EP4689807A1

EP4689807A1 - Method for controlling an industrial process

Info

Publication number: EP4689807A1
Application number: EP23716241.7A
Authority: EP
Inventors: Katharina Stark; Jan Christoph SCHLAKE; Thorsten Schindler; Marco Ulrich; Andreas Zehnpfund
Original assignee: ABB Schweiz AG
Current assignee: ABB Schweiz AG
Priority date: 2023-03-29
Filing date: 2023-03-29
Publication date: 2026-02-11
Also published as: WO2024199651A1; CN120898181A

Abstract

A method for controlling an industrial process using a model of the industrial process, the method comprising selecting a production policy from a set of policies, wherein a policy is a set of high-level configurations that are applied in a high-level framework of the industrial process under a respective predefined state, simulating the high-level configurations of the selected production policy, evaluating at least one performance indicator of the selected policy, optionally modifying the selected policy or selecting another production policy based upon the at least one performance indicator, accepting and implementing the accepted policy.

Description

METHOD FOR CONTROLLING AN INDUSTRIAL PROCESS

TECHNICAL FIELD

[0001] Embodiments of the present disclosure relate to a method of controlling an industrial process. Further embodiments relate to an industrial control system for controlling an industrial process. Particularly, methods and control systems according to the embodiments of the present disclosure may relate to pulp and/or paper processing.

BACKGROUND

[0002] In industrial process controlling, energy consumption is getting a more and more important additional driving criterium. This shows the need for holistic and dynamic policies for planning and scheduling of the industrial process. To be able to reduce energy consumption during the industrial process, one needs to know in which process step how much energy is consumed to produce a specific amount of product as basis for combined planning and scheduling. In many industrial processes, particularly in pulp and/or paper making, some sort of energy generation can be found on site. With the energy typically being sourced from multiple sources, on one level considering electricity generation on-site and electricity provided from a public grid, but also considering the sources of electricity in the public grid as well as heat or steam production, particularly using natural gas or hydrogen, a feedback mechanism from the industrial process to the energy management and vice versa is lacking. In particular, there is a desire for a production policy implementation mechanism that incorporates these interactions.

DISCLOSURE OF THE INVENTION

[0003] In the view of the foregoing, the present disclosure is directed to a method and an industrial control system for controlling an industrial process that allows to employ, adapt and test production policies on a high-level framework, that are applied under a predefined state, in particular an environmental or energy-related state. [0004] According to an aspect of the present disclosure, a method of controlling an industrial process is provided. The industrial process is being carried out by a production plant, a distributed control system, a manufacturing execution system, and/or an energy management system. The method employs a model of the industrial process, and comprises multiple steps. The method includes selecting a production policy from a set of policies, simulating the high-level configurations of the selected production policy, evaluating at least one performance indicator of the selected production policy, optionally, modifying the selected production policy or selecting another production policy based upon the at least one performance indicator, accepting the selected production policy based upon the at least one performance indicator and implementing the accepted production policy and controlling the industrial process by the associated low-level operational instructions of the production policy.

[0005] According to another aspect of the present disclosure, an industrial control system for controlling an industrial process is provided. The system comprises a distributed control system module configured for implementing the accepted production policy and the associated low-level operation instructions, a manufacturing execution system module configured for providing an overview over a current production status and order fulfilment, an energy management system module configured for controlling energy flows in the industrial process and an analytics module configured for modeling the industrial process, wherein energy flows influence material flows and material flows influence energy flows.

MODEL OF INDUSTRAL PROCESS

[0006] According to some embodiments of the present disclosure, the model of the industrial process may be built upon a digital twin model of the industrial process, herein also referred to as “model”. In particular, the digital twin model may comprise interactions and dependencies of different steps of the industrial process. In embodiments, the digital twin model or a part of the digital twin model may be generated based upon historical data. In particular, historical data may allow to find implicit interactions between different steps of the industrial process. In embodiments, the digital twin may be generated by a machine learning method using historical data. The digital twin model may consist of production- related models which allow to build a model for the entire industrial process, in particular tracking energy flows and tracking production flows. Production flows may also be named material flows. The general concept and possible implementations of a digital twin model are described, for example, in WO2022144082A1. A digital twin model may be named “Material Flow and Energy Digital Twin” if it incorporates material flows and energy flows. The digital twin model may be generated based on process descriptions, particularly on a piping and instrumentation diagram (P&ID), and may be optimized using historical data or live data.

[0007] The model of the industrial process may comprise asset models of assets of the industrial process, comprising models of single units and/or groups of machinery involved in the industrial process. Further, the model of the industrial process may comprise process structures of the industrial process.

[0008] In some embodiments, the model of the industrial process, in particular a digital twin model, may be initialized from current process values. Current process values may be obtained from a distributed control system (DCS) that controls the industrial process on a low level or from an edge device collecting and providing process-related signals. In particular, the DCS and/or the edge device may comprise sensors that measure process values of the industrial process. In some embodiments, the model of the industrial process may be initialized from the production schedule, in particular provided by a planning and scheduling module. More particularly, the model of the industrial process may be initialized from planned production states. In some embodiments, the model of the industrial process may be initialized from storage levels, in particular provided by an inventory management. Storage levels may comprise at least one of product storage levels, energy storage levels, raw or intermediate material storage levels, auxiliary materials storage levels, operating materials storage levels. In some embodiments, the model of the industrial process may be initialized from the production status, in particular provided by a manufacturing execution system (MES). In some embodiments, the model of the industrial process may be initialized from historical data, in particular historical sensor data, historical storage level data, historical production status data and/or historical cost data. In embodiments, the model of the industrial process may be initialized by a combination of the aforementioned initialization paths. [0009] According to embodiments, the model of the industrial process may comprise a simulator. The simulator may use the production-related models to provide predictions of future process values, storage levels, and production status. The simulator may comprise production constraints and/or resource constraints. Production policies may be implemented in the simulator. The simulator may be used to optimize production policies.

[0010] In embodiments, the digital twin model, in particular the Material and Energy Flow Digital Twin (MEFDT), may model the effect of changes in the material flow on the energy flows. Further it may model the effect of changes in the process parameters and/or a process set-up on the energy flows. The MEFDT may model the energy flows based on the material flow and the steps of the industrial process necessary to achieve a particular material flow. In particular, the MEFDT models the energy flows based on the precise low-level operation instructions associated with a particular production policy. In particular, differences in the low-level operation instructions of different production policies may be reflected in the energy flows of the MEFDT.

[0011] According to some embodiments, the digital twin model, in particular the Material and Energy Flow Digital Twin (MEFDT), may model the effect of energy buffers. Energy buffers may be explicit by energy and/or heat storage machinery, particularly electrical batteries. Energy buffers may be implicit by implicit storages, particularly machinery wherein a temperature can be set within a range.

INDUSTRIAL PROCESS/PREAMBLE CLAIM 1

[0012] In embodiments, the model of the industrial process, in particular the digital twin model, more in particular the MEFDT, may be incorporated in an analytics module of the industrial process. The analytics module may further comprise a value chain model to generate performance indicators from predicted material and energy flows simulated in the MEFDT. In particular, performance indicators may be generated based on a selected production policy, of which the low-level operational instructions are employed to simulate the industrial process in the MEFDT.

[0013] According to some embodiments, the analytics module may further comprise a value chain optimization module. The value chain optimization module may amend production policies to optimize performance indicators obtained by the value chain model. An amended production policy may be simulated in the model of the industrial process, in particular in the digital twin model, more in particular in the MEFDT.

[0014] In embodiments, the analytics module may further comprise aggregated data from historical data and/or current data from the DCS.

[0015] The industrial process is being carried out by the production plant, the distributed control system, the manufacturing execution system and the energy management system. The production plant may comprise machinery required in production steps, transport systems to move material from one machine to another machine, buildings housing the machinery and/or energy supply infrastructure.

[0016] In embodiments, the distributed control system is used by an operator to control the plant. In some embodiments, the production plant may further comprise edge devices installed at a production site. The edge devices may collect and exchange information and data with the DCS. The DCS may be configured to allow the operator to control the operation of the plant. Further, the DCS may use sensor devices to obtain information on low-level process values. The DCS may implement set-points for specific process values. The DCS may control controllers of machinery of the industrial process to achieve these set points. Controllers of machinery may use control laws, in particular proportional control laws, proportional integral control laws or proportional-integral-derivative (PID) control laws to run the machinery. The DCS may provide alarm signals to an operator if a specific controller and/or sensor measures a process value beyond a predefined range. The DCS may implement operator decisions, in particular high-level configuration decisions, into lower-level operation instructions. The operator operating the production plant by the DCS may be a person or another module of the industrial process.

[0017] In embodiments, the manufacturing execution system (MES) may provide a production plant manager and/or the production planner an overview of the current production status and a status of order fulfilment. In particular, an MES may control the plant operations on a high-level framework. A plant manager and/or the production planner may use the MES to optimize the execution and/or planning of the industrial process. In particular, the MES may comprise and use information from a maintenance plan and a production plan. The MES may use information from an inventory management system. The MES may exchange information with an enterprise resource planning module (ERP).

[0018] In embodiments, the energy management system (EMS) may manage energy flows in the industrial process. Energy flows may comprise electricity flows, steam flows, heat flows and/or coolant flows. The EMS may receive information on energy flow parameter. Energy flow parameter may comprise availability of an energy flow, current price of energy flow, future price of energy flow and/or source of energy flow. The source of energy flow may comprise a provider of an energy flow and/or the production method of an energy flow and the amount of CO2 equivalent needed to produce it. The EMS may exchange information with the MES and/or the DCS and/or a planning and scheduling system.

[0019] The analytics module may exchange information with the MES and/or the DCS and/or the planning and scheduling system. In particular, the MES may exchange information with the analytics module to test and/or verify and/or optimize high-level configurations on a lower level. The DCS may exchange information with the analytics module, in particular with the digital twin model, to initiate the digital twin model with current data and to select and/or optimize a suitable production policy on a higher level. The energy management system may exchange information with the analytics module to test the influence of changes in the energy flows on the production process or to test the influence of changes in the production process, in particular of changes in production polices, on the energy flows.

PRODUCTION POLICY

[0020] In embodiments, a production policy is a set of high-level configurations that are applied in a high-level framework of the industrial process under a respective predefined state. In particular, the production policy may be described as follows: if a specific predefined state is fulfilled (i.e., an actual state corresponds to a predefined state), a specific action is taken. Further a production policy may be aimed at optimizing at least one specific performance indicator. For a predefined state, multiple production policies may be available, in particular multiple production policies optimized for different performance indicators. [0021] In embodiments, the high-level framework of the industrial process may be a generalized and systematic description of the industrial process and interactions within the industrial process. In particular, generalized flows of material and energy may be described in a high-level framework. Further, interactions between multiple steps of the industrial process may be described in a high-level framework. Within the high-level framework, a production policy may be structured such that the predefined state and the specific action taken is easily conceivable by an experienced operator.

[0022] According to some embodiments, the high-level configurations may comprise a description, in the high-level framework, of a generalized operation mode of at least the distributed control system, the production process and the energy management system. In particular, high-level configurations may be focused on an operational goal to be achieved in the production process. A high-level configuration is thus an aggregated and abstract description of the industrial process and of the operation of the production plant which is understandable to a human operator, at an aggregation and abstraction level above the DCS. For example, the high-level configuration can be a production plan. Low-level operation instructions may be derived from the high-level configurations by the DCS and/or by an experienced operator.

[0023] In embodiments, a predefined state of the production policy may comprise at least one of an environmental state, an energy supply state, a process state, a stock level state, an energy source state, a market state. A predefined state may be a condition of an internal or external variable that can be determined reliably. A predefined state can be a historical state, a present state or a future state. A predefined state may be a discrete property or a continuous property. A predefined state may be composed of multiple conditions of internal or external variables. The condition of an internal or external variable may be expressed in a high-level framework.

[0024] In embodiments, an environmental state may comprise a weather state, particularly a temperature, a wind speed and/or a cloudiness level, an astrological state, particularly a sunrise and/or a sunset time, and/or a nature state, particularly a water level and/or a water quality of a body of water. An energy supply state may comprise a price of a unit of energy, in particular the price of a unit of electricity and/or a unit of heat, an availability of energy, in particular the availability of electricity, heat or steam. A process state may be every reliably determinable state of the industrial process, comprising conditions of specific machinery, load levels, operator interactions, malfunctions of parts of the industrial process or machinery and/or reliably determinable properties of a product of the industrial process. A stock level state may comprise a stock level of a product and/or a waste product of the industrial process and/or a stock level of a raw material and/or a material required for the operation of the industrial process and/or a stock level of chemicals and/or available storage space to store a product. A stock level state may comprise a stock level of a source of energy, in particular a stock level of oil, oil products, liquified or pressurized gas, hydrogen, coal, and/or wooden pellets, pulp or chips. An energy source state may comprise the source of electricity, in particular the amount of electricity produced on-site, the share of renewable electricity generation in a unit of electricity, the share of non-fossil electricity generation in a unit of electricity and/or the amount of CO2-equivalent emission of a unit of electricity. A market state may comprise a price of a unit of raw material, particularly of a unit of pulp or old paper, an availability of raw materials, a price of chemicals, an availability of chemicals, a price of unit of product, in particular the price of unit of product at a specific point in time and/or the demand of product.

[0025] According to some embodiments, a predefined state may comprise multiple of the states mentioned before. A predefined state may be calculated based upon indicators of multiple of the states mentioned before.

SELECTING POLICY

[0026] In embodiments, selecting a production policy may comprise selecting a production policy in live conditions, in particular based on an operational requirement. An operational requirement may comprise a malfunction of machinery, a change in predefined state and/or a change in a production plan. In some embodiments, a production policy may be selected for a future state, in particular to plan a production. In some embodiments, a production policy may be selected for a historical state, in particular to compare different production policies, optimize production policies and/or to conduct fault analysis. The set of policies may comprise at least one production policy. In embodiments, the set of policies may only comprise production policies with the same predefined state. [0027] In embodiments, the set of policies may be restricted to production policies in which a predefined state resembles a present state. The present state may be derived using the DCS, the EMS, and/or other input paths. These input paths may include the Internet, e.g., for obtaining environmental information. In particular, the predefined state may resemble the present state if the conditions of internal and/or external variables of a state are similar. In embodiments, multiple states of the industrial process may be considered similar if, under the same operating conditions, the difference in the variables describing the states does not increase overall.

[0028] In embodiments, the production policies from the set of policies can be derived by a policy generator. Particularly, an initial production policy, that may be optimized, can be derived by a policy generator. The policy generator combines a predefined state and at least one high-level operating instruction. The policy generator may derive production policies from historical data. The policy generator may further derive production policies from a set of historical data using machine learning methods, in particular finding implicit operating procedures and/or interdependencies in the industrial process. Historical data may comprise at least one of cost, in particular energy cost, energy consumption, machine usage, energy source and/or storage level. A production policy may be derived from an operational experience of an operator. A production policy may be derived from explicit standard operating procedures. A production policy may be derived from an educated guess on how different aspects of the production process are connected and/or interacting.

[0029] In embodiments, the policy generator may combine a user-defined state with at least one high-level configuration from historical data. In embodiments, a policy generator may derive a production policy for a present state by comparing historical data in which the present state has occurred with other present states. In particular, a production policy in which a historical state resembles the present state may be used to derive a production policy for the present state.

SIMULATING POLICY

[0030] In embodiments, the high-level configurations of the selected production policy are translated into low-level operation instructions by the model of the industrial process. Low- level operation instructions may comprise instructions that specifically set machinery settings and/or set-points of controller devices. Low-level operation instructions may be implemented using the DCS. Suitable low-level instructions may be translated from the high- level configurations by the model of the industrial process employing information on the process structure and material and energy flows. In embodiments, historical data may be used to implement high-level configurations into low-level operation instructions, in particular using machine learning methods. Additionally or alternatively, an optimization algorithm may be used to translate the high-level configurations into the low-level operation instructions. In particular, a reinforcement learning method, as exemplarily described in DE102021004426A1, may be employed to obtain low-level operation instructions. After implementing the high-level configurations into low-level operation instructions, the industrial process may be simulated (“simulating the high-level configuration”). In the digital twin model, material and energy flows may be derived. A value chain model may be used to derive at least one performance indicator. A performance indicator may also be named key performance indicator (KPI).

[0031] In embodiments, a performance indicator may comprise at least one of a cost indicator, an energy use indicator, a production yield indicator, a product quality indicator, an efficiency indicator, a sustainability indicator, a greenhouse gas indicator, an environmental impact indicator, a wear and tear indicator, a safety indicator. A performance indicator may be an indicator of a single property or may be generated from multiple properties. In particular, the performance indicator may be calculated from multiple performance indicating properties. A performance indicator may be a value of the digital twin model. In embodiments, a performance indicator may be calculated from at least one value of the digital twin model.

[0032] In embodiments, a cost indicator may comprise an overall cost associated to a unit of a product of the industrial process, marginal cost of an additional unit of a product of the industrial process and/or energy cost. A cost indicator may further comprise cost for waste paper, in particular for paper that does not meet quality standards. An energy use indicator may comprise the amount of energy used for a unit of a product of the industrial process, the efficiency of the energy used, waste heat and/or the amount of energy used from different sources of energy, particularly the amount or ratio of energy used from on-site energy sources. A production yield indicator may comprise a manufacturing defect indicator, an indicator of the utilization rate of the overall production plant and/or of specific machinery. An efficiency indicator may comprise the energy required per unit of product, the amount of raw material required per unit of product and/or the amount of time required per unit of product. A sustainability indicator may comprise the amount of renewable energy in the energy mix used. An environmental impact indicator may comprise the amount of waste products per unit of product and/or the composition of waste products per unit of product, in particular the amount of waste water per unit of product. An environmental impact indicator may further comprise the use of waste paper, in particular the amount of recycled waste paper. A greenhouse gas indicator may comprise the amount of CO2-equivalent emission per unit of product. A greenhouse gas indicator may further comprise the amount of CO2- equivalent emission of waste paper associated with the production of a unit of product. A wear and tear indicator may comprise the amount of wear and tear per unit of product, the expected risk of a malfunction due to wear and tear, the expected lifetime of the production plant and/or specific machinery due to wear and tear and/or the amount of repair time required per unit of product due to wear and tear. A safety indicator may comprise the risk of exceeding safe operation conditions.

EVALUATING POLICY

[0033] In embodiments, evaluating a production policy comprises comparing the at least one performance indicator of the selected production policy with an acceptance criterion. In particular, an acceptance criterion may be a predefined threshold for at least one performance indicator and/or a predefined threshold for a difference in performance indicator, more in particular if evaluating comprises a comparison with another production policy. In embodiments, evaluating the at least one performance indicator of the selected production policy may comprise presenting the selected production policy and the at least one performance indicator of the selected production policy. Presenting the selected production policy and the at least one performance indicator of the selected production policy may comprise visualizing the selected production policy and the at least one performance indicator of the selected production policy for an operator. In embodiments, evaluating the at least one performance indicator of the selected production policy may comprise presenting the production policy with its high-level configurations next to the at least one performance indicator. In embodiments, evaluating the at least one performance indicator of the selected policy may be performed by an operator. The at least one performance indicator may be selected together with the selected production policy before simulating the selected production policy. One or more performance indicators may be generated and evaluated for each selected production policy. The selected production policy may be evaluated isolated. In embodiments, the at least one performance indicator of the selected production policy is presented in conjunction with at least one other production policy and the associated at least one performance indicator. Multiple production policies may be presented in a way that a preferable production policy is marked according to the performance indicator. Based on the performance indicator generated and evaluated, amending the production policy may be suggested. An acceptance criterion, upon which a selected production policy is to be accepted, for the at least one performance indicator may be defined. An acceptance criterion may be represented by a threshold value of at least one performance indicator. An acceptance criterion may comprise a plurality of thresholds for different performance indicators. In embodiments, evaluating the selected production policy may not require user interaction. In embodiments, evaluating the at least one performance indicator of the selected policy may be automated. In embodiments, evaluating the at least one performance indicator of the selected production policy may comprise saving the selected production policy and the respective at least one performance indicator in a computer file, particularly in a temporary computer file and/or a log file.

MODIFYING POLICY

[0034] In some embodiments, after evaluating, the evaluated production policy may be amended. The amendments may be focused on at least one performance indicator selected upon evaluation of the selected production policy and/or evaluation of the at least one respective performance indicator. In embodiments, the optimization of the evaluated production policy may be conducted by a value chain optimization. The value chain optimization may use optimization algorithms, particularly Gradient Decent, to optimize the evaluated production policy. In embodiments, the predefined state of the production policy may be amended. In particular, the condition of a variable of the predefined state may be amended. More in particular, the value upon which a predefined state is considered fulfilled may be amended. In embodiments, the high-level configurations that are applied under a predefined state may be amended. [0035] In embodiments, amending the evaluated production policy may be restricted to the high-level configurations of the evaluated production policy. A restriction of amendments to the high-level configurations may reduce the complexity of optimization and provide a more foreseeable optimization result for an operator and/or production planer and/or process planer. Further, an operator and/or production planer and/or process planer may select a production policy based on an already materialized predefined state and may only seek to optimize the production policy selection and the production policy to be implemented in reaction to the materialized state. In embodiments, a production policy may be optimized automatically if a predefined state occurs.

[0036] In embodiments, after an amendment to the production policy has been performed, the amended production policy may be simulated. From the simulation of the amended production policy, performance indicators may be generated. The generated performance indicators of the amended production policy may be the same performance indicators as in the initial simulation. In embodiments, new performance indicators may be selected before modifying the selected production policy.

[0037] The amended production policy and the at least one performance indicator of the amended production policy may be compared with the initial production policy and the at least one performance indicator of the initial production policy. In embodiments, the at least one performance indicator of the amended production policy may be compared with the at least one performance indicator of the acceptance criterion. If the predefined threshold defined in the acceptance criterion is reached, the amended may be accepted. If the predefined threshold is not reached, the acceptance criterion may be evaluated iteratively until the threshold is reached.

[0038] In embodiments, the amended production policy and the at least one performance indicator of the amended performance indicator may be compared with another production policy of the set of policies. If the other production policy is better than the amended production policy, the other production policy may be employed and accepted or further adapted. A production policy may be considered better than another production policy if the performance indicator of the production policy is closer to the threshold value than the performance indicator of the other production policy. [0039] In embodiments, a production policy may be compared with at least one other production policy in a what-if-analysis. A what-if-analysis to compare production policies may comprise a scenario that is reflected in a state and simulating a production policy in the model of the industrial process initialized with process values associated with the scenario. At least one performance indicator may be generated for each production policy to be compared. In embodiments, multiple scenarios may be employed to compare production policies. At least one performance indicator may be generated from the at least one performance indicator of multiple scenarios. A production policy may be accepted based upon a comparison in a what-if-analysis.

[0040] In embodiments, a what-if analysis may comprise a Queuing Chain Model, in particular for foreseeable breaks, in particular for maintenance breaks.

[0041] In embodiments, a production policy may be tested in a what-if-analysis to test the robustness of a production policy. Testing the robustness of a production policy may comprise setting scenarios that are outside standard operating scenarios. In embodiments, testing the robustness of a production policy may comprise unintended production disruptions, particularly sheet brakes. In embodiments, testing the robustness of a production policy may comprise unplanned maintenance breaks. In embodiments, testing the robustness of a production policy may comprise unexpected changes in product demand, particularly additional product demand or less product demand, unexpected changes in energy prices, particularly higher peak energy prices or negative energy prices. In embodiments, an indicator of the robustness of a production policy may be uptime or downtime of machinery and/or the production process, a mean time between unplanned production stops and/or failures.

ACCEPTING POLICY

[0042] In embodiments, a selected production policy may be accepted based upon the at least one performance indicator. In particular, a selected production policy may be accepted if a threshold of the at least one performance indicator is passed. An evaluated production policy may be accepted if a selected production policy may be considered better than another production policy, in particular than a production policy in use or designated to be used. In embodiments, a production policy may be considered better than another production policy if its at least one performance indicator is better than the at least one performance indicator of the other production policy. In embodiments, accepting a selected production policy may be performed manually by a user. In embodiments, a selected production policy is accepted automatically once a predefined threshold is reached.

IMPLEMENTING POLICY

[0043] In embodiments, an accepted production policy may be implemented. Implementing an accepted production policy may comprise implementing the associated low-level instructions to control the industrial process by the DCS. In particular, implementing may comprise changing controller settings of controllers of machinery, amending set points of controllers of machinery, amending ranges for alarm signals and/or amending instructions for operators.

[0044] In embodiments, implementing an accepted production policy may comprise amendments in the energy management system. In particular, low-level instructions for energy procurement may be amended. In embodiments, energy storage low-level instructions may be amended.

[0045] In embodiments, implementing an accepted production policy may comprise amendments in the inventory management system. In particular, low-level instructions for replenishing inventory may be amended.

[0046] In embodiments, implementing an accepted production policy may comprise amendments in the planning and scheduling system. In particular, the execution order and/or a time of a process step may be amended.

[0047] In embodiments, implementing an accepted production policy may comprise amendments in the manufacturing execution system. In particular, high-level configurations to be employed after unplanned states may be amended and/or set.

MISCELLANEOUS

[0048] In embodiments, the industrial process controlled by the method may comprise a pulp production. In embodiments, the industrial process controlled by the method may comprise a paper production. In embodiments, the industrial process controlled by the method may comprise mining processes, minerals processing, food and beverage production, hydrogen generation and/or steam generation. In embodiments, the industrial process controlled by the method may be a chemical industry process. In embodiments, the method may be used to control water networks.

[0049] In embodiments, the method may be employed to amend production policies during a production process. In particular, a used production policy may be selected and simulated in the model of the industrial process, wherein the model of the industrial process is initialized with a current production state. In embodiments, the amendment process during production is implemented continuously during production. In embodiments, the production policy is amended with the method in regular intervals, in particular every hour, every two hours, every four hours, every eight hours or every day. In embodiments, the production policy is amended with the method if at least one predefined state changes. In embodiments, the production policy is amended in regular intervals or if at least one predefined state changes. In embodiments, the production policy is amended with the method if the digital twin model of the industrial process is amended, particularly if the industrial process is amended.

INDUSTRIAL CONTROL SYSTEM

[0050] The method according to the embodiments described herein, particularly according to the methods described herein, may be performed in an industrial control system. The industrial control system may comprise a distributed control system module, a manufacturing control system module, an energy management system module and an analytics module.

[0051] In embodiments, the distributed control system module may comprise a plurality of sensor systems, a plurality of controller systems and/or edge devices. In particular, the distributed control system may be distributed over the entire production plant. In embodiments, low-level operational instructions of a production policy may be implemented in edge devices and/or controllers. In embodiments, low-level operational instructions of a production policy may be processed centrally. In particular, sensor data may be transmitted to a central control infrastructure and controller settings are amended due to centrally performed analysis.

[0052] In embodiments, the analytics module of the industrial control system may comprise a digital twin model of the industrial process, in particular a Material Flow and Energy Digital Twin. In embodiments, the digital twin may be accessed via cloud computing infrastructure.

[0053] Embodiments of the present disclosure may allow to control an industrial process by amending production policies in a digital twin model. In particular, production policies may be amended to incorporate energy supply related external states. Embodiments may allow to improve the sustainability of a unit of product, in particular the CO2-equivalent emission per unit of product. Embodiments may allow to control the industrial process in a way to optimize the production and/or production policies along flexible spot electricity prices.

BRIEF DESRICPTION OF THE DRAWINGS

[0054] The accompanying drawings relate to embodiments of the disclosure and are described in the following:

FIG 1 schematically illustrates a method for controlling an industrial process according to embodiments described herein;

FIG 2 schematically describes part of a method, in particular amending a selected policy, for controlling an industrial process according to embodiments described herein;

FIG 3 schematically describes part of a method, in particular amending a selected policy, for controlling an industrial process according to embodiments described herein;

FIG 4 schematically describes a model of the industrial process according to embodiments described herein;

FIG 5 schematically describes the interactions of a policy generator according to embodiments described herein;

FIG 6 schematically describes interactions of a material flow and energy digital twin according to embodiments described herein;

FIG 7 schematically describes interactions of a material flow and energy digital twin according to embodiments described herein;

FIG 8 schematically describes interactions of a material flow and energy digital twin according to embodiments described herein;

FIG 9 schematically illustrates an industrial control system according to embodiments described herein; and

FIG 10 schematically illustrates an implementation of production policies according to embodiments described herein. DETAILED DESCRIPTION OF EMBODIMENTS

[0055] Reference will now be made in detail to the various embodiments of the disclosure, one or more examples of which are illustrated in the figures. Generally, only the differences with respect to individual embodiments are described. Each example is provided by way of explanation of the disclosure and is not meant as a limitation of the disclosure. Further, features illustrated or described as part of one embodiment can be used on or in conjunction with other embodiments to yield a further embodiment. It is intended that the description includes such modifications and variations.

[0056] Fig 1 schematically illustrates a method for controlling an industrial process using a model of the industrial process. The method comprises multiple steps. In a first step a production policy is selected from a set of policies. The set of policies comprises at least one policy. In Fig. 1, the set of policies comprises four production policies (ppi, pp2, pp3, pp4). Exemplarily, pp2 is selected. The selected production policy is simulated in the model of the industrial process. Within the model of the industrial process, high-level configurations of the selected production policy are translated into low-level operational instructions. Further, performance indicators, also named key performance indicators (KPI), are generated within the model. At least one of the KPI of the selected production policy is evaluated. Based on the result of the evaluation of the at least one of the KPI of the selected production policy, the selected production policy is either accepted, adapted, or another production policy is selected. Once a production policy is accepted, the production policy is implemented.

[0057] Fig. 2 schematically shows the adaption of a selected production policy in more detail. Adapting a policy comprises amending a selected policy. The amended policy is simulated in the model of the industrial process. Within the model of the industrial process, high-level configurations of the amended production policy are translated into low-level operational instructions. Further, performance indicators are generated within the model. For each performance indicator, the value of the performance indicator is compared with a predefined threshold. If the threshold of the performance indicators is reached, the amended policy is accepted. If the threshold of the performance indicators is not reached, the amended policy is amended further, iteratively repeating the described process until the threshold is reached. Amending the policy may be performed manually by an operator and/or production planer. Amending the policy may be performed automatically by an optimization algorithm.

[0058] Fig. 3 schematically shows an embodiment of the adaption of a selected production policy. The selected policy is amended and simulated as in the embodiment of Fig. 2. The performance indicators of the amended policy are compared to the performance indicators of the originally selected policy. If from a comparison of the performance indicators, the amended policy is considered, the amended policy is accepted. Otherwise the amended policy is amended further, iteratively repeating the described process until the amended policy is better than the selected policy. The embodiment of Fig. 3 is particularly suited for an optimization of an already implemented production policy.

[0059] Fig. 4 schematically illustrates interactions within the model of the industrial process. The model of the industrial process, in particular as a digital twin model, comprises energy flows and material flows. Within the model, influences of the energy flows on the material flows, in particular influences of changes of the energy flows on the material flows, are described. Further, within the model, influences of the material flows on the energy flows, in particular influences of changes in the material flows on the energy flows, are described.

[0060] Fig. 5 schematically shows interactions of a policy generator according to embodiments herein. A state of an industrial process and planned actions within the industrial process are input properties of a simulator module. Within the simulator module, high-level configurations of a production policy are generated. An input state may comprise a current state, a historical state or a future state. Planned actions may be actions that have been performed in the past as a reaction upon a historical state, and/or that could have been performed in the past as a reaction upon a historical state. Planned actions may further be actions according to best practice rules, standard operating procedure, operator experience and/or educated guess.

[0061] Fig. 6 schematically illustrates interactions of a material flow and energy digital twin according to embodiments herein. Information from a distributed control system (DCS) is used to initiate the material flow and energy digital twin with a current state of the production plant. A selected policy is implemented in the material flow and energy digital twin. Within the material flow and energy digital twin, the production process is simulated and at least one performance indicator is generated. An operator may use the performance indicator to decide whether to use the selected policy in the industrial process.

[0062] Fig. 7 schematically illustrates interactions of a material flow and energy digital twin according to embodiments herein. Information from a manufacturing execution system (MES) and a selected production policy are implemented in the material flow and energy digital twin. Within the material flow and energy digital twin, the production process is simulated and at least one performance indicator is generated. An operator may use the performance indicator to amend the MES planning if the selected policy is to be maintained. Alternatively or additionally, the selected policy may be amended.

[0063] Fig. 8 schematically illustrates interactions of a material flow and energy digital twin according to embodiments herein. Planned actions from an integrated planning and scheduling (IPS) system and a selected production policy are implemented in the material flow and energy digital twin. Within the material flow and energy digital twin, the production process is simulated and at least one performance indicator is generated. An operator may use the performance indicator to amend the planned action planned in the IPS if the selected policy is to be maintained. Alternatively or additionally, the selected policy may be amended.

[0064] Fig. 9 schematically illustrates an industrial control system and its interactions with an industrial process. The industrial control system comprises an analytics module, an energy management system (EMS), a manufacturing execution system (MES) and a distributed control system (DCS). The distributed control system controls the industrial process and receives data from sensors and controllers of the industrial process. The analytics module may comprise a digital twin of the industrial process. The digital twin may be initialized using data from the distributed control data. The analytics module may receive data from the EMS to simulate energy states and/or to reflect an energy market state. Analysis data from the digital twin may be used by the energy management system to amend energy procurement. The MES may exchange information with the energy management system on a high-level framework. In particular, information comprising the general availability conditions of energy may be exchanged. The MES provides the analytics module with high-level configurations to be translated in low-level operating instructions. Data from the analytics module, in particular performance indicators, may be employed to adapt and/or optimize high-level framework instructions and/or high-level configurations.

[0065] Fig. 10 schematically illustrates an implementation of production policies according to embodiments described herein. In particular, the industrial process illustrated in Fig. 10 is a paper making process. An optimization of production policies acts upon material flows in the paper machine as well as on energy flows, particularly on energy flows and energy management in a steam generation machine. The interaction of energy flows and material flows is optimized for the steam drying phase in Fig. 10. For other phases of the industrial process, a similar optimization may be performed.

Claims

1. A method for controlling an industrial process, the industrial process being carried out by a production plant, a distributed control system, a manufacturing execution system, and an energy management system, using a model of the industrial process, the method comprising:

(a) selecting a production policy from a set of policies, wherein a policy is a set of high-level configurations that are applied in a high-level framework of the industrial process under a respective predefined state, and wherein the high-level framework is a generalized and systematic description of the industrial process and interactions within the industrial process, the high-level configurations comprise a description, in the high-level framework, of a generalized operation mode of at least the distributed control system, the production process and the energy management system, and a state comprises at least one of an environmental state, an energy supply state, a process state, a stock level state, an energy source state;

(b) simulating the high-level configurations of the selected production policy, using the model of the industrial process, wherein simulating comprises translating the high-level configuration of the selected production policy into low-level operation instructions implemented in the industrial model, thereby generating at least one performance indicator;

(c) evaluating the at least one performance indicator of the selected production policy.

(d) optionally, modifying the selected production policy or selecting another production policy based upon the at least one performance indicator;

(e) accepting the selected production policy based upon the at least one performance indicator; and

(f) implementing the accepted production policy and the associated low-level operation instructions to control the industrial process by the distributed control system.

2. The method according to claim 1, wherein the at least one performance indicator comprises at least one of a cost indicator, an energy use indicator, a production yield indicator, a quality indicator, an efficiency indicator, an environmental impact indicator, particularly a greenhouse gas emission indicator.

3. The method according to claim 1 , wherein modifying the evaluated production policy comprises iteratively evaluating an acceptance criterion for the at least one performance indicator, until the predefined threshold of the at least one performance indicator is reached;

4. The method of claim 3, wherein evaluating an acceptance criterion for the at least one performance indicator comprises:

- amending the production policy,

- simulating the high-level configurations of the amended production policy, using the model of the industrial process, wherein simulating comprises translating the high-level configuration of the amended production policy into low-level operation instructions implemented in the industrial model, thereby generating at least one performance indicator;

- comparing the generated at least one performance indicator with a predefined threshold of the at least one performance indicator

5. The method according to claim 4, wherein amending the evaluated production policy comprises amending the predefined state.

6. The method according to claim 4, wherein the scope of amending the evaluated production policy is restricted to the high-level configurations.

7. The method according to claim 1, wherein

- the model of the industrial process includes energy and material flows; - influences of changes in the energy flows onto the material flows and influences of changes in the material flows onto the energy flows are described in the model of the industrial process.

8. The method according to claim 1, wherein at least one of the production policies from the set of policies is derived by a policy generator.

9. The method of claim 8, particularly by a machine learning process, wherein the policy generator generates policies employing a machine learning process based upon historical data.

10. The method according to claim 1, wherein selecting another production policy based upon the at least one performance indicator comprises a comparison with at least one other production policy.

11. The method according to claim 10, wherein the comparison with at least one other production policy comprises comparing different production policies in a what-if-analysis.

12. The method according to claim 1, further comprising initializing the model of the industrial process with a production state, wherein the production state is derived from at least one of historical data, current data or planned production states.

13. The method according to claim 1, wherein the industrial process comprises paper and/or pulp production.

14. An industrial control system for controlling an industrial process by the method of any of the preceding claims.

15. An industrial control system according to claim 14, comprising a distributed control system module configured for implementing the accepted production policy and the associated low-level operation instructions; a manufacturing execution system module configured for providing an overview over the current production status and order fulfilment; an energy management system module configured for controlling energy flows in the industrial process; and an analytics module configured for modeling the industrial process, wherein energy flows influence material flows and material flows influence energy flows.