System for optimising the operation of an automation system, and method for optimising the operation of an automation system

Abstract

The invention relates to a system for modelling the dependency of process output variables on process variables of a process in an automation system, comprising: a server platform (SP) with a computer unit (CE) and a control unit (SE); first measuring devices (M1) for detecting measured values of the process variables (W1); second measuring devices (M2) for detecting measured values of the process output variables (W2); a closed communication network (KN) for connecting the measuring devices (M1, M2) and server platform (SP); wherein the first measuring devices (M1) detect measured values of the process variables (W1) and transmit them as input values (I) to the server platform (SP) by means of the closed communication network (KN); wherein the second measuring devices (M2) detect measured values of the process output variables (W2) and transmit them as output values (O) to the server platform (SP) by means of the closed communication network (KN); wherein the server platform (SP) stores transmitted input values (I) and output values (O) at least partially as time series (ZR); wherein the server platform (SP) trains an algorithm (ML) with stored time series (ZR) by means of computer unit (CE); wherein the server platform (SP) calculates values (WO) for the process output variables by means of the trained algorithm (ML) on the basis of the transmitted input values (I).

Description

[0001] System for optimizing the operation of an automation technology plant and method for optimizing the operation of an automation technology plant

[0002] The invention relates to a system for optimizing the operation of an automation technology system and a method for optimizing the operation of an automation technology system.

[0003] Field devices are already known from the state of the art and are used in industrial plants. They are widely employed in process automation as well as in manufacturing automation. In principle, field devices are defined as all devices used close to the process that provide or process process-relevant data or information. Thus, field devices are used to acquire and / or influence process variables. Measuring devices or sensors are used to acquire process variables. These are used, for example, for measuring pressure and temperature, conductivity, flow rate, pH, level, etc., and acquire the corresponding process variables such as pressure, temperature, conductivity, pH value, level, and flow rate. Actuators are used to influence process variables.These include, for example, pumps or valves that can influence the flow of a liquid in a pipe or the fill level in a container. In addition to the aforementioned measuring devices and actuators, field devices also include remote I / Os, radio adapters, and generally any devices located at the field level.

[0004] A large number of such field devices are produced and distributed by the Endress+Hauser Group.

[0005] In modern industrial plants, field devices are typically connected to higher-level units via communication networks such as fieldbuses (Profibus®, Foundation® Fieldbus, HART®, etc.). These higher-level units are, for example, PLCs (programmable logic controllers) or DCSs (distributed control systems). The higher-level units are used, among other things, for process control and commissioning the field devices. The measured values ​​acquired by the field devices, especially sensors, are transmitted via the respective bus system to one (or possibly several) higher-level units, which may process the measured values ​​further and forward them to the plant's control center. The control center is used for process visualization, process monitoring and control, as well as diagnostics and data storage, etc., via the higher-level units.In addition, data transmission from the higher-level unit to the field devices via the bus system is required, particularly for configuring and parameterizing field devices and controlling actuators. Increasingly, such fieldbuses are being replaced by Ethernet-based communication networks. Field devices generate a wide variety of data. Besides the aforementioned sensor measurement data, which provides plant operators with information about the current process values ​​at their plant's measuring points, this data includes control data, such as for the position control of an actuator. Furthermore, the data includes diagnostic, historical, and / or status data, which informs the plant operator about problems with the field devices or their current status, as well as calibration / parameterization data.

[0006] In the context of Industry 4.0, or ILOT ("Industrial Internet of Things"), data generated by field devices is often collected directly from the field using so-called data conversion units, such as "edge devices" or "cloud gateways," and automatically transmitted to a central cloud-enabled service platform where an application resides. This application, which offers functions for visualizing and further processing the data stored in the database, can be accessed by users via the internet. This essentially corresponds to the concept of the NAMUR Open Architecture.

[0007] Automation systems often execute a process largely autonomously over extended periods. To ensure the automation system performs and controls the process as optimally as possible, it is necessary for the entire system to be operated and monitored by qualified personnel from the system operator, and, if necessary, regularly serviced and / or maintained. State-of-the-art technology includes simplified monitoring and, in some cases, even the operation of an entire system using a computer-aided method.

[0008] The method taught in EP3655828A1 describes the monitoring of an automation system with multiple field devices that, among other things, acquire process variables. The acquired data is transmitted to a higher-level unit with a database and an analysis unit, which is part of a higher-level database and analysis platform, in particular a cloud-computing-based platform. The higher-level unit monitors the system by comparing the acquired secondary environmental variables with reference values ​​to detect any abnormal conditions in the system.

[0009] Document WO2021089461 A1 describes a resource management system for an automation plant with field devices for acquiring and / or influencing process variables. The field devices are connected via an edge device to a cloud-based service platform, which provides, among other things, diagnostic functions for the field devices. Depending on the available computing and storage resources in the field devices, the edge device, and the service platform, the respective devices, and thus the plant, can be optimized. The service platform is suitable for modeling the relationships between measured values ​​and the quality of, for example, a product from the plant. Such modeling enables the optimal adjustment of the plant's operating parameters, allowing the plant to produce an optimized product or perform a task in an optimized manner.

[0010] However, the problem with the platforms mentioned above is that these platforms have components located at least partially outside the automation system, for example as part of a computing cloud, so that process and plant-specific data can be accessed by unauthorized persons via the internet.

[0011] Based on this problem, the invention aims to present a system that allows the modeling of a process and the monitoring of a plant to be ensured by a platform without making it vulnerable to unauthorized access.

[0012] The invention solves the problem by means of a system according to independent claim 1.

[0013] The system according to the invention for modeling the dependence of at least one process output variable on at least one process variable of a process in an automation system comprises: a server platform comprising a computer unit and a control unit; at least one first measuring device configured to acquire measured values ​​of the process variables; at least one second measuring device configured to acquire measured values ​​of the process output variables; a closed communication network configured to connect the at least one first measuring device, the at least one second measuring device, and the server platform; wherein the at least one first measuring device is configured to transmit acquired measured values ​​of the process variables as input values ​​to the server platform via the closed communication network;wherein at least one second measuring device is configured to transmit recorded measured values ​​of the process output variables as output values ​​to the server platform via the closed communication network; wherein the server platform is configured to store the transmitted input values ​​and the transmitted output values ​​at least partially as time series; wherein the server platform is configured to train an algorithm with the stored time series using the computer unit; wherein the server platform is configured to calculate at least one value for the at least one process output variable using the trained algorithm based on the transmitted input values.

[0014] In a further development of the system according to the invention, the closed communication network has access control lists for controlling access by users and / or devices.

[0015] In a further development of the system according to the invention, the closed communication network is separated from other networks, in particular the Internet; wherein the closed communication network optionally has a segmentation into subnetworks, and / or firewalls for monitoring data traffic and / or firewalls for monitoring data traffic.

[0016] In a further development of the system according to the invention, the server platform is configured to subject the transmitted input values ​​and / or the transmitted output values ​​to a check.

[0017] In a further development of the system according to the invention, the verification includes a comparison of the transmitted input values ​​and / or output values ​​with the stored time series.

[0018] In a further development of the system according to the invention, the control unit is configured to transmit data to a receiver outside the closed communication network; wherein the data to be transmitted includes a selection of status information from the at least one first measuring device, status information from the at least one second measuring device, the transmitted input values, the transmitted output values, the stored time series, the at least one calculated value for the process output variable, and optionally a result of the verification; wherein the control unit is configured to control the transmission of the data to be transmitted, in particular to interrupt and resume it, and preferably to physically interrupt and resume it.

[0019] In a further development of the system according to the invention, the receiver is configured to perform analyses concerning the state of the at least one first measuring device based on the state information of the at least one first measuring device; wherein the receiver is configured to perform analyses concerning the state of the at least one second measuring device based on the state information of the at least one second measuring device. In a further development of the system according to the invention, a result of the analyses is transmitted to the closed communication network by means of the control unit.

[0020] In a further development of the system according to the invention, information transmission between the at least one first measuring device and / or the at least one second measuring device and the server platform can be interrupted at any time at the control unit;

[0021] So that no data can be transmitted between the at least one first measuring device and / or the at least one second measuring device, and the server platform.

[0022] In a further development of the system according to the invention, the calculated values ​​for the process output variables are a prediction for the values ​​of the process output variables.

[0023] In a further development of the system according to the invention, the automation system has control parameters; wherein the computer unit calculates control parameters based on the trained machine learning system; wherein the calculated control parameters serve to operate the process in such a way that the process output variable assumes a certain setpoint.

[0024] In a further development of the system according to the invention, the computer unit and / or the control unit and / or the at least one first measuring device and / or the at least one second measuring device are configured to execute programs; wherein the programs are containerized, in particular by means of Docker, Kubernetes, or Open Container Initiative.

[0025] In a further development of the system according to the invention, the algorithm comprises one or more sub-algorithms; wherein a sub-algorithm comprises a machine learning system, a deep learning algorithm, a convolutional neural network, a convolutional neural network with long short term memory, a dilated convolutional neural network, deterministic models, or a hybrid model comprising any combination of the aforementioned models.

[0026] The inventive method for a system according to the invention for modeling the dependence of at least one process output variable on at least one process variable of a process in an automation system or a further development thereof, comprises at least the following steps: Acquiring measured values ​​of the process variables with the at least one first measuring device; Acquiring measured values ​​of the process output variables with the at least one second measuring device; Transmitting the measured values ​​of the process variables to the server platform as input values; Transmitting the measured values ​​of the process output variables to the server platform as output values; Storing the input values ​​and the output values ​​as time series, at least in segments; Training a machine learning system with the time series; Calculating at least one value for the at least one process output variable using the trained machine learning system.

[0027] A further development of the method according to the invention includes: checking the input values ​​and / or the output values, in particular by comparing the input values ​​and / or the output values ​​with the stored time series.

[0028] In a further development of the method according to the invention, the process steps are carried out at least partially in parallel.

[0029] The system according to the invention has the advantage that an unauthorized user cannot access the measured values, input values, output values, the calculated value for a process output variable, and the status information of the measuring devices from outside the closed communication network. In particular, the invention offers the advantage that the security of the data transmitted and / or stored in the closed communication network and / or on the server platform is ensured. Furthermore, the invention has the advantage that a connection to a receiver, especially via the internet, can be interrupted at the control unit.The invention further has the advantage that an algorithm, in particular a machine learning system, can be trained, and that at least one value for the at least one process output variable can be calculated using the trained algorithm, even if the closed communication network is not connected to the internet and / or a receiver. The invention further has the advantage that the input values ​​and / or the output values ​​can be verified, even if the closed communication network is not connected to the internet and / or a receiver. Furthermore, the invention has the advantage that the at least one first measuring device comprises field devices, which are configured, for example, for pressure and temperature measurement, conductivity measurement, flow measurement, pH measurement, level measurement, etc., and which acquires the corresponding process variables pressure, temperature, conductivity, pH value, level, and flow rate.Furthermore, the advantage is that the second measuring device is designed for the analysis of liquids, in particular for the examination of a medium involved in the process, which can analyze samples of the medium involved in the process and can be located in a separate room.

[0030] Furthermore, the system according to the invention has the advantage that it can be built with containerized applications, for example in Docker, Kubernetes, or the Open Container Initiative, and with special software for orchestrating the method according to the invention. This makes it scalable and allows it to be housed in an edge computer within the system. This includes the aspect of using containerized software applications, in particular Python-based analysis functions, which are not limited to open-source platforms for deploying and managing containerized software applications.

[0031] Furthermore, there is the advantage that a specific algorithm — a so-called orchestrator — can be set in the computer unit, which receives the data recorded by the measuring devices via at least one programming interface, transmits data to the algorithm for verification to suitable software applications, monitors and / or controls the software applications, and transmits the results of checks, calculated values, process control parameters, and other process-relevant information to the communication network and, if applicable, to the receiver.

[0032] Furthermore, the advantage is that a check of the transmitted input values ​​and / or the transmitted output values ​​by the computer unit can provide information about an aging process of the measuring instruments, or information about a property of the process or a subprocess, in particular information about the stability of the process.

[0033] Furthermore, the advantage is that connecting the system according to the invention with a possibly cloud-based receiver enables the receiver to perform analyses of at least some of the data collected in the system, whereby the receiver can, for example, use additional information about a device type of the first and / or second measuring device.

[0034] Furthermore, there is the advantage that the integration of the data recorded by the measuring devices, the results of checks, and other process-relevant information can be efficiently made available via defined interfaces in the closed communication network.

[0035] Furthermore, the computer unit has the advantage that it can calculate control parameters for controlling the system, which serve to optimize a process output, so that a process output variable assumes a specific setpoint that can be detected by at least one second measuring device.

[0036] The invention is explained with reference to the following figures. Fig. 1 shows a schematic view of an embodiment of the system according to the invention.

[0037] Fig. 2 schematically shows the sequence of one embodiment of the method according to the invention.

[0038] The schematic view shown in Fig. 1 of an embodiment of the system according to the invention for modeling the dependence of at least one process output variable on at least one process variable of a process in an automation system comprises a first measuring device M1 and a second measuring device M2. In this specific case, the measuring device M1 is a flow meter suitable for mining or open-pit mining, which measures, for example, a flow velocity, a density, or an electromagnetic property of a medium involved in the process. In this specific case, the measuring device M2 is a sensor in a laboratory which measures one or more properties of a medium involved in the process. This measurement can, among other things, determine the composition of wastewater flowing from the system outside the system and at regular intervals, for example, from several hours to days.In this specific case, at least one second measuring device M2 is located at a different location than the system, and the measurement data measured by this second measuring device M2 are fed into the communication network at regular intervals. The system can include further first measuring devices MT and further second measuring devices M2' (shown as dashed lines in Fig. 1).

[0039] The at least one first measuring device M1 is connected to the closed communication network KN, for example via Ethernet / IP or a fieldbus of automation technology (Profibus®, Foundation® Fieldbus, HART®, etc.), which closed communication network KN is connected to the server platform SP. The closed communication network can be wired or wireless, for example using a wireless fieldbus standard.

[0040] In this configuration, the server platform SP is connected to the closed communication network KN via the control unit SE. The server platform SP includes a computer unit CE and a machine learning system. The server platform SP can have multiple computer units CE.

[0041] Access to the closed communication network (KN) by users and devices is controlled by access control lists (ACLs). In this configuration, the closed communication network (KN) is not isolated from other networks, particularly the internet, but is connected to the receiver (E) via another network, such as the internet, by the control unit (SE). Access to the closed communication network (KN) and the server platform (SP) can, for example, occur locally and / or by the control unit. The closed communication network can optionally be segmented into subnetworks and / or have firewalls for monitoring data traffic. Implementing access control lists and / or firewalls in the control unit can be advantageous.

[0042] In this configuration, the time series ZR, the input values ​​I, the output values ​​O, and the status information of measuring devices M1 and M2 can be transmitted from the control unit SE to the receiver E. Receiver E is a cloud-based server platform and can contain an additional machine learning system and further data, and can, for example, analyze the data transmitted by the control unit SE. Among other things, the receiver can perform analyses regarding the status of measuring devices M1 and M2 based on the transmitted status information. It can be advantageous for receiver E to transmit the result of an analysis to the control unit SE so that the result is available in the closed communication network KN and / or on the server platform SP.

[0043] In this configuration, the SP server platform features a CE computer unit designed as an edge computer with the following technical specifications: Debian Linux operating system (optional ctrIX OS); Intel® i7-7600U 2.8 GHz (max. 3.90 GHz) processor; TPM 2.0 chip security module; 16 GB DDR4 2133 MHz main memory; 256 GB SATA 2.5" SSD internal flash storage; full-size mPCIe slot for memory expansion; 1 x DisplayPort 1.2, 1 x HDMI v1.4, Intel® HD Graphics interfaces; 3 LED indicators; DC 24 V (10 ... 36 V) supply voltage; 2292 mA typical input current at nominal load (24 V); 3967 mA maximum input current (24 V); 55 W typical, 95.2 W maximum operating power.

[0044] The schematic flow diagram of an embodiment of the method according to the invention, shown in Fig. 2, depicts the measured values ​​W1 acquired by a first measuring device M1 and the measured values ​​W2 acquired by a second measuring device M2. These measured values ​​W1 and W2 are transmitted by the measuring devices M1 and M2 as input values ​​I and output values ​​O via the communication network KN to the server platform SP, where they are stored as time series ZR. In particular, the time series ZR comprise measured values ​​W1 and W2 that were transmitted within the last year, preferably within the last four months. In this embodiment of the method, the computer unit CE performs a so-called "time stamp reconciliation" after storing the time series ZR, so that the time series can be used by machine learning systems.In this configuration, the computer unit CE is set up to train the machine learning system ML using the stored time series ZR, the input values ​​I, and the output values ​​O, and to calculate at least one value WO for the at least one process output variable using the trained machine learning system and the input values ​​I, and to make these available to the server platform SP.

[0045] In this configuration, the computer unit CE is still set up to check the input values ​​I and the output values ​​O. In this configuration, a check Ü can be an "input check," where the input values ​​I and the output values ​​O are compared with the stored time series ZR. Alternatively or additionally, the check Ü can be a "context check," where, for example, the input values ​​I and the output values ​​O are compared with the stored time series ZR and other information. In this configuration of the procedure, all process steps can take place in parallel and can be orchestrated by a central program.

[0046] Reference symbol list

[0047] SP Server Platform

[0048] CE Computer Unit SE Control Unit

[0049] M1, MT First measuring device

[0050] M2, M2' Second measuring device

[0051] W1 Measured values ​​of the process variables

[0052] W2 Measured values ​​of the process output variables KN Communication network

[0053] I Input values

[0054] O Output values

[0055] ZR time series

[0056] ML Algorithm Review

[0057] WO calculated value of the process output variable

Claims

Patent claims 1. System for modeling the dependence of at least one process output variable on at least one process variable of a process in an automation plant, comprising: • a server platform (SP), comprising a computer unit (CE) and a control unit (SE); • At least one first measuring device (M1) set up to record measured values ​​of the process variables (W1); • At least one second measuring device (M2) set up to record measured values ​​of the process output variables (W2); • a closed communication network (CN) set up to connect at least one first measuring device (M1), at least one second measuring device (M2), and the server platform (SP); • wherein the at least one first measuring device (M1) is set up to transmit recorded measured values ​​of the process variables (W1) as input values ​​(I) to the server platform (SP) via the closed communication network (KN); • wherein at least one second measuring device (M2) is set up to transmit recorded measured values ​​of the process output variables (W2) as output values ​​(O) to the server platform (SP) via the closed communication network (KN); • wherein the server platform (SP) is set up to store the transmitted input values ​​(I) and the transmitted output values ​​(O) at least section by section as time series (ZR); • wherein the server platform (SP) is set up to train an algorithm (ML) with the stored time series (ZR) using the computer unit (CE); • wherein the server platform (SP) is set up to calculate at least one value (WO) for the at least one process output variable using the trained algorithm (ML) based on the transmitted input values ​​(I).

2. System according to claim 1, • wherein the closed communication network (CN) includes access control lists to control access by users and / or devices.

3. System according to one of claims 1 or 2, • wherein the closed communication network (CN) is separate from other networks, in particular the Internet; • wherein the closed communication network (CN) optionally includes segmentation into subnetworks, and / or firewalls for monitoring data traffic and / or firewalls for monitoring data traffic.

4. System according to one of claims 1 to 3, • wherein the server platform (SP) is set up to perform a check (Ü) on the transmitted input values ​​(I) and / or the transmitted output values ​​(O).

5. System according to claim 4, • where the verification (Ü) includes a comparison of the transmitted input values ​​(I) and / or output values ​​(O) with the stored time series (ZR).

6. System according to any one of claims 1 to 5, • wherein the control unit (SE) is configured to transmit data to a receiver (E) outside the closed communication network (CN); • wherein the data to be transmitted includes a selection of status information from at least one first measuring device (M1), status information from at least one second measuring device (M2), the transmitted input values ​​(I), the transmitted output values ​​(O), the stored time series (ZR), the at least one calculated value (WO) for the process output variable, and, if applicable, a result of the verification (Ü); • wherein the control unit (CE) is configured to control the transmission of the data to be transmitted, in particular to interrupt and resume it, and preferably to physically interrupt and resume it.

7. System according to one of claim 6, • wherein the receiver (E) is equipped to perform analyses concerning a state of the at least one first measuring device (M1) using the state information of the at least one first measuring device (M1); • wherein the receiver (E) is equipped to perform analyses concerning the state of at least one second measuring device (M2) using the state information of the at least one second measuring device (M2).

8. System according to claim 7, • wherein a result of the analyses is transmitted to the closed communication network (CN) via the control unit (SE).

9. System according to any one of claims 1 to 8, • Whereby information transmission between the at least one first measuring device (M1) and / or the at least one second measuring device (M2), and the server platform (SP) can be interrupted at any time at the control unit (SE); So that no data can be transmitted between the at least one first measuring device (M1) and / or the at least one second measuring device (M2), and the server platform (SP).

10. System according to any one of claims 1 to 9, • where the calculated values ​​(WO) for the process output variables are a prediction for values ​​of the process output variables.

11. System according to any one of claims 1 to 10, • where the automation system has control parameters; • wherein the computer unit (CE) calculates control parameters based on the trained machine learning system; • where the calculated control parameters serve to operate the process in such a way that the process output variable assumes a specific target value.

12. System according to any one of claims 1 to 11, • wherein the computer unit (CE) and / or the control unit (SE) and / or the at least one first measuring instrument (M1) and / or the at least one second measuring instrument (M2) are configured to execute programs; • where the programs are containerized, in particular using Docker, Kubernetes, Open Container Initiative.

13. System according to any one of claims 1 to 12, • wherein the algorithm (ML) comprises one or more sub-algorithms; • wherein a sub-algorithm comprises a machine learning system, a deep learning algorithm, a convolutional neural network, a convolutional neural network with long short term memory, a dilated convolutional neural network, deterministic models, or a hybrid model comprising any combination of the foregoing models.

14. Method for a system according to any one of claims 1 to 13, comprising at least the following steps: • Acquisition of measured values ​​of the process variables (W1) with the at least one first measuring device (M1); • Acquisition of measured values ​​of the process output variables (W2) with at least one second measuring device (M2); • Transmitting the measured values ​​of the process variables (W1) to the server platform (SP) as input values ​​(I); • Transmitting the measured values ​​of the process output variables (W2) to the server platform (SP) as output values ​​(O); • At least partially storing the input values ​​(I) and the output values ​​(O) as time series (ZR); • Training a machine learning system (ML) with the time series (ZR); • Calculating at least one value (WO) for the at least one process output variable using the trained machine learning system (ML).

15. The method of claim 14, further comprising: • Checking the input values ​​(I) and / or the output values ​​(O), in particular by Comparing the input values ​​(I) and / or the output values ​​(O) with the stored time series (ZR).

16. Method according to one of claims 14 or 15, • wherein the method steps are carried out at least partially in parallel.

Images