Edge computing devices for processing plant process data

Abstract

The present invention relates to an edge computing device (120) for processing data (112) acquired in relation to a production process of an industrial plant including a plant control system (110), the control system including a server (111). The device includes a first unit (125) configured to be communicatively coupled to the server for receiving data from the system. The second unit (121) is configured to provide a container runtime environment (122) configured to execute a container in the second unit. A process container (123) is configured to execute in the environment, the process container including a program configured to process the acquired data in relation to the production process of the industrial plant when executed inside the process container. The first unit is communicatively coupled to the second unit for providing the received data to the environment, the program configured to process the provided data.

Description

[Technical field]

[0001] FIELD OF THEINVENTION The present invention relates to an edge computing device, a computing framework, a method and a computer program product for processing plant process data acquired in relation to a production process of an industrial plant. [Background technology]

[0002] 2. Background of the Invention Today, the optimization of complex production processes in industrial plants, for example in the chemical or pharmaceutical industry, is often used to improve production reliability, quality and quantity. With the increasing complexity of the production processes, the analysis and optimization programs as well as the IT infrastructure running these programs often need to be updated, for example by adding additional hardware and / or software. Typically, such analysis and optimization programs are based on processing plant process data received from a plant control system controlling the industrial plant and are often also adapted to provide, as output, respective optimized control data back to the plant control system. In order to improve the data security input and output of the analysis and optimization programs, the respective analysis and optimization programs are often deployed on-site, i.e. are part of the hardware and / or software structure of the plant. Therefore, it is often difficult and inefficient to adapt, update or replace the analysis and optimization programs in multiple industrial plants. In many cases, even manual manipulation of the software and / or hardware of the production plants is required on-site. This makes the development and deployment process of the analysis and optimization programs of the production plants inefficient and time-consuming. It would therefore be advantageous if the analysis and optimization programs of multiple industrial plants could be more efficiently managed, i.e. adapted, updated and / or replaced. Summary of the Invention [Problem to be solved by the invention]

[0003] Summary of the Invention It is an object of the present invention to provide an edge computing device, a computing framework, a method and a computer program product that allows for more efficient management, i.e. updating, adapting and / or replacing, of analysis and optimization programs of industrial plants. [Means for solving the problem]

[0004] In a first aspect of the present invention, an edge computing device for processing plant process data acquired for a production process of an industrial plant including a plant control system is presented, the plant control system including an interface server. The edge computing device includes: a) an interface computing unit configured to be communicatively coupled to the interface server for receiving plant process data from the plant control system, b) a container computing unit configured to provide a container runtime environment configured to execute a container on the container computing unit, and c) a process container configured to execute on the container runtime environment, the process container including a processing program configured to process the plant process data acquired for the production process of the industrial plant when executed inside the process container, the interface computing unit is communicatively coupled to the container computing unit for providing the received plant process data to the container runtime environment, the processing program configured to process the provided plant process data.

[0005] The edge computing device includes a container computing unit and a process container configured to run in a container runtime environment, the process container including a processing program configured to process plant process data and, for example, to analyze and optimize a plant process based on the plant process data, such that the processing program can be executed independently of the respective hardware and software provided as part of the industrial plant. Furthermore, the independence of the runtime environment of the processing program from the hardware and / or software provided by the industrial plant allows management of the processing program independent of the IT infrastructure of the industrial plant. Thus, the efficiency of management, i.e., adaptation, updating and / or replacement, of the processing program for optimizing the process of the industrial plant can be improved.

[0006] In general, an edge computing device refers to a computing device adapted to follow the edge computing paradigm, which refers to distributed computing, which brings the computation closer to where the results of the computation are needed, as opposed to the cloud computing paradigm. The use of edge computing generally allows for improved security of data transmission, improved response times, and allows for secure network bandwidth. Furthermore, edge computing in this application allows for a physical island mode of operation of the combination of plant automation systems and connected IT infrastructure, e.g. edge devices. Therefore, it is preferred that the edge computing device is provided on-site, i.e. as part of the industrial plant, and is directly connected to the computing hardware and / or software of the industrial plant.

[0007] The acquired plant process data relating to the production process of the industrial plant may refer to any data acquired by the plant control system during the production process. For example, the plant process data may refer to sensor measurements of a number of sensors provided in the plant to monitor the production process. The process data may also refer to settings utilized by one or more production equipment of the industrial plant during the production process, such as settings of valves, industrial heaters, pumps, etc. The plant control system may refer to any system used to control the production process in the plant and is realized as hardware and / or software in the industrial plant. The plant control system includes an interface server adapted to interface with other computing units to provide, for example, the acquired plant process data to the other computing units.

[0008] The interface computing unit is configured to be communicatively coupled to the interface server to receive plant process data from the plant control system. For example, the interface computing unit of the edge computing device can be communicatively coupled to the interface server of the plant control system via a wired or wireless connection. Preferably, to improve security and availability of data transmission, the interface computing unit is communicatively coupled to the interface server via a wired connection. However, a wireless connection, for example based on the 5G standard, may also be preferred if a wireless connection is available in a secure and reliable manner.

[0009] The container computing unit is configured to provide a container runtime environment, in which the container is configured to execute in the container runtime environment. The container runtime environment refers to a computing environment that enables the execution of containers. A container in this context can generally be defined as an instance of an operating system that is isolated from other instances of the operating system and can independently manage the computing resources of the edge computing device available to each container. Thus, a container bundles all resources down to the kernel level and the runtime dependencies of a computer program running within the container. Furthermore, due to limitations on the computing resources of the computing system that a container can access, each container preferably includes each piece of software required to execute each program containerized in the container. For example, a container preferably includes all libraries of a computing framework required by each processing program.

[0010] The edge computing device further includes a process container configured to execute in the container runtime environment. The process container includes a processing program containerized in the process container. The processing program, when executed inside the container, is generally configured to process plant process data acquired for a production process of the industrial plant. Processing preferably refers to the analysis of the plant process data with respect to the optimization of the production process of the industrial plant. Furthermore, processing may refer to generating optimized plant control data that can be used to control the industrial plant to perform the optimized production process. The processing program may refer to any known already existing processing program or personalized processing program that is specifically adapted to process the plant process data of a particular production plant. For example, the processing program may refer to a soft sensor. The soft sensor may be realized, for example, as a Kalman filter that reconstructs the internal state of the industrial plant that may be difficult to measure at a physical level. Such reconstructed internal state may then be fed back to a plant control system as plant control data, where this form of plant control data may then be treated as an actual physical measurement. In another example, the processing program may refer to a model predictive controller. Such a controller may be used, for example, to optimize throughput through a distillation column based on a state space model of the underlying process.

[0011] The interface computing unit is further communicatively coupled to the container computing unit so as to be able to provide the received plant process data received from the interface server to the container runtime environment. The communicative coupling can be realised in the form of any wired or wireless connection between the interface computing unit and the container computing unit. Preferably, the interface computing unit and the container computing unit are part of the same computing hardware and / or software. In this case, the communicative coupling may refer to a virtual coupling, which allows the container computing unit to access the data provided by the interface computing unit and vice versa. In general, the container runtime environment is adapted to provide access to the received plant process data of the processing program containerised in the process container as a resource to the container. The processing program is therefore adapted to access and in particular process the provided plant process data.

[0012] In one embodiment, the interface computing unit is communicatively coupled to the interface server via a protocol converter configured to separate and tunnel the plant process data and convert the plant process data from a data format utilized by the interface server to a data format utilized by the interface computing unit. In many applications, the plant control system, and therefore the interface server, provides data, particularly plant process data, in a particular format, while the edge computing device, and particularly the interface computing unit, is configured to use a different data format, e.g., a more recent data format. In this case, a protocol converter provided between the interface computing unit and the interface server converts data from the interface server to a data format utilized by the interface computing unit, thus enabling a more efficient exchange of data. The protocol converter may be provided as part of the interface server, part of the interface computing unit, or both the interface server and the interface computing unit to perform this task. For example, the protocol converter may be provided as hardware and / or software in the interface server, the interface computing unit, or both.

[0013] Additionally, the protocol converter is configured to segregate and tunnel the plant process data. Segregation of data generally refers to the processing of data such that only the intended data is transferred. Tunneling of data generally refers to the repackaging of transferred data into different data protocols during transfer and the transportation of data from a physical source to each recipient. In a preferred embodiment, the protocol converter is adapted to tunnel the plant process data by encrypting the plant process data using encryption.

[0014] The protocol converter converts the plant process data from the data format used by the interface server to the data format used by the interface computing unit, and in the process is adapted to isolate and tunnel the plant process data, thereby not only making communication between the interface server and the interface computing unit more efficient, but also greatly improving the security of this communication.

[0015] In a preferred embodiment, the protocol converter is configured to convert the plant process data provided in the legacy data format in the interface server into a modern data format. For example, due to stable working performance and also due to the large amount of expenses often required to implement new hardware or software in an already running plant control system, multiple legacy software and hardware systems are still active in many plant control systems. Thus, such legacy software or hardware systems also provide respective legacy data, i.e. data provided in a legacy data format, where a legacy data format refers to a data format that is obsolete or outdated, generally no longer used by modern applications. In this regard, the term "modern data format" refers to a data format that is still usable by current modern software and hardware applications. Since many legacy data formats are not only not usable by modern hardware and software systems, but may further include additional drawbacks such as security issues, reduced storage performance, etc., it is advantageous to convert the plant process data provided in the legacy data format into a modern data format in order to transfer the plant process data to the interface computing unit. For example, the legacy data format may refer to an Open Platform Communications Data Access (OPC DA) format and the modern data format refers to an Open Platform Communications Unified Access (OPC UA) format, in which case the protocol converter preferably includes an Open Platform Communications Data Access / Unified Access (OPC DA / UA) converter configured to convert the plant process data provided in the interface server in the OPC DA format by applying an OPC UA wrapper to the plant process data.However, if the plant process data is already provided in a modern data format, the protocol converter is preferably adapted to act as a transparent communication tunnel, i.e., the protocol converter does not convert the plant process data, but only isolates and tunnels the plant process data.

[0016] In one embodiment, the processing program is configured to provide plant control data for controlling or optimizing the industrial plant based on the processing of the plant process data, the container computing unit being communicatively coupled to the interface computing unit for providing the plant control data to the interface computing unit, and the interface computing unit being communicatively coupled to the interface server for providing the plant control data to the interface server for implementation in a plant control system of the industrial plant. In this embodiment, the processing program is adapted to determine, based on the plant process data, plant control data that can be used for controlling or optimizing the industrial plant. In particular, the plant control data can be the result of an analysis of the plant process data. For example, such an analysis can indicate that the production of a particular product of the production plant can be improved if certain hardware of the production process, e.g. a heater, is controlled differently, e.g. controlled to provide less heat to the product at a particular production stage. Thus, in this example, the provided plant control data can include data indicating to the plant control system of the industrial plant to adapt the control of the respective hardware, in this example the heater, accordingly. The plant control data provided by the processing program is then provided to the interface computing unit via the communicatively coupled container computing unit, and further provided from the interface computing unit to the interface server, such that the plant control data can be implemented in the plant control system of the industrial plant. In this embodiment, not only is it possible to analyze and process data provided by the industrial plant, but it is also possible to directly influence, and in particular optimize, the control of the industrial plant.

[0017] Also in this embodiment, the interface computing unit is preferably communicatively coupled to the interface server via a protocol converter, which is configured to separate and tunnel the plant control data and convert the plant control data from a data format used by the interface computing unit to a data format used by the interface server. As already mentioned above and also in this case, when the plant control system still includes software and / or hardware that uses a legacy data format, it is advantageous for the protocol converter to be configured to convert the plant control data provided in a modern data format to a legacy data format and provide the plant control data to the plant control system. For example, the protocol converter can include an OPC DA / UA converter, which is configured to convert the plant control data in OPC UA format to an OPC DA format and provide the plant control data to the plant control system.

[0018] In one embodiment, the edge computing device further includes an abstraction container configured to run in the container runtime environment, the abstraction container including an abstraction program configured to pre-process the plant process data before the plant process data is provided to the process container. Providing an abstraction container including an abstraction program to pre-process the plant process data can increase the flexibility of the system regarding how the plant process data is provided by the interface computing unit to the container computing unit in view of how the plant process data is utilized in the processing program. The pre-processing performed by the abstraction program may refer to any kind of processing of the plant process data that allows the plant process data received from the interface computing unit to be provided in a format expected by the processing program as input. For example, pre-processing may refer to conversion of the plant process data, in which case the abstraction program includes respective data or protocol converters. However, pre-processing may also refer to sorting or converting one physical unit to another physical unit, such as converting all physical units to SI units. Furthermore, the abstraction program may also be adapted to filter the plant process data, for example, to provide only a portion of the plant process data to the process container and thus the processing program. Furthermore, the abstraction program may also be adapted to perform operations on the time series data, for example, to calculate an aggregate average of values ​​over a time window. Such aggregated time series data is particularly suitable for use in machine learning models.

[0019] Because an abstraction container containing an abstraction program is provided, the system becomes more flexible and has the advantage that not only the abstraction program is provided in the form of an abstraction container, but furthermore it can be managed, i.e. updated, adapted or replaced, in the same easy and efficient way as the processing container, thus allowing easy adaptation of any changes that occur in the processing program. Furthermore, when an abstraction program is utilized for pre-processing of the plant process data, the processing container, and in particular the respective processing programs running in the processing container, can also be more easily shared across multiple industrial plants, so that every industrial plant does not require a customized processing program. This generally further improves the effectiveness of managing the processing of the plant process data.

[0020] In one embodiment, the interface computing unit and / or the container computing unit are configured to run a virtual Windows environment and / or a virtual Linux environment. In particular, the container computing unit is preferably configured to run a Linux environment as an operating system or to run a virtual Linux environment capable of providing a container runtime environment. Furthermore, the interface computing unit is preferably configured to run a Windows environment as an operating system or is adapted to run a virtual Windows environment when running with a different operating system. In general, the interface computing unit and the container computing unit can be adapted to run with any suitable operating system and then, if necessary, to provide a respective virtual environment in which another operating system runs in order to run specific software as is advantageous for the respective tasks of the interface computing unit and / or the container computing unit.

[0021] Configuring the interface computing unit and / or container computing unit to run respective virtual operating system environments provides further flexibility to the edge computing device with respect to, for example, the operating systems running on each computing unit of the edge computing device.

[0022] In one embodiment, the container runtime environment is configured to enable deployment of the processing programs running in the process containers via a continuous integration and continuous delivery (CI / CD) workflow. The container runtime environment is configured to enable deployment of the processing programs via a CI / CD workflow, which allows the processing programs to be managed very efficiently even when multiple processing programs of different industrial plants need to be managed. Furthermore, if the container runtime environment further includes an abstraction container with an abstraction program, it is further preferred that the abstraction program can also be deployed via a CI / CD workflow.

[0023] In a further aspect of the present invention, a computing framework for processing plant process data acquired relating to a production process of an industrial plant including a plant control system is presented, the computing framework comprising: a) an edge computing device as described above; and b) an interface server that is part of the plant control system, communicatively coupled to the interface computing unit and for providing plant process data from the plant control system to the interface computing unit of the edge computing device and vice versa.

[0024] In a further aspect of the invention, a method is presented for processing plant process data acquired for a production process of an industrial plant including a plant control system, the plant control system including an interface server, the method including: a) receiving plant process data from the plant control system via the interface server, b) providing a container runtime environment configured to execute a container, the container including a processing program configured to process the plant process data of the industrial plant when executed within the container, and c) providing the received plant process data to the container runtime environment and processing the provided plant process data with the processing program.

[0025] In a further aspect of the present invention there is provided a computer program product for processing plant process data acquired relating to a production process of an industrial plant including a plant control system, the computer program product comprising program code means for causing an edge computing device as described above to perform the method as described above.

[0026] It is to be understood that the above mentioned edge computing device, the above mentioned computing framework, the above mentioned method and the above mentioned computer program product have similar and / or identical preferred embodiments, in particular as defined in the dependent claims.

[0027] It is to be understood that a preferred embodiment of the invention can be any combination of the dependent claims or the above-mentioned embodiments with the respective independent claims. These and other aspects of the invention will be apparent and elucidated with reference to the embodiments described hereinafter. [Brief description of the drawings]

[0028] [Figure 1] 1 illustrates, in a schematic and exemplary manner, one embodiment of a computer framework for processing plant process data. [Diagram 2]1 illustrates, diagrammatically and exemplarily, a process diagram of a continuous deployment and delivery workflow. [Diagram 3] 1 illustrates, generally and exemplarily, a flow chart of a method for processing plant process data. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0029] Detailed Description of the Preferred Embodiments FIG. 1 illustrates a schematic and exemplary computing framework 100 for processing plant process data 112 acquired regarding a production process of an industrial plant that includes a plant control system 110 .

[0030] Generally, an industrial plant may refer to any technological platform used for industrial purposes. The industrial purpose may be the manufacture or processing of one or more industrial products, i.e. the manufacturing process or processing carried out by the industrial plant. For example, the industrial purpose may refer to the production of industrial products. The industrial products may be any physical products, such as, for example, chemical, biological, pharmaceutical, food, beverage, textile, metal, plastic or semiconductor. Additionally or alternatively, the industrial products may even be service products, such as electricity, heating, air conditioning, waste treatment, such as recycling, chemical processing, such as disintegration or melting, or incineration. Accordingly, the industrial plant may be one or more of a chemical plant, a process plant, a pharmaceutical plant, a fossil fuel processing facility, such as an oil well and / or a natural gas well, a refinery, a petrochemical plant, a fractionation plant, etc. The industrial plant may even be either a distillery, an incinerator or a power plant. Furthermore, the industrial plant may even be a combination of any of the above examples.

[0031] To carry out the production process, the industrial plant includes a technology platform that can be controlled by control parameters, which can be considered as part of the plant control data implemented by the process control system on the technology platform to control the industrial plant. The technology platform can include equipment or process units such as any one or more of the following: heat exchangers, towers such as fractionators, furnaces, reaction chambers, catalytic cracking units, storage tanks, settlers, pipelines, stacks, filters, valves, actuators, transformers, circuit breakers, machinery, such as turbines, generators, grinders, compressors, fans, pumps, rotating heavy equipment such as motors, etc. Each of this technology platform can provide monitoring data that monitors the functioning of the technology platform. Furthermore, the industrial plant typically includes a number of sensors that allow the operating parameters of the technology platform to be measured. The operating parameters measured by the sensors can relate to various process parameters and / or parameters related to the equipment or process units. For example, sensors may be used to measure process parameters such as flow rate in a pipeline, liquid level inside a tank, temperature of a furnace, chemical composition of a gas, some sensors may be used to measure turbine vibration, fan speed, valve operation, corrosion of a pipeline, voltage across a transformer, etc. Both functional parameters and operational parameters may be considered as part of the plant process data 112 obtained regarding the production process of an industrial plant.

[0032] The plant process data 112 may generally be part of the plant control system 110, and may be acquired and / or stored by the plant control system 110, for example. Additionally, the plant control system 110 may also be adapted to further process the plant process data 112, for example to monitor, analyze, and / or optimize the plant process data 112. However, in many applications, the plant control system 110 is not configured to provide modern deep analysis and possible optimization of the plant process data 112, and often even references legacy systems that do not provide the technological hardware and / or software foundation for modern data analysis of the plant process data 112. It is therefore desirable to provide software and / or hardware that allows modern processing of the plant process data 112, while being easy to deploy and manage.

[0033] To solve this above problem, a computing framework 100 is provided together with an edge computing device 120. The edge computing device 120 is communicatively coupled to an interface server 111 of a plant control system 110, which may also be considered as part of the computing framework 100. In particular, the edge computing device 120 includes an interface computing unit 125 that provides a technological infrastructure, e.g., hardware and / or software, for communicating with the interface server 111 of the plant control system 110. Preferably, the interface computing unit 125 is adapted to receive the plant process data 112 from the interface server 111 through the communicative coupling. In a preferred embodiment, as shown in FIG. 1, the interface computing unit 125 is communicatively coupled to the interface server 111 via a protocol converter 130. Preferably, the protocol converter 130 is configured to separate and tunnel the plant process data 112 from the interface server 111 to the interface computing unit 125. This allows the plant process data 112 to be securely and efficiently transferred from the plant control system 110 to the edge computing device 120. Furthermore, the protocol converter 130 may be adapted to convert the plant process data 112 from a data format used by the plant control system 110 in general to a data format used by the edge computing device 120, in particular the interface computing unit 125 of the edge computing device 120. When the plant control system 110 refers to a legacy system utilizing a legacy data format, the protocol converter 130 is preferably adapted to convert the plant process data 112 provided to the interface server 111 in a legacy data format to a modern data format that can be utilized by the interface computing unit 125.An example of a legacy data format that can be converted to a modern data format is the OPC DA data format still utilized by many plant control systems. In this case, the protocol converter 130 preferably includes an OPC DA / UA converter that can convert the OPC DA data to the OPC UA data format. However, if the plant control system 110 provides the plant process data 112 in the same data format utilized by the interface computing unit 125, the protocol converter 130 may refer to a transparent communication unit that does not provide any conversion to the plant process data 112, but only provides isolation of the plant process data 112 and tunneling between the interface server 111 and the interface computing unit 125. In general, the protocol converter 130 can be provided as part of the interface server 111, part of the interface computing unit 125, or part of both the interface server 111 and the interface computing unit 125 as illustrated in FIG. 1 with schematic portions of the protocol converters 131, 132 provided at each end of the communicative coupling.

[0034] The edge computing device 120 further includes a container computing unit 121 configured to provide a container runtime environment 122 on which the process container 123 is provided. The container computing unit 121 may be, for example, part of the operating system of the edge computing device 120, or may be provided with a virtual operating system that is different from the general-purpose operating system of the edge computing device 120 depending on the operating system running on the edge computing device 120. For example, if the edge computing device 120 is generally provided with a Windows operating system, the container runtime environment 122 is mostly provided as a Linux-based runtime environment, so that the container computing unit 121 may be configured to run a virtual Linux environment operating system.

[0035] Thus, the container runtime environment 122 is adapted to provide a process container 123 including a processing program executing inside the process container 123. The processing program is adapted to process plant process data 112 acquired regarding a production process of the industrial plant. For example, the processing program may be adapted to analyze the plant process data 112 and, based on the analysis, provide optimized process control data enabling optimized control of the industrial plant.

[0036] Optionally, the edge computing device 120 further includes an abstraction container 124 executing in the container runtime environment 122, the abstraction container 124 including an abstraction program. The abstraction program may be provided to pre-process the plant process data 112 before the plant process data 112 is provided to the process container 123 including the processing program. For example, if the plant process data 112 is provided by the interface computing unit 125 in a format that cannot be processed by the processing program, the abstraction program may be adapted to pre-process the plant process data 112 so that it can be processed by the processing program. Such pre-processing may include filtering the plant process data 112, converting units of parameters that are part of the plant process data 112, reorganizing the plant process data 112, selecting the plant process data 112, etc., depending on the respective configuration and interaction of the interface server 111 with the interface computing unit 125 regarding the provision and reception of the plant process data 112.

[0037] In a preferred embodiment, the processing program executed in the process container 123 is adapted to provide plant control data to be implemented in the plant control system 110 based on an analysis of the plant process data 112. The process container 123 executing in the container runtime environment 122 of the container computing unit 121 is then adapted to provide the plant control data to the interface computing unit 125. The interface computing unit 125 can then provide the plant control data to the interface server 111 of the plant control system 110, preferably via a protocol converter 130. In this case, the protocol converter 130 can be adapted to convert the plant control data from a data format utilized by the interface computing unit 125 to a data format utilized by the interface server 111, in particular the plant control system 110. The plant control system 110 can then implement the received plant control data to control the industrial plant.

[0038] In a preferred embodiment, the containerized programs running in the container runtime environment 122 are deployed by a continuous integration and continuous delivery (CI / CD) workflow. The workflow is illustrated by workflow systems 140 to 142. In particular, the CI / CD workflow is based on providing a control system and a container registry 140, for example on an external cloud-based server. The control system may be adapted to deploy and manage all versions of the processing programs and / or abstraction programs, for example by checking whether a more updated version of the processing program is available and then providing the update to the container runtime environment 122, for example replacing the currently running processing program with the updated version of the processing program. Furthermore, a container registry may be provided that respectively registers all containerized programs running on edge computing devices of multiple industrial plants to provide a global view of all currently utilized containerized programs. Optionally, the CI / CD workflow may further include software and / or hardware 142 adapted to continuously receive and monitor outputs and inputs to the processing programs, for example plant process data 112 as inputs and plant control data as outputs. This external software / hardware 142 may then be adapted to analyze and monitor the execution of the process program, for example using an artificial intelligence system. Based on this monitoring and analysis, the software / hardware 141 may be provided to continuously improve the process program, for example providing an updated version of the process program to the control system 140.

[0039] FIG. 2 shows such a CI / CD workflow in a schematic and exemplary manner. Such a workflow may include, for example, a process of first building, testing and then merging parts of processing programs and / or abstraction programs to be utilized in one or more industrial plants. This building, testing and merging may be performed, for example, on an external server or in a cloud-based environment, allowing a safe and efficient development of each processing program. If a processing program or, for example, an updated version of a processing program is available, the process may include automatically releasing the new version of the processing program to a repository in which the new version of the processing program is registered and stored. From the repository, the new version of the processing program may then be automatically deployed to an edge computing device, for example together with the next update, and then implemented in a container runtime environment of a container computing unit of the edge computing device. This allows a very efficient and improved workflow for the deployment of processing programs in an industrial plant.

[0040] 3 refers to a schematic and exemplary illustrated flow chart of a method for processing plant process data. The method 300 can be configured to be utilized with the computing framework 100 described with respect to FIG. 1. In particular, the method 300 comprises in a first step 310 receiving plant process data from a plant control system via an interface server. In step 320, a container runtime environment configured to execute a container is provided, the container including a processing program configured to process the plant process data of the industrial plant when executed inside the container. In a final step 330, the method 300 comprises providing the received plant process data to the container runtime environment and processing the provided plant process data by the processing program.

[0041] The optimization of production processes in the chemical industry can be considered as important for the improvement of the reliability and quality of the production process. Therefore, process optimization utilizing the edge paradigm has been used for many years. Usually, the implementation of a processing program at the edge is done by installing an additional IT infrastructure in the industrial plant. The software stack of the computing framework of such a processing program can consist of components that, for example, i) read data from a plant control system interface, i.e. an interface server, often based on OPC DA, for example with reference to a DCS interface, ii) process this data in a programming language suitable for optimization and advanced data analysis, for example Matlab, and iii) write output data to the DCS interface. Maintaining and deploying such edge intelligence realized by the processing program often requires manual steps, which means that a person goes to the plant and manually copies files to a server in the plant until the models at the edge can be adapted, retrained or replaced. This can slow down the development and deployment process and lead to inefficient workarounds.

[0042] Part of the solution to this problem, for example discussed with respect to FIG. 1, provides for containerization of the processing programs. For example, the processing programs can be deployed as Linux containers containing Python, R or Scala code. Containerization enables a true "build it, run it anywhere" paradigm, thus allowing the processing programs to be built, tested and deployed automatically from the cloud to the edge. Therefore, as a further part of the solution, as already described above with respect to FIG. 1, it is advantageous to allow container-based workloads to be deployed physically close to the interface servers, for example DCS interfaces, and to be able to seamlessly and securely read and write process data from and to the interface servers. This can be achieved, for example, with the computing framework described with respect to FIG. 1, with one of the main differences from existing setups being that the processing programs can be deployed as containers and can be swapped and retrained very fast and agilely, based on full automation, based on CI / CD pipelines.

[0043] For example, the edge computing device may be a spare device, e.g., an IPC, located physically close to the interface server and connected, i.e. communicatively coupled, using, e.g., highly available, possibly redundant, network connections. In the interface server, a protocol converter, e.g., an OPC DA / UA converter, may be installed that locally isolates and tunnels older, potentially insecure, e.g., OPC DA-based communications, optionally with an OPC UA-based wrapper. However, other protocols may be contemplated. For example, the protocol in the interface server may be a legacy protocol, e.g., a COM-based OPC DA. The protocol in the interface computing unit may be a more modern protocol, e.g., MQTT or OPC UA, that allows communication with containerized computer programs. In the edge computing device, an interface server, e.g., an OPC UA server, is available that wraps all access to the interface server. In general, the protocol converter may run on the interface server, may be distributed to the interface server and the interface computing unit, e.g., running an optional virtual Windows environment, or may run only on the interface computing unit.

[0044] Furthermore, a virtual Windows environment and / or a Linux environment can be installed on the interface computing unit and the container computing unit, respectively. For example, the host operating system can be Windows with a Linux virtualization layer or Linux with a Windows virtualization layer. The virtual Windows environment allows the installation of legacy software and protocols, for example, OPC DA is only available on Windows OS. In the Linux environment, a container runtime environment of abstractions and process containers can be deployed.

[0045] The abstraction container can be adapted to further abstract the access to the interfacing computing unit. For example, the optional abstraction container can convert and / or aggregate the plant process data in a format that can be processed by the process container, for example by converting plant specific units to standardized processing program input units, for example from cm to m. The process container can include processing programs for advanced analysis, such as optimizers or trained AI models. For example, the processing programs can be adapted to act as soft sensors providing additional measurements or to emulate missing instruments, for example based on a Kalman filter. The processing programs can also refer to AI models that are pre-trained using, for example, a highly scalable cloud infrastructure, such as big data analysis tools, such as Databricks. Furthermore, any kind of real-time optimizer can be utilized as a processing program. Furthermore, the processing programs can also refer to model predictive controllers, and soft real-time and high availability requirements for the networking stack can be met by the setup.

[0046] Thus, the deployment process can be controlled from cloud services, such as IoT Hub and IoT Edge, using modern CI / CD and DevOps processes, such that an update of a processing program logical cloud side can lead to a fully automatic deployment of this processing program on all associated edge computing devices.

[0047] Generally, the invention as described above has the advantage that processing programs are deployable in an island mode, providing flexible functionality for reading, processing and writing process data to and from an interface server.

[0048] Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

[0049] In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality.

[0050] A single unit or device may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

[0051] The procedures such as receiving plant process data, providing a container runtime environment, executing a processing program, processing plant process data, etc., performed by one or several units or devices may be performed by any other number of units or devices. These procedures may be implemented as program code means of a computer program and / or as dedicated hardware.

[0052] The computer program product may be stored / distributed on a suitable medium, such as an optical storage medium or a solid-state medium, may be supplied together with or as part of other hardware, but may also be distributed in other forms, for example via the Internet or other wired / wireless communication systems.

[0053] Any reference signs in the claims should not be construed as limiting the scope of the invention.

[0054] The present invention relates to an edge computing device for processing data acquired with respect to a production process of an industrial plant including a plant control system, the control system including a server. The device includes a first unit configured to be communicatively coupled to the server for receiving data from the system. The second unit is configured to provide a container runtime environment configured to execute a container in the second unit. A process container is configured to execute in the environment, the process container including a program configured to process the data acquired with respect to the production process of the industrial plant when executed inside the process container. The first unit is communicatively coupled to the second unit for providing the received data to the environment, the program configured to process the provided data.

Claims

1. An edge computing device (120) for processing plant process data (112) acquired regarding a production process of an industrial plant including a plant control system (110), the plant control system (110) including an interface server (111), the edge computing device (120) comprising: an interface computing unit (125) configured to be communicatively coupled to said interface server (111) for receiving said plant process data (112) from said plant control system (110); a container computing unit (121) configured to provide a container runtime environment (122) configured to execute a container on said container computing unit (121); a process container (123) configured to be executed in said container runtime environment (122), said process container (123) including a processing program configured, when executed inside said process container (123), to process plant process data (112) acquired for a production process of said industrial plant; the interface computing unit (125) is communicatively coupled to the container computing unit (121) to provide the received plant process data (112) to the container runtime environment (122), and the processing program is configured to process the provided plant process data (112).

2. 2. The edge computing device (120) of claim 1, wherein the interface computing unit (125) is communicatively coupled to the interface server (111) via a protocol converter (130), the protocol converter (130) configured to separate and tunnel the plant process data (112) and convert the plant process data (112) from a data format utilized by the interface server (111) to a data format utilized by the interface computing unit (125).

3. 2. The edge computing device (120) of claim 1, wherein the protocol converter (130) is configured to convert plant process data (112) provided in the interface server (111) in a legacy data format into a modern data format.

4. 3. The edge computing device of claim 2, wherein the protocol converter comprises an Open Platform Communications Data Access / Unified Access (OPC DA / UA) converter configured to convert the plant process data provided in the interface server in an Open Platform Communications Data Access (OPC DA) data format by applying an Open Platform Communications Unified Access (OPC UA) wrapper to the plant process data.

5. The edge computing device (120) of any one of claims 1 to 4, wherein the processing program is configured to provide plant control data for controlling or optimizing the industrial plant based on the processing of the plant process data (112), the container computing unit (121) is communicatively coupled to the interface computing unit (125) for providing the plant control data to the interface computing unit (125), and the interface computing unit (125) is communicatively coupled to the interface server (111) for providing the plant control data (112) to the interface server (111) for implementation in the plant control system (110) of the industrial plant.

6. 6. The edge computing device (120) of claim 5, wherein the interface computing unit (125) is communicatively coupled to the interface server (111) via a protocol converter (130), the protocol converter (130) configured to separate and tunnel the plant control data and convert the plant control data from a data format utilized by the interface computing unit (125) to a data format utilized by the interface server (111).

7. 7. The edge computing device (120) of claim 6, wherein the protocol converter (130) is configured to convert the plant control data provided in a modern data format to a legacy data format and provide the plant control data to the plant control system (110).

8. 8. The edge computing device of claim 7, wherein the protocol converter includes an OPC DA / UA converter configured to convert plant control data in an OPC UA format to an OPC DA format and provide the plant control data to the plant control system.

9. 9. The edge computing device (120) of claim 1, further comprising an abstraction container (124) configured to run in the container runtime environment (122), the abstraction container (124) including an abstraction program configured to pre-process the plant process data (112) before the plant process data (112) is provided to the process container (123).

10. The edge computing device (120) of any one of claims 1 to 9, wherein the interface computing unit (125) and / or the container computing unit (121) are configured to run a virtual Windows (registered trademark) environment and / or a virtual Linux (registered trademark) environment.

11. The edge computing device (120) of any one of claims 1 to 10, wherein the container runtime environment (122) is configured to enable deployment of the processing program running in the process container (123) via a continuous integration and continuous delivery (CI / CD) workflow.

12. A computing framework (100) for processing plant process data (112) acquired in relation to a production process of an industrial plant including a plant control system (110), comprising: An edge computing device (120) according to any one of claims 1 to 11, an interface server (111) that is part of the plant control system (110) and is communicatively coupled to the interface computing unit (125) to provide plant process data (112) from the plant control system (110) to the interface computing unit (125) of the edge computing device (120) and vice versa; A computing framework (100) comprising:

13. 1. A method (300) for processing plant process data (112) acquired for a production process of an industrial plant including a plant control system (110), the plant control system (110) including an interface server (111), the method (300) comprising: receiving (310) plant process data (112) from said plant control system (110) via said interface server (111); providing (320) a container runtime environment (122) configured to execute a container, the container including a processing program configured to process the plant process data (112) of an industrial plant when executed within the container; providing (330) the received plant process data (112) to the container runtime environment (122) and processing (112) the provided plant process data by the processing program; The method (300).

14. A computer program for processing plant process data (112) acquired in relation to a production process of an industrial plant including a plant control system (110), the computer program comprising a program code for causing an edge computing device (120) according to any one of claims 1 to 11 to perform the method (300) according to claim 13.

Images