One embodiment of the method provides that another participant in the communication network, in particular another field device or a network device, carries out the step of detecting the transmitted data telegrams. Such a network device is, for example, a gateway or edge device connected to the communication network.

In an advantageous embodiment of the method, it is provided that the steps of carrying out the Fourier transformation and/or the comparison are carried out by a further network participant, in particular a local PC, wherein the recorded data are transmitted to the further network participant via the communication network.

The steps of capturing the transmitted data telegrams and performing the Fourier transformation and/or comparing them can of course also be carried out by a single device, for example by the field device itself or the other network participant.

In an alternative advantageous embodiment of the method it is provided that the steps of carrying out the Fourier transformation and/or the comparison are carried out by a cloud-based platform, with the recorded data being transmitted to the cloud-based platform via the Internet. A cloud-based platform (also simply called a “cloud”) is a server on which data can be sent and received via the Internet and on which one or more applications are located. These applications, which can contain, for example, the AI or machine learning algorithm, can be accessed by a user via the Internet.

One embodiment of the method provides that the field device sends out physical measured variables of a process engineering process and/or diagnostic data as data telegrams. It is known to those skilled in the art that a field device can generate a large number of different data, which can be sent out as the content of data telegrams as part of the method.

One embodiment of the method provides that an automation technology fieldbus is used as the communication network. In principle, any common protocol of a wired fieldbus network, in particular a fieldbus network for automation technology, such as Foundation Fieldbus ^® , Profibus ^® , Profinet ^® , HART ^® , Modbus ^® , etc. can be used. An alternative embodiment of the method provides that an Ethernet-based network, for example Industrial Ethernet or Profinet®, is used as the communication network.

• 1: eine erste Netzwerkanordnung zur Durchführung des erfindungsgemäßen Verfahrens;
• 2: Beispiele für ausgesendete Datentelegramme im zeitlichen Raum und ein daraus Fourier-transformiertes Frequenzspektrum; und
• 3: eine zweite Netzwerkanordnung zur Durchführung des erfindungsgemäßen Verfahrens.

The invention is explained in more detail using the following figures. It shows

• 1 : a first network arrangement for carrying out the method according to the invention;
• 2 : Examples of transmitted data telegrams in temporal space and a Fourier-transformed frequency spectrum from them; and
• 3 : a second network arrangement for carrying out the method according to the invention.

1 shows an arrangement consisting of a field device FG1 and an operating device BE. The field device is a device for detecting or influencing a physical quantity of a process engineering process. For example, the field device FG1 is a pressure measuring device. Further examples of field devices have already been listed in the introductory part of the description.

The operating device BE is, for example, a PC or laptop with a corresponding communication modem or a mobile operating unit. The operating device BE and the field device FG1 are connected to one another using second communication lines and form a communication network that uses the HART communication protocol. Using this protocol, the operating device BE sends requests to the field device FG1, the field device FG1 responds to the operating device BE using data telegrams DT. The field device FG1 can be set in such a way that it independently transmits data telegrams DT to the operating unit BE at regular intervals without the operating unit BE having to send out requests for this. The data telegrams DT contain data generated by the field device, for example process variables and/or status data.

The condition of a field device FG1 can deteriorate over time. For example, electronic components can gradually develop failures or change their characteristics. The condition of the communication network KN itself can also change, for example due to damage to the cables or corrosion on the contacts between the field device FG and the cables. Such changes in condition cannot always be recognized directly using the status displays of the field device or from the data generated by the field device.

For this reason, a further participant TN of the communication network KN is provided, for example in the form of a passive device (e.g. a switch with additional electronics), which is physically arranged between the field device FG1 and the operating unit BE. Alternatively, it can be provided that the operating unit BE or the field device FG1 itself carries out the functionalities of the further participant TN of the communication network KN described below.

The further participant TN records the data telegrams DT of the field device FG1 for a predetermined time interval.

In 2 The upper diagram shows a sequence of three data telegrams DT that were recorded over the predetermined time interval. The diagram shows the time course t on the abscissa and a current value I in mA on the ordinate. In the HART protocol, data telegrams are designed in such a way that they are modulated onto a base current value (4-20 mA standard), which indicates the current size of the primary process variable recorded by the field device FG1. A digital "1" is displayed with a frequency of 1.2 kHz (1200 Hz) and a "0" is displayed with a frequency of 2.2 kHz (2200 Hz). In the diagram at hand, no malfunctions can be seen in this representation.

The other participant TN then carries out a Fourier transformation of the data telegrams recorded in the predetermined time interval. Here, it carries out a fast Fourier transformation (FFT) algorithm or a similarly suitable algorithm. This transfers the recorded data telegrams DT from the temporal space to the frequency space. 2 The diagram below shows the specific frequency spectrum FS of the Fourier-transformed data telegrams DT. The ordinate represents a frequency band f, the abscissa shows the magnitude of a size of the Fourier-transformed data telegrams.

The specific frequency spectrum FS is unique to the field device FG1 and is therefore also referred to as a "fingerprint". If you carry out such a Fourier transformation on a digital signal from an electrical device with a sampling rate that is significantly higher than the communication frequency itself, you can identify significances in the frequency spectrum that are unique to the electronics. Every electronic device has its own fingerprint, so to speak.

The specific frequency spectrum FS is then transmitted to a cloud-based platform. An application is run on this platform that executes a Kl or machine learning algorithm. This Kl or machine learning algorithm compares the specific frequency spectrum with a predetermined frequency spectrum for deviations.

It must be taken into account that a fingerprint depends not only on the field device FG1 itself, but also on the overall impedance of the communication network KN. It is also possible that the fingerprint can change depending on internal device processes. For example, a sensor system of the field device FG1 could be in the limit load has an impact on the fingerprint. The fingerprint therefore has a certain dynamic that must first be learned. Kl or machine learning algorithm is therefore trained in advance for this dynamic using a large number of frequency spectrums, or is continually trained during operation (e.g. using feedback from the user) and can therefore make very precise statements after a while as to whether a Change actually took place or not.

In the present case, the AI detects deviations in the upper frequency range of the specific frequency spectrum. The frequency spectrum FS shows typical spikes in the range of 1200 Hz and 2200 Hz, which show the modulated “1” and “0” signals of the HART telegram. The frequency component below 1200 Hz shows the slower change of the 4-20 mA process variable. However, the frequency component above 2200 Hz shows an atypical curve compared to the reference spectrum. The Kl or machine learning algorithm detects this as a deviation and detects (by comparing it with other reference spectra that show typical error patterns) a gradual, age-related change in the electronics of the field device FG1. This is communicated to the user via a notification.

Instead of the cloud-based platform, the operating device BG, the other participant TN or the field device FG can also carry out the comparing step. The prerequisite for this is that these instances have sufficient resources to run and train the KL or machine learning algorithm.

In 3 Another application of the method is shown. Shown here is a communication network KN, which consists of several field devices FG1, FG2, FG3, FG4, a control unit SE, several workstation PCs PC1, PC2 and another participant TN in the form of an edge device or gateway. The communication network KN uses a modern fieldbus protocol (for example Foundation Fieldbus ^® or Profibus ^® ) or is designed to be Ethernet-based.

The further participant TN records data telegrams from one or more of the field devices FG1, FG2, FG3, FG4 and converts the data telegrams recorded at specific time intervals into specific frequency spectra using Fourier transformation. The Kl or machine learning algorithm implemented on the cloud-based platform CP examines these specific frequency spectrums for deviations. Not only can (hardware-related) errors in the field devices be detected - due to the uniqueness of the frequency spectrums, it can be detected whether a data telegram actually comes from the addressee contained in the telegram or whether there has been manipulation by a third party (e.g. through an exchange of field devices FG1 , FG2, FG3, FG4 or lines, or by introducing data telegrams from a third-party device). The data telegrams from the SE control unit can also be checked in this way.

List of reference symbols

BE

Control unit

C.P

cloud-based platform

DT

Data telegrams

FG1, FG2, FG3, FG4

Field devices

FS

Frequency spectrum

AI

Kl or Machine Learning Algorithm

CN

Communication network

PC1, PC2

Workstation PCs

SE

Control unit

TN

other participants in the communication network

Claims (13)

Method for checking a field device (FG1, FG2, FG3, FG4) of automation technology, wherein the field device (FG1, FG2, FG3, FG4) is connected to a communication network (KN), wherein the field device (FG1, FG2, FG3, FG4) receives and sends out data telegrams (DT) via the communication network (KN), the field device (FG1, FG2, FG3, FG4) sending out the data telegrams (DT) in a predetermined communication frequency band, comprising: - detecting the data from the field device (FG1, FG2, FG3, FG4) sent data telegrams (DT) over at least a predetermined period of time; - Carrying out a Fourier transformation on the data telegrams (DT) acquired over the predetermined period of time to obtain a specific frequency spectrum (FS), the Fourier transformation being carried out at a sampling rate which is higher than the maximum frequency of the communication frequency band; - Comparing the specific frequency spectrum (FS) with at least one predetermined frequency spectrum; and - Creation of a notification in the event of a deviation of the specific frequency spectrum (FS) from the at least one predetermined frequency spectrum. Procedure according to Claim 1 , wherein a Kl or machine learning algorithm (KI) is used for the step of comparing the specific frequency spectrum (FS) with the predetermined frequency spectrum. Procedure according to Claim 2 , whereby the Kl or machine learning algorithm is taught with a large number of frequency spectrums of the field device (FG1, FG2, FG3, FG4) and creates the predetermined frequency spectrum from this. Procedure according to Claim 3 , where a user provides feedback after creating a notification as to whether an undesirable deviation has occurred. Procedure according to Claim 3 , where the Kl or machine learning algorithm determines whether an unwanted change has occurred and feeds this back to itself. Procedure according to Claim 4 or 5 , where the Kl or machine learning algorithm adjusts the predetermined frequency spectrum based on the feedback. Method according to one or more of the preceding claims, wherein the field device (FG1, FG2, FG3, FG4) itself carries out the step of detecting the transmitted data telegrams (DT). Method according to one or more of the preceding claims, wherein a further participant (TN) of the communication network (KN), in particular a further field device (FG1, FG2, FG3, FG4) or a network device, carries out the step of detecting the transmitted data telegrams (DT). . Procedure according to Claim 7 or 8th , wherein the steps of performing the Fourier transformation and/or comparing are carried out by a further participant (TN) of the communication network (KN), in particular a further field device or a local PC, wherein the recorded data are transmitted to the further network participant via the communication network (KN). Procedure according to Claim 7 or 8th , Cloud, wherein the steps of performing the Fourier transform and/or comparing are performed by a cloud-based platform (CP), wherein the collected data is transmitted to the cloud-based platform (CP) via the Internet. Method according to one or more of the preceding claims, wherein the field device (FG1, FG2, FG3, FG4) sends out physical measured variables of a process engineering process and/or diagnostic data as data telegrams (DT). Method according to one or more of the preceding claims, wherein an automation technology field bus is used as the communication network (KN). Method according to one or more of the preceding claims, wherein an Ethernet-based network is used as the communication network (CN).

Images