Method for monitoring and / or controlling one or more chemical plants

Abstract

A method is disclosed for monitoring and / or controlling a chemical plant (12) having a plurality of assets via a distributed computing system (10) having more than two deployment layers (14, 16, 30, 32, 34) wherein the deployment layers (14, 16, 30, 32, 34) comprise at least two of a first processing layer (14), a second processing layer (16, 32, 34) and an external processing layer (30), the method comprising the steps of:-providing (60) a containerized application (48, 50), the method comprises the steps of:-deploying (62) a containerized application (48, 50) comprising an asset or plant template specifying input data, output data and an asset or plant model,-deploying (62) the containerized application (48, 50) for execution on at least one of deployment layers (14, 16, 30, 32, 34), where the deployment layers (14, 16, 30, 32, 34) are allocated based on the input data, load indicators or system layer tags, and-deploying (62) the containerized application (48, 50) for execution on the at least one of the deployment layers (14, 16, 30, 32, 34) on the basis of the distributed deployment layers (14, 16, 30, 32, 34). The containerized application (46, 52, 54) is executed on the controller (34) to generate output data for controlling and / or monitoring the chemical plant (12),-the generated output data is provided (66) for controlling and / or monitoring the chemical plant (12).

Description

[0001] This application is a divisional application of the patent application with application number 202080086066.0 and invention title "Method for Monitoring and / or Controlling One or More Chemical Plants", filed on December 8, 2020. Technical Field

[0002] The present disclosure relates to a method for monitoring and / or controlling a chemical plant having a plurality of assets via a distributed computing system having a plurality of deployment layers. Background Art

[0003] Chemical production is a highly sensitive production environment, especially in terms of safety. A chemical plant typically includes a plurality of assets for producing chemical products. A plurality of sensors are distributed in such plants for monitoring and control purposes and for collecting a large amount of data. Thus, chemical production is an environment with a large amount of data. However, so far, the benefits of improving the production efficiency of one or more chemical plants from these data have not been fully utilized.

[0004] Therefore, there has been a great deal of interest in applying new technologies in cloud computing and big data analytics. However, different from other manufacturing industries, the process industry needs to comply with very high safety standards. For this reason, the computing infrastructure is usually isolated, and access to the monitoring and control systems is highly restricted. Due to such safety standards, latency and availability considerations conflict with simply migrating the so-far embedded control systems to, for example, cloud computing systems. Bridging the gap between highly proprietary industrial manufacturing systems and cloud technologies is one of the major challenges.

[0005] WO2016065493 discloses a client device and system for data acquisition and preprocessing of process-related massive data from at least one CNC machine or industrial robot and sending the process-related data to at least one data receiver, such as the cloud-based server. The client device includes at least one first data communication interface with at least one controller of the CNC machine or industrial robot, the data communication interface for continuously recording hard real-time process-related data via at least one real-time data channel, and for recording non-real-time process-related data via at least one non-real-time data channel. The client device further includes at least one data processing unit that maps at least the recorded non-real-time data to the recorded hard real-time data to aggregate a set of context-ualized process-related data. In addition, the client device includes at least one second data interface for sending the set of context-ualized process-related data to the data receiver and for further data communication with the data receiver.

[0006] WO2019138120 discloses a method for improving a chemical production process. At corresponding chemical production facilities, a plurality of derivative chemical products are produced through a derivative chemical production process based on at least some derivative process parameters. Each of these chemical production facilities includes a separate corresponding in-facility intranet. At least some of the corresponding derivative process parameters are measured from the derivative chemical production process by corresponding production sensor computer systems within each in-facility intranet. A process model for simulating the derivative chemical production process is recorded in a process model management computer system external to the in-facility intranet.

[0007] US20160320768A1 discloses an example network environment for monitoring a factory process, in which a system computer operates as a root cause analyzer. The system computer communicates with a data server to access data of measurable process variables collected from a historical database. The data server is communicatively coupled to a distributed control system (DCS), which in turn transmits the collected data to the data server through a communication network.

[0008] The object of the present invention relates to a highly scalable and flexible method for monitoring and / or controlling a chemical plant in a process industry, which follows high safety standards and allows enhanced monitoring or control. Summary of the Invention

[0009] A method for monitoring and / or controlling a chemical plant having a plurality of assets via a distributed computing system having more than two deployment layers is proposed. The deployment layers include at least two of a first processing layer, a second processing layer, and an external processing layer. The method includes the following steps:

[0010] - Providing a containerized application including an asset or plant template specifying input data, output data, and an asset or plant model,

[0011] - Deploying the containerized application to execute on at least one of the deployment layers, wherein the assignment of the deployment layer depends on the input data, a load indicator, or a system layer tag, and executing the containerized application on the corresponding deployment layer to generate output data for controlling and / or monitoring the chemical plant,

[0012] - Providing the generated output data for controlling and / or monitoring the chemical plant.

[0013] A system for monitoring and / or controlling a chemical plant having a plurality of assets by using more than two deployment layers is proposed, wherein the deployment layers include at least two of a first processing layer, a second processing layer, and an external processing layer, and the system is configured to:

[0014] - Providing a containerized application including an asset or plant template specifying input data, output data, and an asset or plant model,

[0015] - Deploy containerized applications to execute on at least one of the deployment layers in the deployment layer, wherein the deployment layer is assigned based on input data, a load indicator, or a system layer tag, and the containerized applications are executed on the assigned one or more deployment layers to generate output data for controlling and / or monitoring the factory.

[0016] - Provide the generated output data for controlling and / or monitoring the chemical plant.

[0017] The present invention further relates to a distributed computer program or a computer program product having computer-readable instructions that, when executed on one or more processors, cause the one or more processors to perform a method for monitoring and / or controlling one or more chemical plants described herein. The present invention further relates to a computer-readable non-volatile or non-transitory storage medium having computer-readable instructions that, when executed on one or more processors, cause the one or more processors to perform a method for monitoring and / or controlling one or more chemical plants as described herein.

[0018] The proposed method allows for efficient application processing in a distributed computing system for controlling and / or monitoring a chemical plant. By introducing different process and storage layers, the massive data transfer, orchestration, and execution of applications can be distributed across different layers, enabling flexible application processing. In addition, the concept of three system layers allows for highly available and secure monitoring and / or control because the second processing and external management layers are redundant. In other words, more critical tasks can be assigned to local computing resources that do not depend on an external network, while less critical tasks can be assigned to external computing resources. Another advantage is that, based on context, the method is capable of automating application deployment and automatically identifying the need for retrofitting additional sensors or IoT sensors.

[0019] Furthermore, the proposed method can be adapted to multiple chemical plants via the second processing layer or the external processing layer. Thus, the method enables highly scalable application orchestration for more reliable and enhanced monitoring and / or control of chemical plants. In particular, the orchestration of containerized applications in different application environments can be organized to meet the specific requirements of the process industry. For example, the deployment of containerized applications for obtaining input data can be simplified for multiple assets even in multiple plants. In addition, based on the specific data required by the containerized application and the computing resources required to run such an application, an appropriate processing layer can be selected, thereby adhering to the high availability standards of the chemical plant. For example, a computationally intensive application specific to a plant can be executed on the second processing layer, while a process application that obtains asset- or process-specific data and requires low latency can be executed on the first processing layer. Other criteria can be defined for when and on which deployment layer to orchestrate.

[0020] The following description relates to the above - mentioned systems, methods, computer programs, and computer - readable storage media. In particular, the systems, input units, computer programs, and computer - readable storage media are configured to perform the method steps as described above and further described below.

[0021] In the context of the present invention, a chemical plant refers to any manufacturing facility based on chemical processes, such as using chemical processes to convert raw materials into products. Compared with discrete manufacturing, chemical manufacturing is based on continuous or batch processes. Thus, the monitoring and / or control of a chemical plant is time - related and thus based on large time - series data sets. A chemical plant can include more than 1000 sensors that generate measurement data points every few seconds. Such dimensions result in the need to process terabytes of data in systems used for controlling and / or monitoring chemical plants. A small chemical plant can include several thousand sensors that generate data points every 1 to 10 seconds. For comparison, a large chemical plant can include tens of thousands of sensors, e.g., 10000 to 30000, that generate data points every 1 to 10 seconds. Contextualizing such data results in processing hundreds of gigabytes to terabytes of data.

[0022] A chemical plant can produce products via one or more chemical processes that convert raw materials into products via one or more intermediate products. Preferably, a chemical plant provides an encapsulated facility for producing products that can be used as raw materials in subsequent steps of a value chain. A chemical plant can be a large - scale plant, such as an oil and gas facility, a gas purification plant, a carbon dioxide capture facility, a liquefied natural gas (LNG) plant, a refinery, a petrochemical facility, or a chemical facility. For example, in petrochemical production, upstream chemical plants include steam cracking units that start with the processing of naphtha into ethylene and propylene. These upstream products can then be provided to further chemical plants to derive downstream products, such as polyethylene or polypropylene, which can again be used as raw materials for chemical plants that derive further downstream products. Chemical plants can be used to manufacture discrete products. In one example, a chemical plant can be used to manufacture precursors for polyurethane foam. Such precursors can be provided to a second chemical plant to manufacture discrete products, such as insulation panels containing polyurethane foam.

[0023] The value - chain production from various intermediate products to final products can be dispersed at different locations or integrated into an integrated site or chemical industrial park. Such an integrated site or chemical industrial park includes a network of interconnected chemical plants where products produced in one plant can be used as raw materials for another plant.

[0024] A chemical plant may include multiple assets such as heat exchangers, reactors, pumps, pipelines, distillation or absorption towers, etc. In a chemical plant, some assets may be critical. Critical assets are those that, when interrupted, will seriously affect the plant operation. The interruption of such assets may cause damage to the manufacturing process. It may lead to a decline in product quality or even a halt in manufacturing. In the worst-case scenario, fire, explosion, or release of toxic gases may be the result of such an interruption. Therefore, depending on the chemical process and the chemicals involved, such critical assets may require more stringent monitoring and / or control than other assets. To monitor and / or control the chemical process and assets, multiple actors and sensors may be embedded in the chemical plant. Such actors or sensors can provide process or asset-specific data related to, for example, the status of individual assets, the status of individual actors, the composition of chemicals, or the status of the chemical process. In particular, the process or asset-specific data includes one or more of the following data categories:

[0025] - Process operation data, such as the composition of raw materials or intermediate products,

[0026] - Process monitoring data, such as flow rate, material temperature,

[0027] - Asset operation data, such as current, voltage, and

[0028] - Asset monitoring data, such as asset temperature, asset pressure, vibration.

[0029] In the context of the present disclosure, an asset may include any component of a chemical plant, such as equipment, instruments, machines, processes, or process components. Therefore, an asset model may relate to a machine, equipment, instrument, process, or process component model.

[0030] Process or asset-specific data refers to data that is related to a specific asset or process and is contextualized with respect to such specific asset or process. Process or asset-specific data may be contextualized only with respect to individual assets and processes. Process or asset-specific data may include measurement values, data quality measurements, time, measurement units, an asset identifier of a specific asset, or a process identifier of a specific process part or stage. Such process or asset-specific data may be collected at the lowest processing layer or the first processing layer and is contextualized with respect to a specific asset or process in an individual plant. Such contextualization may involve the context available at the first processing layer. Such context may be related to an individual plant.

[0031] Plant-specific data refers to process or asset-specific data that is contextualized for one or more plants. Such plant-specific data can be collected at a second processing layer and contextualized for multiple plants. Specifically, the contextualization can involve the context available at the second processing layer. Via contextualization, context such as plant identifiers, plant types, reliability indicators, or alarm limits of the plant can be added to process or asset-specific data points. In a further step, the technical asset structure of one or more plants, integrated sites, or chemical industrial parks, other asset management structures (such as asset networks), or application contexts (such as model identifiers, third-party exchanges) can be added. This overall context can be derived from functional locations or digital twins, such as digital piping and instrumentation diagrams, 3D models, or scans with the xyz coordinates of plant assets. Additionally or alternatively, local scans from mobile devices linked to, for example, piping and instrumentation diagrams can be used for contextualization.

[0032] In particular, plant-specific data related to the interfaces between chemical plants in a manufacturing chain can be provided at the second processing layer or an external processing layer. As a result, monitoring and / or control, such as via anomaly detection, setpoint guidance, and optimization in a chain across multiple plants, can be enhanced. To monitor and / or control a chain across multiple plants, process applications with online input / output data profiles can be used. Such data and process applications can be transferred between plants. In combination with the quality and energy balances that can be monitored, such process applications can optimize the entire chain across multiple chemical plants, rather than individual plants in the chain.

[0033] The contexting process refers to linking data points available in one or more storage units. One or more such units can be persistent or non-volatile storage. The data points may relate to measurement values or context information. The (one or more) storage units can be part of a first processing layer, a second processing layer, an external processing layer, or distributed over two or more of these layers. The linking can be generated dynamically or statically. For example, predefined or dynamically generated scripts can generate dynamic or static links between information data points in one processing layer or across processing layers. Links can be established by generating a new data object that includes the link data itself and storing such new data object in a new instance. Any data points of the stored data may be actively deleted if a copy is stored elsewhere. Thus, any data points of the new data object copied from one storage unit to the same or another storage unit can be deleted to reduce storage space. Additionally or alternatively, links can be established by generating a metadata object with embedded links to address or access the corresponding data points in one or more distributed storage units. Thus, any data points that can be addressed or accessed through the metadata object can remain in their original storage unit. For example, on the external processing layer, it is still possible to perform the linking of this information to form new data objects. For the retrieval of data, either the data object is accessed directly or the metadata object is used to address or access data distributed over one or more storage units. Any operations on such data (such as an application) can access such data directly or access a non-persistent mirror of such data, such as from a cache memory or a persistent copy of the data.

[0034] In this context, a containerized application refers to a process application that can be executed in an encapsulated runtime environment independent of the host operating system. Thus, the application can be considered to run in a sandbox. A containerized application can be based on a container image that contains the application. The container image can include software components, such as a hierarchical tree of software components required to execute the corresponding application in the encapsulated runtime environment. Such containerized applications can be stored in or associated with a registry in the second processing layer or the external processing layer.

[0035] To deploy a containerized application, an orchestration application associated with the second processing layer or the external processing layer can manage the execution of the containerized application. Such management can include general runtime environment configuration, such as storage or networking for running the containerized application. Such management can further include host allocation, which defines the distribution between a central master node or one or more compute nodes to execute the application on the first processing layer, the second processing layer, or the external management layer. Specifically, such allocation of computing resources depends on input data, load indicators, or system layer tags.

[0036] The input data can include real-time data from sensors (such as wireless monitoring devices or IoT devices), non-real-time data, or output data of deployed containers or executed applications. Such data may relate to machinery (such as mechanical types or sensor data regarding mechanical measurements), chemicals (such as chemical types or sensor data regarding chemical composition measurements processed in a chemical plant), processes (such as chemical process types or sensor data measured relative to chemical processes executed in a chemical plant), and / or factories (such as factory types or sensor data regarding measurements in a chemical plant, such as environmental measurement data).

[0037] The asset or factory model can include data-driven or kinetic models, such as providing health status, operation prediction, event prediction, or event triggers. The asset or factory model can be based on a pure data-driven model, a hybrid model combining data-driven and kinetic models, or a pure kinetic model. The asset or factory model can further be based on a scenario matrix that maps input data (such as sensor data) to specific events. The asset model can reflect the physical behavior of a single or multiple assets. The factory model can reflect the physical behavior of a part of one or more factories, a complete factory, or multiple factories.

[0038] The output data can include key performance indicators related to assets, factories, input data, asset model performance, or factory model performance. The asset model performance or factory model performance can be embedded in the asset or factory model hosted by a containerized application. Any generated output data of the method can be used as input data in one or more further containerized applications. In this way, a chain of containerized applications can be realized to build a system covering the system and use the generated output data to control and / or monitor one or more chemical plants. Such chemical plants can be parallel manufacturing plants or plants connected along the value chain.

[0039] The containerized application can include one or more operations to obtain input data, provide the input data to the corresponding asset or factory model to generate output data, and provide the generated output data for controlling and / or monitoring the chemical plant. Such output data can be passed to a persistent instance after the application execution. In particular, such output data can be passed to a control instance, such as on the first processing layer of the chemical plant. Additionally or alternatively, such output data can be passed to a monitoring instance on the first processing layer, the second processing layer, or an external processing layer. The output data can be passed to a client application, such as for display to an operator, or a further containerized application for execution.

[0040] In one aspect, the second processing layer includes greater storage and computing resources than the first processing layer. Alternatively or additionally, the external processing layer may include greater storage and computing resources than the second processing layer. This stacked resource structure helps bridge the gap between the embedded control systems in chemical plants and the available cloud technologies. In particular, the embedded control systems in chemical plants are limited in terms of storage and processing capabilities. Expanding such resources allows for enhanced monitoring and / or control.

[0041] The first processing layer and the second processing layer may be hosted, placed, configured within a secure network or internally. The first processing layer may be communicatively coupled to the second processing layer. The first processing layer may include at least one core process system associated with a chemical plant or an individual chemical plant. Preferably, the first processing layer is configured to control and / or monitor chemical processes and assets at the asset level of a separate plant. Thus, the first processing layer monitors and / or controls the chemical plant at the lowest level. Further preferably, the first processing layer is configured to monitor and control critical assets. Critical assets are those assets that, when interrupted, have a severe impact on plant operations. Such asset interruptions may cause the manufacturing process to be compromised. There may be a decline in product quality or even a halt in manufacturing. In the worst-case scenario, fires, explosions, or toxic gas releases may be the result of such interruptions. Therefore, depending on the chemical process and the chemicals involved, such critical assets may require more stringent monitoring and / or control than other assets.

[0042] Additionally or alternatively, the second processing layer may be configured to provide data to an external network, for example, via an interface with the external network. The second processing layer may be communicatively coupled to the external processing layer via the external network. Adding stacked processing layers within a secure network or internally can allow for compliance with the high safety standards of the chemical industry. In particular, this architecture allows the method to be executed completely independently of the external management layer, thus enabling an island mode for one or more chemical plants. Here, the island mode refers to monitoring and / or controlling a chemical plant without accessing the external network.

[0043] On the other hand, the first processing layer is configured to provide asset - or process - specific data, while the second processing layer is configured to provide plant - specific data. The second processing layer may be configured to contextualize the asset - or process - specific data. In this way, the performance of the first processing layer is not affected. Since the typical core process systems of the first processing layer, especially in older plants, do not have the required computing power, adding a further system with higher performance can even enable the contextualization. Additionally, the second processing layer allows data contextualization at the plant level rather than the asset level. Data contextualization in this context involves adding context information to the asset - or process - specific data or reducing the data size by pre - processing the asset - or process - specific data. Adding context may include adding one or more further information tags to the asset - or process - specific data. Pre - processing may include filtering, aggregating, normalizing, averaging, or inferring the asset - or process - specific data.

[0044] In one aspect, the first processing layer is associated with one or a single chemical plant. The first processing layer may be a core process system including one or more processing devices and storage devices. Such a layer may include one or more distributed processing and storage devices that form a programmable logic controller (PLC) system or a distributed control system (DCS) with control loops distributed throughout the chemical plant. Preferably, the first processing layer is configured to control and / or monitor chemical processes and assets at the asset level. Thus, the first processing layer monitors and / or controls the chemical plant at the lowest level. Additionally, the first processing layer may be configured to monitor and control critical assets. Additionally or alternatively, the first processing layer is configured to provide process - or asset - specific data to the second processing layer. Such data may be provided directly or indirectly to the second processing layer.

[0045] In another aspect, the second processing layer is associated with more than one chemical plant. The second processing layer may include a process management system having one or more processing and storage devices. Preferably, the second processing layer or the process management system is configured to manage data transmission to and / or from the first processing layer. Further preferably, the second processing layer or the process management system is configured to host and / or orchestrate process applications. Such process applications may monitor and / or control one or more chemical plants or one or more assets. The process management system may be associated with one or more chemical plants. In other words, the process management system may be communicatively coupled to a plurality of first processing layers associated with one or more chemical plants.

[0046] On the other hand, the second processing layer may include an intermediate processing system and a process management system. Here, the intermediate processing system may be communicatively coupled to the first processing layer, preferably the core process system, and the process management system may be communicatively coupled to the intermediate layer. Preferably, the first processing layer and the process management system are coupled or communicatively coupled via the intermediate processing system. The intermediate processing system may be configured to collect process or asset-specific data provided by the first processing layer. The process management system may be configured to provide plant-specific data of one or more chemical plants to an interface with an external network. The intermediate processing system may be associated with one or more chemical plants. In other words, the intermediate processing system may be communicatively coupled to the first processing layer of a chemical plant or the multiple first processing layers of multiple plants. The process management system may be communicatively coupled to one or more intermediate processing systems. Adding the intermediate processing level to the second processing layer adds a further security layer. It completely separates the virus-infected first processing layer from any external network access. In addition, the intermediate layer allows for further enhanced data processing capabilities by reducing the data transfer rate to the external processing layer via preprocessing and allows for improved data quality through contextuality. The intermediate processing system and the process management system may include one or more processing and storage devices.

[0047] On the other hand, a secure network is an isolated network that includes more than two secure zones separated by firewalls. Such firewalls may be network-based or host-based virtual or physical firewalls. The firewall may be hardware-based or software-based to control incoming and outgoing network traffic. Here, predefined rules in the sense of a whitelist may define the allowed traffic via access management or other configuration settings. Depending on the firewall configuration, the secure zones may follow different security standards.

[0048] On the other hand, the first processing layer is hosted, placed, or configured in or within the first security zone via the first firewall, while the second processing layer is hosted, placed, or configured in or within the second security zone via the second firewall. To securely protect the first processing layer, the first security level may follow a higher security standard than the second security level. The security levels may follow common industry standards, such as those described in the Namur document IEC 62443. The second processing layer may provide further isolation via a security zone. For example, an intermediate processing system may be hosted, placed, or configured in or within the third security zone via the third firewall, while a process management system may be configured in the second security zone via the second firewall. The third and second security zones may also be staggered in terms of security standards. For example, the third security zone may follow a higher security standard than the second security zone. This allows for a higher security standard on the lower security zone of the first processing layer and a lower security standard on the higher security zone of the second processing layer. In one embodiment, the first processing layer is inside the first security zone, the process management system is inside the second security zone, and the intermediate processing system is inside the third security zone.

[0049] The second processing layer may be configured to contextualize process or asset-specific data. This way, the performance of the first processing layer is not affected. Since typical core process systems in old factories do not have the required computing power, adding a further system with higher performance can even enable contextualization. Additionally, the second processing layer, especially the intermediate processing system, allows for data contextualization at the factory level rather than the asset level. Data contextualization in this context involves adding context information to process or asset-specific data or reducing the data size by preprocessing process or asset-specific data. Adding context may include adding one or more further information tags to process or asset-specific data. Preprocessing may include filtering, aggregating, normalizing, averaging, or inferring process or asset-specific data.

[0050] On the other hand, one-way or two-way communication can be implemented for the data flow between different processing layers, such as data transfer or data access. In other words, the system can be configured to allow one-way or two-way communication, such as data transfer or data access between different processing layers. A data flow can include process or asset-specific data from a first processing layer, which is passed to a second processing layer and contextualized via the second processing layer and then transmitted to an external processing layer. Contextualization can be performed on the second processing layer, the external processing layer, or both. In other words, the second processing layer, the external processing layer, or both can be configured to contextualize process or asset-specific data or plant-specific data. Additionally, depending on the criticality of the process or asset-specific data or plant-specific data, such data can be allocated for one-way or two-way communication. In other words, the system can be configured to allocate one-way or two-way communication to process or asset-specific data or plant-specific data based on the criticality of the process or asset-specific data or plant-specific data. For example, by implementing a diode-type communication channel, data communication from the second or external processing layer to a critical asset can be prohibited. Such communication can only allow one-way communication from the critical asset to the processing layer and vice versa.

[0051] On the other hand, critical or non-critical data can be assigned to the data flow. In other words, the system can be configured to assign critical or non-critical data tags. Critical data refers to data that is critical to the operation of a chemical plant, such as short-term data from which the operating points of the chemical plant are derived. Such critical data can cover a short-term range, such as a few hours or days out of one or more weeks, which is required for the plant to operate in its optimized state. Non-critical data refers to data that is not critical to the operation of a chemical plant, such as medium- and long-term data for monitoring the chemical plant based on medium- and long-term behavior. Such non-critical data can cover a medium- and long-term time range, such as weeks or months to a year or several years, which is required, for example, to monitor and / or control one or more assets or one or more plants over a period of time. Such data can also be referred to as cold, warm, and hot data, where hot data corresponds to critical data, warm data corresponds to medium-term non-critical data, and cold data corresponds to long-term non-critical data.

[0052] On the other hand, data contextualization is interleaved across the system layer, the processing layer, or processing systems included in such a processing layer, with each layer mapping the context information available in the corresponding layer. In other words, the system can be configured to interleave data contextualization across the system layer, the processing layer, or processing systems included in such a processing layer, where each layer maps the context information available in the corresponding layer. The interleaving can include the contextualization of asset- or process-specific data at different levels, thereby adding context information to the single-plant level and / or the multi-plant level. In a layered system architecture, the context information available in one layer can be mapped to the data provided by a lower layer or a processing system. Here, lower means closer to the chemical plant data access. For example, the process- or asset-specific data provided by the first processing layer can include context information at the asset level. In other words, the first processing layer can be configured to provide process- or asset-specific data that includes context information at the asset level. Such context information can be related to real-time information (such as measurements, measurement quality, product quality, batch-related data, or measurement time). The context information at the asset level can further be related to asset-specific information, such as asset identifiers, internal logistics, or measurement unit identifiers. The intermediate system and the process management system can be configured to add further context information to or contextualize such process- or asset-specific data. Such context information can be related to the plant level rather than the asset level. For example, the context information relates to plant context such as plant identifier, plant type, reliability indicator, alarm limit, etc., or application context such as model identifier, third-party exchange identifier, confidentiality identifier, etc. In this way, the data quality can be improved to maximize the context, and the data management via process applications and the resulting monitoring and / or control capabilities can be enhanced.

[0053] On the other hand, the intermediate processing system is configured to contextualize data by mapping heterogeneous process- or asset-specific data to a homogeneous data format at the plant level. In this context, heterogeneous refers to data that is individually related to the asset level or multiple assets, while homogeneous data refers to data that is related to a portfolio of assets or the same type of assets in a plant. The intermediate processing system can be configured to provide such plant-specific data to the process management system. The process management system can further be configured to contextualize the plant-specific data provided by the intermediate processing system, preferably at the multi-plant level. Such contextualization can include adding context information at the multi-plant level or the site level, such as multi-plant or site context, including the technical asset structure of one or more plants or asset management information such as an asset network. Additionally or alternatively, such contextualization can include adding application context, such as model identifier, third-party exchange identifier, or confidentiality identifier.

[0054] Additionally or alternatively, the first processing layer may be configured to provide asset or process specific data to the second processing layer. Such data may be provided directly or indirectly to the second processing layer. The second processing layer may be associated with one or more factories. The second processing layer may include a process management system and an optional intermediate processing layer. The first processing layer may include a factory specific core process system. One or more core process systems may optionally be communicatively coupled to the process management system via the intermediate processing layer. The second processing layer (in particular the process management system) and the external processing layer may be configured to contextualize, store or aggregate data from one or more chemical plants, and / or to orchestrate process models for one or more chemical plants.

[0055] In another aspect, the intermediate processing system is configured to contextualize data by mapping heterogeneous process or asset specific data to a homogeneous data format at the factory level. In this context, heterogeneous refers to data that is individually related to the asset level or multiple assets, while homogeneous data refers to data that is related to a portfolio of assets or the same type of assets in a factory. The intermediate processing system may be configured to provide such factory specific data to the process management system. The process management system may further be configured to contextualize the factory specific data provided by the intermediate processing system, preferably at the multi-factory level. Such contextualization may include adding context information at the multi-factory level or site level, such as multi-factory or site context, including the technical asset structure of one or more factories or asset management information such as an asset network. Additionally or alternatively, such contextualization may include adding application context, such as model identifiers, third party exchange identifiers or confidentiality identifiers.

[0056] The second processing layer, preferably the process management system, may be communicatively coupled to the external processing layer via an external network. The second processing layer, preferably the process management system, may be configured to manage the data transfer to and / or from the external processing layer in real time or on demand. The second processing layer, preferably the process management system, may be configured to provide factory specific data to an interface with the external network, for example based on an identifier added by contextualization. Such an identifier may be a confidentiality identifier, based on which such data is not provided to the interface of the external network.

[0057] The external processing layer may be a computing or cloud environment that provides virtualized computing resources such as data storage and computing power. The external processing layer may provide a private, hybrid, public, community or multi-cloud environment. Cloud environments are advantageous as they provide on-demand storage and computing power. Additionally, in cases where multiple chemical plants operated by different parties are to be monitored and / or controlled, data or process applications affecting the chemical plants may be shared in such a cloud environment.

[0058] On the other hand, a second processing layer, preferably a process management system, is configured to provide plant-specific data from one or more chemical plants to an external processing layer. The second processing layer, preferably the process management system, may further be configured to delete at least a portion of the data transmitted to the external processing layer. The external processing layer may be configured to store historical data from one or more chemical plants. The external processing layer may be configured to aggregate, store, or contextualize plant-specific data from multiple chemical plants and / or store historical data from more than one chemical plant. Aggregation here refers to grouping data via aggregation functions such as summation, average, or modulo. Thus, aggregation is related to the function of reducing dimensionality or storage space. In this way, data storage can be externalized, the required local storage capacity can be reduced, and historical transmissions become redundant. Additionally, due to the flexible computing and storage resources of the external processing layer and the fact that data is available in the external processing layer, process applications can be built, trained, tested, or modified in the external processing layer.

[0059] On the other hand, a second processing layer, preferably a process management system, is configured to manage data transmission to and / or from the external processing layer in real time or on demand. Real-time transmission may be buffered according to the network and computing load on the interface with the external network. On-demand transmission may be triggered in a predefined or dynamic manner. Preferably, data transmission to the external processing layer is managed in real time, and transmission from the external processing layer is managed on demand.

[0060] On the other hand, a second processing layer, preferably a process management system, is configured to store or manage access to historical data, real-time data, and scheduled data. On the other hand, a second processing layer, preferably a process management system, is configured to store or manage access to historical data for a first time window, and the external processing layer is configured to store historical data for a second time window, where the first time window is shorter than the second time window. Here, the first time window may correspond to a critical time window that allows the system to monitor and / or control a chemical plant in island mode without an external network connection. The first time window may be regarded as a hot window, and the historical data for this window is required to safely control and / or monitor the chemical plant. The first time window or hot window may be determined based on the storage capacity of the second processing layer (preferably the process management system), or preferably by the process application and the historical data required to execute on the process application in island mode without an external network connection. This approach always ensures the availability of the system for monitoring and / or control.

[0061] On the other hand, the deployment is managed by an orchestration application that manages the deployment of containerized applications based on input data, load indicators, or system layer tags. Additionally or alternatively, the orchestration application is hosted by a second processing layer and / or an external processing layer. On the other hand, the orchestration application hosted by the second processing layer at execution manages critical containerized applications. Additionally or alternatively, the orchestration application hosted by the external processing layer at execution manages non-critical containerized applications. This execution-time management can be assigned statically or dynamically. In a dynamic scenario, the second processing layer can host a backup of the critical containerized application, and the orchestration application hosted by the second processing system can access this backup if the external network connectivity is interrupted. Here, critical containerized applications refer to those containerized applications that monitor and / or control critical assets. Therefore, such applications are required if monitoring and / or control needs to operate in island mode.

[0062] On the other hand, the management of critical containerized applications is assigned to the second processing layer based on historical criteria reflecting a time window of historical data available on the first processing layer or the second processing layer. The second processing layer can be configured to store historical aggregated data for a first time window, and the external processing layer can be configured to store historical aggregated data for a second time window, where the first time window is shorter than the first time window. In a preferred embodiment, the first time window is selected such that critical containerized applications can be executed on the first processing layer or the second processing layer. On the other hand, containerized applications are deployed to execute on the second processing layer or the external management layer according to historical criteria reflecting a time window of available historical data. In such an embodiment, the application can be executed at the level where such data is available. Therefore, no further data transfer between processing layers is required, thus reducing communication and processing load. Combining the cross-layer interleaved context concept, processing factory-specific data on the first processing layer would introduce redundant data transfer, once from the first processing layer to the second processing layer for context, and then back to the first processing layer for application execution.

[0063] The deployment can depend on the input data. On the other hand, the assignment of the deployment layer depends on data availability indicators, criticality indicators, or latency indicators.

[0064] The data availability indicator can be related to the input data obtained by the containerized application. Based on such an indicator, the execution can be assigned to the deployment layer where the data is directly available or stored. For example, the first processing layer can be configured to provide asset or process-specific data, while the second processing layer can be configured to provide factory-specific data. Containerized applications that obtain asset or process-specific data can be deployed on the first processing layer. Similarly, containerized applications that obtain factory-specific data can be deployed on the second processing layer. The application can be executed in the processing layer that hosts the data to avoid redundant data transfer and reduce the load.

[0065] The criticality indicator can be a static or dynamic indicator. In the case of a static indicator, the criticality of an asset, asset group, or plant can be predefined. In the case of a dynamic indicator, the criticality indicator can be dynamically assigned based on the output data of a previous application run or other application runs. For example, the criticality indicator can be determined based on the critical performance parameters of an asset (such as its health status). If the health status of an asset becomes critical over time, then the criticality criteria may change, and the containerized application may run in different deployment layers due to this change, thereby reducing, for example, data transfer latency. In one example, if the execution is assigned to the second processing layer instead of the first processing layer, the containerized application obtains the asset or process-specific data of a particular asset, and the criticality indicator indicates completion. If the health condition of a particular asset changes and may require, for example, closer monitoring at a higher frequency, the criticality indicator can be reset to indicate that the criticality indicator is not completed if the execution is assigned to the second processing layer. In this case, the application can be assigned to the first processing layer.

[0066] The latency indicator can be a static or dynamic indicator. In the case of a static indicator, the latency requirements of an asset, asset group, a part of a plant, or a plant can be predefined. In the case of a dynamic indicator, the latency requirements can be dynamically assigned based on the signature of the input data. Such a signature can be related to, for example, the frequency of change of real-time measurement data derived from historical real-time measurement data. In one example, the monitoring signal of a pump can show a higher frequency than the monitoring signal of a heat exchanger. In this case, the containerized application for monitoring the pump can be directly deployed at the asset level on the pump controller or on the core process system of the corresponding plant. For example, the corresponding containerized application can be deployed in the processing layer as close as possible to the pump to reduce latency. Thus, the latency criteria can represent the time urgency of the containerized application.

[0067] The load indicator can be based on the processing and / or network load of the corresponding deployment layer. Additionally or alternatively, the application can be executed on a deployment layer that provides sufficient computing and storage resources to reduce the processing load on other processing layers and ensure that the execution of critical applications is not affected. In this case, the input data can be transferred to the corresponding deployment layer. This is particularly advantageous for applications that require a high computing load or have lower time requirements, thus allowing data transfer latency.

[0068] On the other hand, the deployment can depend on system layer tags associated with the containerized application. The system layer tags can be, for example, configurations in an orchestration application that deploys, executes, and monitors the containerized application. In this case, the containerized application can include an application identifier that the orchestration application can use for identification at deployment. In this way, the deployment can be "hard-wired" to ensure that critical applications are executed in the correct layer. Such a deployment scenario can be particularly relevant for containerized applications that monitor and / or control critical assets. In this context, critical assets refer to assets whose interruption would severely affect plant operations. Here, plant-specific data refers to context-specific assets or process-specific data.

[0069] On the other hand, the containerized application is deployed to multiple assets or multiple factories of the same type. Assets of the same type may be related to assets with similar functions, from the same vendor, and / or with similar characteristics (such as performance characteristics). Factories of the same type can refer to factories that produce similar intermediate or final products, have similar physical asset structures, or are based on similar chemical processes. Here, similar means comparable in the sense that the behavior exhibited by the assets or factories does not deviate by more than the tolerance or can be modeled by a single model. Preferably, the containerized application is associated with an asset or factory identifier. Such an identifier can be a configuration setting of the orchestration application or the containerized application. The factory or asset identifier can be one-dimensional representing one factory or asset type or multi-dimensional representing multiple factories or assets for which the containerized application will be executed. Especially if the first processing layer has different processing unit factories or even asset- or process-specific, such asset identifiers allow for the simultaneous processing of assets associated with different processing units, thus enabling the efficient deployment of the containerized application.

[0070] On the other hand, the containerized application is modified based on the input data and output data of the containerized application executed for multiple assets or factories of the same type. This can be done periodically or dynamically at each defined time. By aggregating this data, the containerized application, especially the factory or asset model, can be verified or improved in terms of accuracy. In this way, the model can be adjusted to reflect the behavior of the physical factory or asset. This is particularly important for the maintenance cycle or the life cycle because the physical factory or asset may change its behavior according to the stages of such cycles. Such changes can be automatically compensated and self-optimized by the proposed system.

[0071] On the other hand, containerized applications are monitored based on the confidence of input data, asset models, or factory models. The containerized applications can provide such confidence as output data. Corresponding operations can be embedded in the containerized applications via the asset or factory models or via separate models. The confidence for the input data can be generated by analyzing patterns in, for example, real-time measurement data. The confidence for the asset or factory models can be generated as part of performing model operations based on the input data.

[0072] On the other hand, if the confidence drops below a threshold confidence, for example if the confidence is below 70%, 80%, or 90%, an event signal is triggered. Such an event signal can indicate a failure in the operation of the asset or application. If an abnormal pattern is detected in the input data and the corresponding confidence is below the threshold, the application execution can be stopped and a failure signal can be passed to, for example, a control instance on the first processing layer in a chemical plant. Additionally or alternatively, such a failure signal can be passed to a monitoring instance on, for example, the first or second processing layer, or to a client application, for example for display to an operator. In this way, sensor failures and retrofit needs can be detected.

[0073] On the other hand, if the confidence exceeds the threshold, a modification of the asset or factory model is triggered. The modification may be triggered automatically. Such a modification can include hardware or software modifications. For example, the modification includes retraining the asset or factory model based on historical data, adjusting input data channels, adjusting a scenario matrix, or triggering an event signal to maintain the hardware of, for example, an Internet of Things (IoT) device.

[0074] On the other hand, the modification of the asset or factory model is performed on the second processing layer, particularly on a process management system or an external processing layer. Since such modifications do not interfere with the containerized applications first, the calculations can be performed on the external processing layer, thereby reducing the processing load for monitoring and / or controlling more critical processing layers, such as the first and second processing layers.

[0075] On the other hand, external containerized applications from a third-party environment are provided and deployed to execute on the external processing layer. Applications from a third-party environment refer to any applications that are not created in the proprietary systems of the first processing layer, the second processing layer, and the external processing layer. This can reduce the risk of third-party containerized applications infecting the monitoring and / or control systems of a chemical plant.

[0076] In another aspect, creation of a new containerized application is performed in an external processing layer. For example, an asset or plant model can be trained on the external processing layer. Preferably, the external processing layer is configured to store aggregated data for multiple chemical plants. A second processing layer can be configured to provide the aggregated data from one or more chemical plants to the external processing layer. The second processing layer can be configured to transmit the aggregated data to the external processing layer in real time or on demand. The second processing layer can be configured to delete at least some of the data transmitted to the external processing layer. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Example embodiments of the present disclosure are illustrated in the accompanying drawings. However, it should be noted that the accompanying drawings only illustrate specific embodiments of the present disclosure and therefore should not be considered to limit the scope thereof. The technical teachings may encompass other equally effective embodiments.

[0078] Figure 1 A first schematic representation of a system for monitoring and / or controlling one or more chemical plants is shown.

[0079] Figure 2 A second schematic representation of a system for monitoring and / or controlling one or more chemical plants is shown.

[0080] Figure 3 A third schematic representation of a system for monitoring and / or controlling one or more chemical plants is shown.

[0081] Figure 4 Shown Figures 1 to 3 Schematic representation of the concept of data contextualization in the shown system.

[0082] Figure 5 A flow chart showing a schematic representation of a method for monitoring and / or controlling one or more chemical plants is shown.

[0083] Figure 6 A schematic representation of a system for monitoring and / or controlling one or more chemical plants via containerized applications is shown.

[0084] Figure 7 A flow chart showing a schematic representation of a method for monitoring and / or controlling a chemical plant having a plurality of assets is shown.

[0085] Figure 8 Shown is a schematic representation of a system for monitoring and / or controlling a plurality of chemical plants in different secure networks, which are configured for data and application transmission. DETAILED DESCRIPTION

[0086] In the petrochemical industry, process industrial production typically starts with upstream products, which are used to derive further downstream products. To date, the production of the value chain via various intermediate products to the final product has been highly restricted and based on isolated infrastructure. This has hindered the introduction of new technologies such as IoT, cloud computing, and big data analytics.

[0087] Unlike other manufacturing industries, the process industry needs to comply with very high standards, especially in terms of availability and security. For this reason, the computing infrastructure is typically one-way and isolated, with highly restricted access to the chemical plant monitoring and control systems.

[0088] Generally speaking, chemical production plants are embedded in the enterprise architecture in an isolated manner, with different levels for functional separation between operational technology and information technology solutions.

[0089] Level 0 is related to the physical process and defines the actual physical processes in the plant. Level 1 involves intelligent devices used to sense and manipulate the physical process, for example, via process sensors, analyzers, actuators, and related instruments. Level 2 involves control systems used to supervise, monitor, and control the physical process. Real-time control and software; DCS, human-machine interface (HMI); supervisory control and data acquisition (SCADA) software are typical components. Level 3 involves manufacturing operations systems used to manage the production workflow to produce the required products. Batch management; manufacturing execution / operations management systems (MES / MOMS); laboratory, maintenance, and plant performance management systems, historical data, and related middleware are typical components. The time range for control and monitoring can be shifts, hours, minutes, seconds. Level 4 involves business logistics systems used to manage business-related activities of manufacturing operations. ERP is the main system, which establishes the basic plant production plan, material usage, transportation, and inventory levels. The time range can be months, weeks, days, shifts.

[0090] In addition, such structures follow a strict one-way communication protocol and do not allow data to flow into Level 2 or below. Such architectures do not cover the external Internet of the company or enterprise. However, this model remains a fundamental concept within the field of network security. In this context, the challenge lies in leveraging the advantages of cloud computing and big data while still ensuring the established advantages of the existing architecture: namely, the high availability and reliability of the lower-level systems (Level 1 and Level 2) that control the chemical plant and network security.

[0091] The technical teachings presented here allow this framework to be enhanced in a systematic manner for monitoring and / or control changes to introduce new functions compatible with the existing architecture. The present disclosure specifically relates to a highly scalable, flexible, and available computing infrastructure for the process industry, which simultaneously adheres to high security standards.

[0092] Figure 1 Figure 1 shows a first schematic representation of a system 10 for monitoring and / or controlling a chemical plant 12.

[0093] System 10 includes two processing layers, including a first processing layer in the form of a core process system 14 associated with each of the chemical plants 12 and a second processing layer 16 in the form of, for example, a process management system associated with two chemical plants 12. The core process system 14 is communicatively coupled to the second processing layer 16, thereby allowing unidirectional or bidirectional data transfer. The core process system 14 includes a set of distributed processing units associated with the assets of the chemical plant 12.

[0094] The core process system 14 and the second processing layer 16 are configured in secure networks 18, 20, which include two security zones in the schematic representation. The first security zone is at the core process system 14 level, where a first firewall 18 controls the incoming and outgoing network traffic to and from the core process system 14. The second security zone is on the second processing layer 16, where a second firewall 20 controls the incoming and outgoing network traffic to and from the second processing layer 16. This isolated network architecture can protect vulnerable plant operations from cyberattacks.

[0095] The core process system 14 provides process or asset or process-specific data 22 of the chemical plant 12 to the second processing layer 16. The second processing layer 16 is configured to contextualize the process or asset or process-specific data of the chemical plant 12. The second processing layer 16 is further configured to provide plant-specific data 24 of the chemical plant 12 to an interface 26 of an external network. Here, the plant-specific data can refer to the contextualized process or asset or process-specific data.

[0096] The process or asset or process-specific data can include value, quality, time, unit of measurement, asset identifier. Via contextualization, further context can be added, such as plant identifier, plant type, reliability indicator, or alarm limits of the plant. In a further step, the technical asset structure and other asset management (such as asset network) of one or more plants or sites, as well as application context (such as model identifier, third-party exchange), can be added.

[0097] The second processing layer 16 is communicatively coupled to an external processing layer 30 via an interface 26 to an external network. The external processing layer 30 can be a computing or cloud environment that provides virtualized computing resources such as data storage and computing capabilities. The second processing layer 16 is configured to provide plant-specific data 24 from one or more chemical plants 12 to the external processing layer 30. Such data can be provided in real time or on demand. The second processing layer 16 is configured to manage data transfer to and / or from the external processing layer in real time or on demand. For example, the second processing layer 16 can provide plant-specific data 24 to the interface 26 to the external network based on an identifier added through context. Such an identifier can be a confidentiality identifier, based on which such data is not provided to the interface 26 to the external network. The second processing layer 16 can further be configured to delete at least part of the data transferred to the external processing layer 30.

[0098] The external processing layer 30 is configured to aggregate plant-specific data from multiple chemical plants and / or store historical data from multiple chemical plants. In this way, data storage can be externalized, and the required local storage capacity can be reduced, and historical transfers become redundant. In addition, such a storage concept allows historical data to be stored on the second processing layer 16 for a hot window, which is a critical time window that allows the system 10 to monitor and / or control the chemical plant in island mode without an external network connection. This way always guarantees the availability of the system 10 for monitoring and / or control.

[0099] The second processing layer 16 and the external processing layer 30 are configured to host and / or orchestrate process applications. In particular, the second processing layer 16 can host and / or orchestrate process applications related to core plant operations, and the external processing layer 30 can be configured to host and / or orchestrate process applications related to non-core plant operations.

[0100] In addition, the second processing layer 16 and the external processing layer 30 can be configured to exchange data with a third-party external processing layer, orchestrate data visualization, orchestrate computational process workflows, orchestrate data calculations, orchestrate APIs to access data, orchestrate data storage, transfer, and calculations, provide an interactive plant data work environment for users (such as operators), and verify and improve data quality, for example, via integration with a third-party management system.

[0101] Figure 2 A second schematic representation of a system 10 for monitoring and / or controlling one or more chemical plants 12 is shown.

[0102] Figure 2 The system 10 shown is similar to Figure 1 the system shown. However, Figure 2The system has a second processing layer with a process management system 32 and an intermediate processing system 34. The intermediate processing systems 34.1, 34.2 are configured in a secure zone of a secure network via a firewall 40.

[0103] The intermediate processing systems 34.1, 34.2 can be configured to obtain process or asset or process-specific data 22 from a single or multiple chemical plants 12. Such data is contextualized at the plant level in the intermediate processing systems 34.1, 34.2, and plant-specific data 38 can be provided to the process management system 32, where further contextualization can be performed, for example, across the plant level or base level of integration. In this setting, data contextualization is staggered across different system 10 layers, and each layer 14, 34, 32 maps the context information available in the corresponding layer 14, 34, 32.

[0104] Figure 3 A third schematic representation of a system 10 for monitoring and / or controlling one or more chemical plants 12 is shown.

[0105] Figure 3 The system 10 shown in Figure 1 and Figure 2 is similar to the system shown in Figure 3 However, the system of

[0106] Figure 4 includes monitoring devices 44 that are communicatively coupled to the process management system 32 or an external processing layer 30. The monitoring devices 36 can be configured to transmit monitoring data to the process management system 32 or the external processing layer 30. The process management system 32 or the external processing layer 30 can be configured to manage multiple monitoring devices 44. Since such IoT devices are not considered reliable, the monitoring data provided by the monitoring devices 44 can be unidirectionally tagged, and any control loop associated with the chemical plant 12 can include a filter for such tags. Therefore, this data will not be used for any control of the chemical plant 12. Figures 1 to 3 A schematic representation of the data contextualization concept in the system 10 shown in

[0107] Figures 1 to 3 The system 10 of

[0108] ■Collect process or asset or process-specific data,

[0109] ■Interact with the basic automation system from level 2,

[0110] ■Initial contextization (bottom-up approach), where the context is added based on what is known at levels 2 and 1 and what is known within the distributed edge devices,

[0111] The process management system 32 can be configured as a centralized edge computing layer. Such a layer can be associated with level 4 for multiple factories. The process management system 32 can be configured to:

[0112] ■Integrate data from different distributed edge devices (including the intermediate processing system 34 or the monitoring device 44),

[0113] ■Further contextization (bottom-up approach), where additional context is added based on the pre-processed context distributed in the distributed edge devices.

[0114] The external processing layer 30 can be configured as a centralized cloud computing platform. Such a platform can be associated with level 5 for multiple factories. The external processing layer 30 can be configured as a manufacturing data workspace with complete data integration across multiple factories, including manufacturing data historical transfer and streaming, collection of all data from all edge components. In this way, the complete contextization of all lower-level contexts can be integrated onto the external processing layer 30 for multiple factories. Therefore, the external processing layer 30 can be further configured to:

[0115] ■Run cloud-native apps,

[0116] ■Connect with external PaaS and SaaS tenants,

[0117] ■Integrate machine learning with manufacturing data and processes, training-testing-deployment,

[0118] ■Visualize data, access apps, orchestrate.

[0119] Through the system architecture, the bottom-up contextization concept can be realized. Such a concept is as Figure 4As shown. In the bottom-up concept, all information available at a lower level may already be added as attributes to the data, so that the lower-level context is not lost. Here, the first processing layer 14, which is the lowest context level, may include measurement values 11, and the measurement values 11 are contextualized with respect to the item 13 for which the measurement is made. The intermediate processing system 34 may further contextualize by adding further tags 15 related to individual factories 12. The process management system 32 may further contextualize by adding tags 17 related to multiple chemical factories 12 and / or business information. The external processing layer 30 may further contextualize by adding tags 19 related to multiple factories and / or external context information such as from third parties.

[0120] The contextualization concept can cover at least two basic types of context. One type may be the functional location within the production environment of multiple chemical factories. This may cover information about what and where this data point represents in the production environment. Examples are connections to functional locations, attributes regarding which physical asset the data is collected for, etc. This context can be usefully employed in later applications as it explains which data is available for which factories and assets.

[0121] Another type may be the confidentiality classification. Such tags can be added at the lowest possible level, and this information can be propagated to further processing layers. Such tags can be added automatically or manually. Using technical measures, such as filters embedded in a firewall, data that is "strictly confidential" can be automatically prohibited from being integrated into the external processing layer 30 all the time. Sharing data with the outside will result in an automatic notification that "confidential data" is being shared. Automatic contract checks can be implemented to see if the data can be shared with this external processing layer.

[0122] Overall, the contextualization concept implemented in this way allows for the efficient use of data in process applications deployed on any layer of the system.

[0123] Figure 5 A flowchart showing a schematic representation of a method for monitoring and / or controlling one or more chemical factories is shown.

[0124] Preferably, the method is executed on a distributed computing system as Figures 1 to 3 shown, which includes a first processing layer 14 associated with and communicatively coupled to a second processing layer 16, 32, 34 related to a chemical factory 12. The method may perform all steps described in the Figures 1 to 4 context, including any steps related to contextualization, data processing, process application management, and monitoring device management.

[0125] In a first step 61, processes or assets or process-specific data of the chemical plant 12 are provided to second processing layers 16, 32, 34 via a first processing layer 14.

[0126] In a second step 63, the processes or assets or process-specific data are contextualized via the second processing layers 16, 32, 34 to generate plant-specific data.

[0127] In a third step 65, the plant-specific data of one or more chemical plants 12 are provided to an interface 26 of an external network via the second processing layers 16, 32, 34.

[0128] In a fourth step 67, one or more chemical plants are monitored and / or controlled based on the processes or assets or process-specific data or plant-specific data via the second processing layers 16, 32, 34 or the first processing layer 14. The monitoring and / or control of one or more chemical plants 12 can be performed based on the plant-specific data via the second processing layers 16, 32, 34 or an external processing layer 30. Additionally, the monitoring and / or control can be performed based on the processes or assets or process-specific data via the first processing layer 14. Such monitoring and / or control can be executed by a process application that obtains the corresponding data and provides a monitoring and / or control output, as Figures 6 to 8 further listed in.

[0129] Figure 6 shows a schematic representation of a distributed computing system for monitoring and / or controlling one or more chemical plants having multiple assets via a distributed computing system 10 having more than two deployment layers 14, 16, 30.

[0130] Figure 6 The schematic representation of shows containerized application orchestration in different deployment layers 14, 16, 30. The system 10 includes an external processing system 30, a second processing layer 16, and a first processing layer 14. Here, the second processing layer 16 can include greater storage and computing resources than the first processing layer 14, and / or the external processing layer 30 can include greater storage and computing resources than the second processing layer 16. The architecture and functionality of the system 10 can follow the architecture and functionality described regarding Figures 1 to 3 described. In particular, the first processing layer 14 and the second processing layer 16 can be configured in secure networks 20, 40, 18. The first processing layer 14 can be communicatively coupled to the second processing layer 16, while the second processing layer 16 can be communicatively coupled to the external processing layer 30 via an external network 24.

[0131] Orchestration applications 56, 58 may be hosted by the external processing layer 30 and the second processing layers 16, 32, 34, respectively. Accordingly, containerized applications or container images 48, 50 may be stored in the registries of the external processing layer 30 and the second processing layers 16, 32, 34, respectively. The containerized applications 48, 50 for execution may include one or more operations to obtain input data, provide the input data to corresponding assets or one or more plant models to generate output data and provide the generated output data for controlling and / or monitoring the chemical plant 12. In this way, the external processing layer 30 and the second processing layers 16, 32, 34 act as facilitation layers, thereby reducing the computing and storage resources required on the first processing layer 14 at the asset level.

[0132] Figure 7 A flowchart showing a schematic representation of a method for monitoring and / or controlling a chemical plant 12 having multiple assets via a distributed computing system 10, which flowchart may be executed in Figures 1 to 4 the system 10 shown.

[0133] In a first step 60, containerized applications 48, 50 are provided, which containerized applications include asset or plant templates specifying input data, output data, and assets or plant models. The containerized applications 48, 50 may be created on the external processing layer 30 or may be modified on the second processing layer 30. External containerized applications may be provided from a third-party environment.

[0134] In a second step 62, the containerized applications 48, 50 are deployed to execute on at least one of the deployment layers 30, 32, 16, 34, 14, wherein the deployment layers 30, 32, 16, 34, 14 are assigned based on input data, load indicators, or system layer tags, and the containerized applications 48, 50 may be executed on one or more of the assigned deployment layers 30, 32, 16, 34, 14 to generate output data for controlling and / or monitoring the chemical plant 12. The deployment may be managed by the orchestration applications 56, 50, which manage the deployment of the containerized applications 48, 50 based on input data, load indicators, or system layer tags. The orchestration applications may be hosted by the second processing layers 16, 23, 34 and / or the external processing layer 30. The orchestration applications 56, 58 hosted by the second processing layers 16, 32, 34 manage the critical containerized applications 48, 50, wherein the orchestration applications 56, 58 hosted by the external processing layer 30 may manage the non-critical containerized applications 48, 50. The assignment of the deployment layers 30, 32, 34, 16, 14 may be based on input data depending on data availability indicators, criticality indicators, or latency indicators. Containerized applications from a third-party environment may be deployed to execute on the external processing layer 30.

[0135] Orchestration applications 56, 58 can be hosted by the external processing layer 30 and the second processing layer 16 respectively. The orchestration applications 56, 58 can deploy containerized applications 48, 50 on any of the deployment layers 30, 16, 14. Thus, the containerized applications 48, 50 can be executed on the corresponding deployment layers 30, 16, 14 by running the process applications 46, 52, 54 in a sandbox-type environment. The deployment layers 30, 16, 14 can be allocated based on input data, load indicators, or system layer tags. For example, the management of the critical containerized application 50 can optionally be assigned to the second processing layer 16 based on historical criteria that reflect a time window of available historical data on the first or second processing layer 16. Advantageously, the containerized applications 48, 50 can be deployed to multiple assets or factories of the same type. Additionally, the containerized applications 50, 48 can be modified based on the input data and output data provided by the containerized applications 46, 52, 54 executed for multiple assets or factories of the same type.

[0136] In a third step 64, the containerized applications 48, 50 can be monitored based on the confidence level of the input data, asset model, or factory model during or after each execution. Based on the obtained confidence level, an event signal or modification of the asset or factory model can be triggered. If the confidence level exceeds a threshold, such a trigger can be set. Such a threshold can be predefined or dynamic. If a trigger is set, the modification of the asset or factory model can be executed, for example, on the second processing layers 16, 32, 34 or the external processing layer 30.

[0137] In a fourth step 66, the generated output data is provided for controlling and / or monitoring the chemical plant 12. Such output data can be passed to a persistent instance after the execution of the containerized applications 48, 50. In particular, such output data can be passed to a control instance, such as on the first processing layer 14 of the chemical plant 12. Additionally or alternatively, such output data can be passed to a monitoring instance on the first processing layer 14, the second processing layers 16, 32, 34, or the external processing layer 30. The output data can be passed to, for example, a client application for display to an operator or further containerized applications 48, 50 for execution.

[0138] Figure 8 A schematic representation of a system 10.2, 10.2 for monitoring and / or controlling multiple chemical plants 12.1, 12.2 in different security networks 20.1, 20.2 is shown, and these security networks are configured for data and process application transmission. As an example, Figure 8 is shown including, as an example, the first and second processing layers 14, 16, 32, 34 and the external processing layer 30 Figures 1 to 3System 10. Any other system architecture can be similarly applied to process applications and data transfer. Both systems are associated with separate secure networks 20.1, 20.2 and are communicatively coupled to external networks 24.1, 24.2 via interfaces 26.1, 26.2.

[0139] Systems 10.1, 10.2 are configured to exchange process or asset or process-specific data or process applications based on transfer tags. By adding transfer tags at the earliest possible level (i.e., where the data or application is generated or first enters the system), once a tag is added and accompanies the data or application on its path through systems 10.1, 10.2, the transfer tag becomes an inherent part of any data point or application. This transfer tag enables seamless but secure integration of external data sources or external applications, as well as the transfer of data or applications to external resources.

[0140] In Figure 8 In one case shown, application 48 is exchanged between systems 10.1, 10.2. In this example, the containerized application 48 is transferred via external processing layers 30.1, 30.2 communicatively coupled to both systems 10.1, 10.2. Here, external processing layer 30.1 is communicatively coupled to system 10.1, and external processing layer 30.2 is communicatively coupled to system 10.2. The exchange of the containerized application 48 is performed indirectly through external processing layers 30.1, 30.2. The containerized application is tagged with a transfer tag that includes two transfer settings related to confidentiality settings and / or third-party transfer settings. In this way, the transfer can be prohibited based on a compliance check of external processing layer 30.2, for example, if a transfer with a corresponding third-party identifier is not associated with the third-party identifier stored in the database for allowed third-party transfers for process application 48. Similarly, process or asset or process-specific data can be transferred 72 between systems 10.1, 10.2. Then, further transfer from external processing layers 30.1, 30.2 to the corresponding systems 10.1, 10.2 can be performed after any transfer between systems 10.1, 10.2.

[0141] Additionally, this transport based on transport tags can be directly performed between systems 10.1, 10.2 between the processing layers 32, 16 associated with the secure networks 20.1, 20.1. This transport based on transport tags can be implemented via a secure connection 74 (such as a VPN connection) between such layers 16, 23. Thus, any transport between systems 10.1, 10.2 can be accompanied by a further transport between system components within the secure networks 20.1, 20.2 or to the external processing layers 30.1, 30.2 of the respective systems 10.1, 10.2. By attaching transport tags to any data points and process applications, whether containerized or not, it allows for the secure handling of third-party transports between systems 10.1, 10.2 in separate secure networks 20.1, 20.1.

[0142] Any component for implementing the methods described herein can be in the form of a distributed computer system having one or more processing devices capable of executing computer instructions. The components of the computer system can be communicatively coupled (e.g., networked) to other machines in a local area network, secure network, intranet, extranet, or the Internet. The components of the computer system can operate as peer machines in a peer (or distributed) network environment. Parts of the computer system can be a virtualized cloud computing environment, edge gateway, network device, server, network router, switch, or bridge, or any machine capable of executing a set of instructions (sequentially or otherwise) that specify actions to be taken by that machine. Additionally, it should be understood that the terms "computer system", "machine", "electronic circuit", etc. are not necessarily limited to a single component, but should be understood to include any collection of machines that individually or jointly execute a set (or multiple sets) of instructions to perform any one or more of the methods discussed herein.

[0143] Some or all of the components of such a computer system can be used or accounted for by any component of system 10. In some embodiments, one or more of these components can be distributed between multiple devices, or can be consolidated into fewer devices than shown. Additionally, some components can refer to physical components implemented in hardware, while others can refer to virtual components implemented in software on remote hardware.

[0144] Any processing layer may include general-purpose processor devices, such as microprocessors, microcontrollers, central processing units, etc. More specifically, the processing layer may include a CISC (Complex Instruction Set Computing) microprocessor, a RISC (Reduced Instruction Set Computing) microprocessor, a VLIW (Very Long Instruction Word) microprocessor, or a processor implementing other instruction sets or a combination of instruction sets. The processing layer may also include one or more dedicated processor devices, such as ASICs (Application-Specific Integrated Circuits), FPGAs (Field-Programmable Gate Arrays), CPLDs (Complex Programmable Logic Devices), DSPs (Digital Signal Processors), network processors, etc. The methods, systems, and devices described herein may be implemented as software in a DSP, microcontroller, or any other co-processor or as hardware circuits within an ASIC, CPLD, or FPGA. It should be understood that the term "processing layer" may also refer to one or more processing devices, such as a distributed system of processing devices located on multiple computer systems (e.g., cloud computing), and is not limited to a single device, unless otherwise specified.

[0145] Any processing layer may include a suitable data storage device, such as a computer-readable storage medium, on which one or more sets of instructions (e.g., software) embodying any one or more of the methods or functions described herein are stored. The instructions may also reside, wholly or at least partially, in the main memory and / or in the processor during its execution by a computer system, the main memory, and the processing device, which may constitute a computer-readable storage medium. The instructions may also be sent or received over a network via a network interface device.

[0146] A computer program for implementing one or more embodiments described herein may be stored and / or distributed on a suitable medium, such as an optical storage medium or a solid-state medium provided together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems. However, the computer program may also be presented via a network such as the World Wide Web and downloaded into the working memory of a data processor from such a network.

[0147] The terms "computer-readable storage medium", "machine-readable storage medium", etc. should be understood to include a single medium or multiple media (e.g., a centralized or distributed database, and / or associated caches and servers) storing one or more sets of instructions. The terms "computer-readable storage medium", "machine-readable storage medium", etc. should also be understood to include any transient or non-transient medium capable of storing, encoding, or carrying a set of instructions for execution by a machine and causing the machine to perform any one or more of the methods of the present disclosure. Thus, the term "computer-readable storage medium" should be understood to include, but not be limited to, solid-state memory, optical media, and magnetic media.

[0148] Certain portions of the detailed description may have been presented in terms of algorithms and symbolic representations of operations on data bits within a computer memory. These algorithmic descriptions and representations are the means used by those skilled in the data processing arts to most effectively convey the substance of their work to others skilled in the art. An algorithm is here, and generally, conceived as a self-consistent sequence of steps leading to a desired result. These steps are those requiring physical manipulations of physical quantities. Usually, though not necessarily, these quantities take the form of electrical or magnetic signals capable of being stored, transferred, combined, compared, and otherwise manipulated. It has proven convenient at times, principally for reasons of common usage, to refer to these signals as bits, values, elements, symbols, characters, terms, numbers, etc.

[0149] However, it should be borne in mind that all of these and similar terms are to be associated with the appropriate physical quantities and are merely convenient labels applied to these quantities. From the foregoing discussion, it should be apparent that unless specifically stated otherwise, it is to be understood that throughout the description, discussions using terms such as "receiving", "retrieving", "transmitting", "computing", "generating", "adding", "subtracting", "multiplying", "dividing", "selecting", "optimizing", "calibrating", "detecting", "storing", "executing", "analyzing", "determining", "enabling", "identifying", "modifying", "transforming", "applying", "extracting", etc., refer to the actions and processes of a computer system or similar electronic computing device that manipulate and transform data represented as physical (e.g., electronic) quantities within the registers and memories of the computer system into other data similarly represented as physical quantities within the memories or registers of the computer system or other such information storage, transmission, or display devices.

[0150] It must be noted that the embodiments of the present invention are described with reference to different subjects. In particular, some embodiments are described with reference to method-type claims, while other embodiments are described with reference to system-type claims.

[0151] However, those skilled in the art will appreciate from the above and the following description that, unless otherwise stated, any combination of features related to different subjects is also considered to be disclosed with this application in addition to any combination of features belonging to one subject. However, all features can be combined to provide synergistic effects that are more than just the simple sum of the features.

[0152] Although the present invention has been described in detail in the drawings and the foregoing description, such description and illustration shall be considered illustrative or exemplary rather than restrictive; the present invention is not limited to the disclosed embodiments. Through the study of the drawings, the disclosure, and the appended claims, those skilled in the art can understand and implement other variations of the disclosed embodiments and practice the claimed invention. In some cases, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring the disclosure.

[0153] 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. A single processor or controller or other unit may implement the functions of several items recited in the claims. The fact that certain measures are recited in mutually different dependent claims does not mean that a combination of these measures cannot be used advantageously. Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A method for monitoring and / or controlling a chemical plant (12) having a plurality of assets via a distributed computing system (10) having more than two deployment layers (14, 16, 30, 32, 34), wherein, The deployment layer (14, 16, 30, 32, 34) includes at least two of a first processing layer (14), a second processing layer (16, 32, 34), and an external processing layer (30), and the method includes the following steps: - Providing (60) containerized applications (48, 50) that include an asset or factory template specifying input data, output data, and assets or a factory model. - Deploying (62) the containerized applications (48, 50) to execute on at least one of the deployment layer (14, 16, 30, 32, 34), where the deployment layer (14, 16, 30, 32, 34) is assigned based on the input data, a load indicator, or a system layer label, and executing the containerized applications (46, 52, 54) on the assigned deployment layer (14, 16, 30, 32, 34) to generate output data for controlling and / or monitoring the chemical plant (12). - Providing (66) the generated output data for controlling and / or monitoring the chemical plant (12).

2. The method according to claim 1, wherein The second processing layer (16, 32, 34) includes greater storage and computing resources than the first processing layer (14), and / or the external processing layer (30) includes greater storage and computing resources than the second processing layer (16, 32, 34).

3. The method according to claim 1 or 2, wherein The first and second processing layers (14, 16, 32, 34) are configured inside a secure network (20), where the first processing layer (14) is communicatively coupled to the second processing layer (16, 32, 34), and the second processing layer (16, 32, 34) is communicatively coupled to the external processing layer (30) via an external network.

4. The method according to any one of the preceding claims, wherein, The containerized applications (48, 50) for execution include one or more operations to obtain input data, provide the input data to a corresponding asset or one or more factory models to generate output data, and provide the generated output data for controlling and / or monitoring the chemical plant (12).

5. The method according to any one of the preceding claims, wherein, The deployment is managed by an orchestration application (56, 58) that manages the deployment of the containerized applications (48, 50) based on the input data, the load indicator, or the system layer label.

6. The method according to claim 5, wherein, The orchestration application (56, 58) is hosted by the second processing layer (16, 32, 34) and / or the external processing layer (30).

7. The method according to claim 5 or 6, wherein The orchestration application (58) hosted by the second processing layer (16, 32, 34) manages critical containerized applications (48, 50), where the orchestration application (56) hosted by the external processing layer (30) manages non-critical containerized applications (48, 50).

8. The method according to any one of claims 5 to 7, wherein Based on a historical criterion of a time window of available historical data reflected in the first processing layer or the second processing layer (14, 16, 32, 34), the management of critical containerized applications (56, 58) is assigned to the second processing layer (16, 32, 34).

9. The method according to any one of the preceding claims, wherein, The allocation of the deployment layer (14, 16, 30, 32, 34) based on the input data is dependent on a data availability indicator, a criticality indicator, or a latency indicator.

10. The method according to any one of the preceding claims, wherein, The containerized application is deployed to multiple assets or plants of the same type.

11. The method according to any one of the preceding claims, wherein, The containerized applications (48, 50) are modified based on the input data and the output data provided by the containerized applications (48, 50) executed for multiple assets or plants (12) of the same type.

12. The method according to any one of the preceding claims, wherein, The containerized applications (48, 50) are monitored based on the confidence level of the input data, the asset model, or the plant model.

13. The method according to claim 12, wherein If the confidence level is below a confidence threshold, an event signal is triggered or the asset or plant model is modified.

14. The method according to claim 12 or 13, wherein, The modification of the asset or plant model is performed on the second processing layer (15, 32, 34) or the external processing layer (30).

15. The method according to any one of the preceding claims, wherein, External containerized applications from a third-party environment are provided and deployed to execute on the external processing layer (30).

16. A system (10) for monitoring and / or controlling a chemical plant (12) having a plurality of assets with more than two deployment layers (14, 16, 30, 32, 34), wherein, The deployment layer (14, 16, 30, 32, 34) includes at least two of a first processing layer (14), a second processing layer (16, 32, 34), and an external processing layer (30), and the system (10) is configured to: - Provide (60) containerized applications (48, 50) that include an asset or plant template specifying input data, output data, and an asset or plant model, - Deploy (62) the containerized applications (48, 50) to execute on at least one of the deployment layer (14, 16, 30, 32, 34), where the deployment layer (14, 16, 30, 32, 34) is allocated based on the input data, a load indicator, or a system layer tag, and execute the containerized applications (46, 52, 54) on the allocated one or more deployment layers (14, 16, 30, 32, 34) to generate output data for controlling and / or monitoring the chemical plant (12), - Provide (66) the generated output data for controlling and / or monitoring the chemical plant (12).

Images