Method, apparatus and system for controlling and / or monitoring a configuration and / or operation of a chemical plant

A trained data-driven model addresses the challenge of finding suitable starting points for chemical plant simulations, enhancing efficiency and convergence by interpolating profiles, thus optimizing chemical plant operations.

WO2026233905A1PCT designated stage Publication Date: 2026-11-12BASF SE
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/EP2026/062755
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-05
Filing Date
2026-05-04
Publication Date
2026-11-12

AI Technical Summary

Technical Problem

Existing methods for simulating chemical plant operations face challenges in finding suitable starting points for equation-oriented models, leading to inefficient runtime and convergence issues due to large databases lacking close matches for desired solutions.

Method used

Utilizing a trained data-driven model to generate suitable starting points for chemical plant simulations by interpolating between existing profiles, reducing the need for explicit database searches and improving convergence by predicting profiles based on input data.

Benefits of technology

This approach significantly reduces simulation time and improves convergence efficiency by generating accurate starting points for chemical plant operations, even when database matches are scarce.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure EP2026062755_12112026_PF_FP_ABST
    Figure EP2026062755_12112026_PF_FP_ABST
Patent Text Reader

Abstract

The disclosure is directed to a method (100) for monitoring operation of a chemical plant with unit operations associated with respective unit operation parameters. The method comprises obtaining (102) input data related to a subset of unit operation parameters of the chemical plant and providing (103) the input data to a trained data-driven model for generating (104) output profile data associated with output unit operation parameters. The data-driven model is trained to relate profile data associated with sets of unit operation parameters for operating the chemical plant, and corresponding input parameter data and / or is trained based on historical profile data and corresponding subsets of unit operation parameters. The generated output profile data is provided (106) for controlling the operation of the chemical plant.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] 240629

[0002] METHOD, APPARATUS AND SYSTEM FOR CONTROLLING AND / OR MONITORING A CONFIGURATION AND / OR OPERATION OF A CHEMICAL PLANT

[0003] FIELD OF THE INVENTION

[0004] The present disclosure is related to a method for monitoring and / or controlling a configuration and / or operation of a chemical plant associated with a plurality of unit operation connected by streams. The disclosure is further related to an apparatus and / or a system for monitoring and / or controlling a configuration or / an operation of a chemical plant. In addition, the disclosure related to an apparatus and / or a system for generating a trained data-driven model for monitoring and / or controlling a configuration and / or operation of a chemical plant, to the use of a trained data-driven model for generating output profile data, to the use of output profile data for monitoring and / or controlling the configuration and / or operation of a chemical plant, to a chemical plant and to a computer program.

[0005] BACKGROUND OF THE INVENTION

[0006] In process engineering, flowsheet simulators are used to design new plants, referred to as design case, and to optimize existing operating plants, referred to as rating case. The flowsheet consists of single unit operations (such as mixer, heater / cooler, flash stage, equilibrium stage column, and rate-based column) connected by streams. Unit operations need to be specified for process simulation based on the given problem and are specified by corresponding unit operation parameters associated therewith.

[0007] SUMMARY OF THE INVENTION

[0008] It is an object of the present disclosure to enable an improvement of the significance of selected profile data specifying or defining a chemical plant.

[0009] A first aspect of the present disclosure is formed by a method for monitoring and / or controlling a configuration and / or operation of a chemical plant. The chemical plant is associated with a plurality of unit operations, in particular connected by streams. The unit operations are associated with respective unit operation parameters.

[0010] The method of the first aspect comprises obtaining, in particular receiving or otherwise determining, input data associated with, or otherwise indicative of, at least a subset of unit operation parameters associated with target unit operations of the chemical plant and providing the obtained input data to a data-driven model for generating output profile data associated with output unit operation parameters. The data-driven model is trained to relate profile data associated with sets of unit operation parameters for configuring and / or operating the chemical plant, in particular complete sets of unit operation parameters that specify, define or characterize a chemical plant, and corresponding input parameter data associated with subsets of unit operation parameters and / or is trained based on historical profile data and corresponding subsets of unit operation parameters.240629

[0011] 2

[0012] The method of the first aspect then comprises generating the output profile data associated with the provided input data in accordance with the data-driven model and providing said output profile data for monitoring and / or controlling the configuration and / or operation of the chemical plant.

[0013] The chemical plant whose configuration or operation is to be monitored or controlled comprises a plurality of unit operations that are carried out in a respective unit associated with the chemical plant. A unit operation corresponds to a basic step in a chemical process, and involves or is associated to a physical change or chemical transformation, such as separation, crystallization, evaporation, filtration, polymerization, isomerization, or other reactions.

[0014] Each unit operation is characterized by corresponding unit operation parameters which correspond, for instance, to dimensioning parameters and / or process parameters (e.g., temperature, pressure, in-flow, out-flow, chemical composition, processing time, etc.) associated with the respective unit and / or the processing step. The chemical plant is then defined by the unit operations, the corresponding unit operation parameters and how they are connected to each other by the corresponding streams. The output profile data thus comprises a set of output unit operation parameters for the unit operations of the chemical plant, which specifies, or characterizes the operation of the chemical plant, for instance in term of input materials, output products, energy consumption, etc. such that different sets of output unit operation parameters are associated with chemical plants with different processing conditions.

[0015] The data-driven model is trained and used to obtain the output profile data, e.g., a complete set of unit operation parameters (the output unit operation parameters) for the target chemical plant whose configuration or operation is to be monitored or controlled. The data-driven model has been trained using available data including full sets of unit operation parameters, for instance using historical profile data (e.g., data indicative of a full set of unit operation parameters of a real chemical plant).

[0016] The configuration and / or operation of the chemical plant can then be monitored and / or controlled using the output profile data which includes the values of the unit operation parameters associated to the unit operations of the chemical plant, and which have been obtained using the input data.

[0017] The provided output profile data can be used to monitor a chemical plant, for instance by considering the unit parameters included in the output profile data as target values for the different unit operations, in particular unit operation parameters associated with the operation of the unit operations. In the context of the present disclosure, the expression "monitoring the chemical plant” may be understood broadly, and may refer, for example, to any kind of monitoring of the production processes carried out in the chemical plant. For instance If deviations between current unit operation parameters and the target values are detected, an alarm can be issued. Further, "monitoring the chemical plant" may also include planning the chemical plant, which can be done in a planning phase before implementing the actual chemical process.240629

[0018] 3

[0019] Additionally, or alternatively, the provided output profile data can be used to control the chemical plant, by actively driving the unit operations to operate in accordance with the unit operation parameters indicated in the output profile data.

[0020] Controlling the chemical plant may refer to any controlling measure or intervention within the chemical plant, a production network, a production step, or the like. This may comprise generating a control signal that modifies data within production, e.g. in or via a production control system, production control apparatus, operating system or the like. For example, such a controlling measure may, for example, comprise controlling a process step, e.g. by changing an energy used for the process step, by technical modification of the production step by an modified physical or chemical influence on an input material of the production step, etc. In other words, controlling the chemical plant may change the values of one or more unit operation parameters in the production reality. "Controlling the chemical plant" may also include planning the chemical plant, which can be done in a planning phase before implementing the actual chemical process.

[0021] One of the main challenges in solving equation-oriented models is finding a suitable starting point for the simulation. Without a proper starting point, simulation runs tend to fail to converge or run in to a wrong secondary solution. The closer the starting point is to the solution, the better the chances of a convergent run. The usual approach is selecting a known value or profile from a database. However, this approach has two main problems: firstly, in a large database, the runtime to find the best fitting profile increases significantly; secondly, if the database lacks a profile close enough to the desired solution, it becomes difficult to find a suitable starting point.

[0022] Thus, the disclosed solution involves utilizing a trained data-driven model, trained based on existing profiles (complete set of unit operation parameters), e.g., stored in a database, to generate suitable starting points (e.g., a complete set of unit operation parameter) based on the input data (e.g., an incomplete set of unit operation parameters). The suitable starting point can be then used, for example, in equation-oriented solvers. For instance, the data-driven model can be trained to transform input data for new simulations into a newly generated input file for the equation-oriented solver. This process effectively addresses the two problems mentioned earlier:

[0023] On the one hand, by using a trained data-driven model, the calculation time is moved away from the user experience. For example, available simulation results (or profiles, referred to as profile data) are stored in a database for matching with input data. The data-driven model is trained to generate profiles (the output profile data) matching the stored profiles based solely on the input data, which led to their calculation. Once trained, the network can generate best suitable output profile data from new input data. This significantly reduces time for each user simulation request, as there is no longer a need to run the database search explicitly during the process simulation. All the necessary timeconsuming operations are done in a separate training phase, which the user does not interact with.

[0024] On the other hand, the used of the trained data-driven model enables the generation of profiles (output profile data) not known before for faster convergence. Here, the data-driven model's understanding of unit operation parameter240629

[0025] 4

[0026] dependencies allows it to predict profiles within and around the existing stored profiles. If a simulation is started with input data indicative of a sub set of unit operation parameters that have not been stored in combination before, the solver can use the trained dependencies to generate a starting point (e.g. the provided output profile data) with closer parameters, rather than starting from a point that only fits a few parameters and iterating towards the real solution. Ultimately, this improves the efficiency of generating an output profile in the form of output profile data and with this monitoring and / or controlling a chemical plant based on said output profile.

[0027] The method of the first aspect of the disclosure can be also used for monitoring and / or controlling a design of a chemical plant. This is done by generating and providing output profile data based on input data associated with a subset of unit operation parameters corresponding to a target chemical plant to be designed.

[0028] In the following, embodiments of the method of the first aspect will be disclosed.

[0029] In an embodiment, the unit operation parameters may be associated with operating parameters associated with an operation of the unit operation and / or dimensioning parameters associated with a geometry of the unit operation. The unit operation refers to a step of the chemical process performed by chemical plant The operating parameters may include, but are not limited to, temperature, pressure, time, energy input, material flow rate, mass, etc., at a given unit operation or for materials involved in the unit operation. The dimensioning parameters may include physical dimensions or volumes of the given unit operation.

[0030] In another embodiment, the method of the first aspect may comprises comparing the output profile data with target operation profile data associated with a current chemical plant, and monitoring and / or controlling the operation of the chemical plant based on the comparison. Here, the target operation profile data may be profile data associated to an existing chemical plant operating in an expected way. Thus, this embodiment is directed to a feedback operation mode where the operation of the existing chemical plant is monitored / or controlled.

[0031] In another embodiment, the method may additionally or alternatively comprise comparing the output profile data with target configuration profile data associated with an expected chemical plant and monitoring and / or controlling a configuration of a chemical plant based on the comparison. Here, the target operation profile data may be profile data associated with an expected or planned chemical plant that is expected to operate in a predetermined way. Thus, this embodiment may be directed to a design process of a chemical plant or a simulation process of the operation of a chemical plant, which may be in a design state and not constructed or operated.

[0032] In another embodiment, the input data obtained may be user input data provided by a user, in particular provided via a user interface. The user may then provide, as input data, user input data associated with the subset of unit operation parameters that the users considers to be relevant, or the values based on which the remaining unit operation parameters are to be determined. The user interface may be a graphical interface including fields for selecting the240629

[0033] 5

[0034] corresponding unit operation parameter and another field for introducing the value of the selected unit operation parameter.

[0035] In another embodiment, the input data may additionally or alternatively comprise system or apparatus input data that is obtained, in particular received, from a database and / or using a mathematical model, wherein the mathematical model comprises one or more equations for relating user input data with the system or apparatus input data. This is particularly advantageous when the unit operation parameter value specified by the user is directly correlated with a value of another unit operation parameter, for instance due to a univocal interdependency of both unit operation parameters. The univocal interdependency may be given for example by the one or more equations.

[0036] In particular, in an embodiment, the chemical plant may be a gas treatment plant including, as unit operations, an absorber and a regenerator. In an embodiment, the unit operation parameters may include one or more of a pressure, a type of mass transfer internals, a temperature and acid gas loading in an absorbent depleted of acid gases in a unit operation, in particular in an absorber, a total mass transfer height or a composition specifying a proportion of one or more component processed by the chemical plant, a diameter of a unit operation, in particular an absorber, of the chemical plant or an acceptable hydraulic load of a unit operation, in particular an absorber, of the chemical plant, a flow rate associated with a unit operation, in particular an absorber, of the chemical plant or a loading factor indicating a distance to an equilibrium capture capacity of a treatment solution associated with a unit operation, in particular an absorber, of the chemical plant.

[0037] The unit operation parameters may also include or one more of a pressure, a type of mass transfer internals, a total mass transfer height of a unit operation, in particular a regenerator, a fraction quality of a part of the material processed by the chemical plant or a strip steam ratio or a loading factor indicating the distance to the equilibrium capture capacity of a regenerated treatment solution at an absorber top of the chemical plant or a composition specifying a proportion of one or more component processed by the chemical plant or the vapor outlet temperature at a regenerator top or an acid gas removal rate in the absorber.

[0038] According to a second aspect of the disclosure, a method or generating a trained data-driven model for monitoring and / or controlling a configuration and / or operation of a chemical plant that is associated with a plurality of unit operations associated with respective unit operation parameters is disclosed. The method of the second aspect of the disclosure comprises providing one or more sets of input parameter data, providing, for each set of input parameter data, corresponding profile data associated with a complete set of parameters for the unit operations of the chemical plant and training a data driven model using the plurality the sets of input parameter data and the corresponding profile data associated with the corresponding set of input parameter data for generating the trained data-driven model.

[0039] In an embodiment, the profile data may be obtained by providing the set of input parameter data to a predetermined mathematical model associated with the chemical plant. The predetermined mathematic model may form an equation-based solver.240629

[0040] 6

[0041] In yet another embodiment, the method of the second aspect also may comprise, for one or more sets of input parameter data, generating, using the set of input parameter data, a plurality of subsets of data items of the set of input parameter data, wherein the data items of the subsets are different; and wherein the step of training the data driven model is performed further using the plurality of generated subsets of data items and the respective profile data associated to the set of input parameter data.

[0042] In another embodiment, the data-driven model is trained to relate profile data and input parameter data associated with a plurality of different chemical compositions produced and / or processed by the one or more unit operations of the chemical plant. Additionally, or alternatively, the data-driven model is trained based on historical profile data and corresponding subsets of unit operation parameters associated with a plurality of different chemical compositions produced and / or processed by the one or more unit operations of the chemical plant.

[0043] The input data and / or the input parameter data may comprise an indication of a chemical composition produced and / or processed by the one or more unit operations.

[0044] Preferably, the trained data-driven model may be based on a neural network, e.g. a deep neural, wherein the neural network comprises an input later including a plurality of input neurons, each input neuron associated with a respective data item of the input parameter data, an output layer including a plurality of output neurons, each output neuron associated with a respective data item of the output profile data and one or more hidden layers arranged between the input layer and the output layer and configured to map dependencies between the respective data items.

[0045] A third aspect of the present disclosure is formed by an apparatus or a system, or an arrangement, for monitoring and / or controlling a configuration and / or operation of a chemical plant associated with a plurality of unit operations connected by streams, wherein the unit operations are associated with respective unit operation parameters. The system, apparatus or arrangement of the third aspect comprise an input interface that is configured to obtain, e.g. receive, input data indicative of at least a subset of unit operation parameters associated with unit operations of a chemical plant to be configured and / or operated. The system or apparatus also comprises a processing unit that is configured to implement a data-driven model trained according to the second aspect of the present disclosure to generate output profile data associated with output unit operation parameters, in particular with a complete set of unit operation parameters. The system or apparatus further comprises an output interface that is configured to provide the output profile data for monitoring and / or controlling the configuration and / or operation of the chemical plant.

[0046] The system or apparatus of the third aspect thus shares the advantages of the methods of the first and the second aspects of the present disclosure described above.

[0047] A fourth aspect of the present disclosure is formed by a system, apparatus or an arrangement for generating a trained data-driven model for monitoring and / or controlling a configuration and / or operation of a chemical plant that comprises a plurality of unit operations associated to parameters. The system or apparatus comprises a data input240629

[0048] 7

[0049] interface that is configured to obtaine, e.g., recevice, one or more sets of input parameter data, and, for each set of input parameter data, corresponding profile data associated with a complete set of parameters for the unit operations of the chemical plant, the profile data based on the respective set of input parameter data in accordance with a predetermined mathematical model. The system or apparatus also comprises or is configured to implement, a data driven model that is configured to be trained using the provided plurality of sets of input parameter data and the corresponding profile data associated with the corresponding set of input parameter data to generate the trained data-driven model.

[0050] Preferably, the data-driven model can be based on a neural network, wherein the neural network comprises an input later including a plurality of input neurons, each input neuron associated with a respective data item of the input parameter data, an output layer including a plurality of output neurons, each output neuron associated with a respective data item of the output profile data and one or more hidden layers arranged between the input layer and the output layer and configured to map dependencies between the respective data items.

[0051] Another aspect of the disclosure is directed to the use of a trained data driven model, in particular as described above, for generating output profile data indicative of a complete set of unit operation parameters associated with provided user input data for monitoring and / or controlling the configuration and / or operation of the chemical plant as generated according to the method of the first aspect of the present disclosure.

[0052] Yet another aspect of the present invention is directed to the use of output profile data generated according to the method of the first aspect for monitoring and / or controlling the configuration and / or operation of a chemical plant. In addition, the output profile data generated can be used for designing a chemical production plant.

[0053] Another aspect of the present disclosure is directed to a computer program comprising instruction which, when executed by a system or an apparatus in accordance with the third aspect of the disclosure, cause said system or apparatus to carry out the method of the first aspect.

[0054] Analogously, another aspect of the present disclosure is directed to a computer program comprising instruction which, when executed by a system or an apparatus in accordance with the fourth aspect, cause said system or apparatus to carry out the method of the second aspect of the disclosure.

[0055] Another aspect of the present disclosure is directed to a method for providing output profile data associated with one or more unit operation parameters characterizing or specifying a configuration and / or operation of a chemical plant and / or one or more unit operations associated with the chemical plant. The method of this aspect of the disclosure comprises obtaining, in particular receiving, user input data indicative of at least a subset of unit operation parameters associated with target unit operations of the chemical plant, providing the user input data to a data-driven model for generating output profile data associated with output unit operation parameters, wherein the trained data-driven240629

[0056] 8

[0057] model is trained to relate profile data associated with sets of unit operation parameters for configuring and / or operating the chemical plant, and corresponding input parameter data associated with subsets of unit operation parameters and / or is trained based on historical profile data and corresponding subsets of unit operation parameters, and providing the output profile data.

[0058] A final aspect of the present disclosure is formed by a chemical plant associated with a plurality of unit operations connected by streams, wherein the unit operations are associated with respective unit operation parameters. The chemical plant of this aspect comprises a system or apparatus for monitoring and / or controlling a configuration and / or operation of the chemical plant according to the third aspect of the present disclosure.

[0059] In a preferred embodiment, the chemical plant may be a gas treatment plant associated to unit operations comprising at least one absorber and one regenerator. The absorber can be configured to provide contact between a fluid containing one or more acid gases and an absorbent that is depleted of acid gases or at least with a concentration of acid gases lower than that of the fluid. The provision of contact between the fluid and the absorbent may result in a fluid with a reduced concentration of acid gas and an absorbent at least partially loaded with acid gases. The regenerator can be arranged and configured to regenerate the absorbent at least partially loaded with acid gases in order to obtain a fluid stream comprising acid gases and an at least partially regenerated absorbent.

[0060] In the gas treatment plant, and with respect to the unit operation parameters that specify the gas treatment plant, a fluid can be received that contains one or more acid gases. The unit operation parameters associated with the absorber may be one or more of the flow rate, pressure temperate and composition of the fluid. The absorber can be further specified by the absorber pressure, the type of mass transfer internals, the temperature, the acid gas loading in the absorbent depleted of acid gases and

[0061] i) at least by one of the following parameters of the total mass transfer height or a composition specifying a proportion of one or more depleted gas component(s) in the treated outlet stream;

[0062] ii) at least by one of the following parameters of the diameter or an acceptable hydraulic load indicating an acceptable hydraulic operational regime in the absorber; and

[0063] iii) at least by the flow rate of the absorbent depleted of acid gases or a loading factor indicating a distance to an equilibrium capture capacity of the treatment solution in the absorber.

[0064] In terms of unit operation parameters, the regenerator may be specified by the regenerator pressure, the type of mass transfer internals, the total mass transfer height, and

[0065] at least by one of the following parameters of the reboiler duty or fraction quality of the regenerated treatment solution or a strip steam ratio or a loading factor indicating the distance to the equilibrium capture capacity of the regenerated treatment solution at the absorber top or a composition specifying a proportion of one or more240629

[0066] 9

[0067] depleted gas component(s) in the treated outlet stream or the vapor outlet temperature at the regenerator top or an acid gas removal rate in the absorber; and

[0068] II) at least by one of the following parameters of the diameter or an acceptable hydraulic load indicating an acceptable hydraulic operational regime in the regenerator.

[0069] It shall be understood that the computer implemented methods described above, the systems or apparatuses described above, the chemical plant described above, and the computer program products described above have similar and / or identical preferred embodiments, in particular, as defined in the dependent claims.

[0070] It shall be understood that a preferred embodiment of the present invention can also be any combination of the dependent claim or above embodiments with a respective independent claim.

[0071] These and other aspects of the present invention will be apparent from and elucidated with reference to the embodiments described hereafter.

[0072] BRIEF DESCRIPTION OF THE DRAWINGS

[0073] In the following, the present disclosure is further described with reference to the enclosed figures. The same reference numbers in the drawings and this disclosure are intended to refer to the same or like elements, components, and / or parts.

[0074] In the following drawings:

[0075] FIG. 1 shows a schematic representation of an exemplary chemical plant in accordance with an embodiment of the present disclosure;

[0076] FIG. 2 shows a schematic flowsheet simulation of a chemical plant implemented as a gas treatment plant;

[0077] FIG. 3 shows a schematic flow diagram of an exemplary method for controlling and / or monitoring configuration and / or operation of a chemical plant in accordance with an embodiment of the present disclosure;

[0078] FIG. 4 shows a schematic flow diagram of an exemplary method for generating a trained data driven model for monitoring and / or controlling a configuration and / or operation of a chemical plant in accordance with an embodiment of the present disclosure;

[0079] FIG. 5 shows a schematic block diagram of an exemplary system for monitoring and / or controlling a configuration and / or operation of a chemical plant in accordance with an embodiment of the present disclosure;240629

[0080] 10

[0081] FIG. 6 shows a schematic block diagram of an exemplary system for generating a trained data-driven model for monitoring and / or controlling a configuration and / or operation of a chemical plant in accordance with an embodiment of the present disclosure; and

[0082] FIG. 7 shows a schematic representation of a data driven model implemented as a deep neural network.

[0083] DETAILED DESCRIPTION OF THE EMBODIMENTS

[0084] The following embodiments are mere examples for implementing the methods, apparatuses and systems disclosed herein and shall not be considered limiting.

[0085] FIG. 1 shows a schematic representation of an exemplary chemical plant 500. The chemical plant may receive a material input 550 that may include a plurality of input materials. The chemical plant may processes the input materials and may produce a material output 552, which may also include a plurality of output materials, which may be final products, intermediate products, by-products, waste products etc. The chemical plant 500 may comprise a plurality of unit operations. A unit operation corresponds to a basic step in a chemical process, and involves or is associated to a physical change or chemical transformation, such as separation, crystallization, evaporation, filtration, polymerization, isomerization, or other reactions. Each unit operation is characterized by corresponding unit operation parameters which may correspond, for instance, to dimensioning parameters and / or process parameters associated with the respective unit operation and / or the processing step. The chemical plant 500 may then defined or specified by the unit operations, the corresponding unit operation parameters and how the unit operations are connected to each other by the corresponding streams. Examples of unit operations and unit operation parameters will be discussed with reference to FIG. 2 below.

[0086] The chemical plant 500 may comprise a plant control system 520 that is configured to control and / or monitor the operation of the chemical plant 500. The plant control system 520 may be in signal communication with a plurality of sensors associated with the operation unit and thus may be configured to receive sensor data that is associated with, or otherwise indicative of, the current operational state of the different operation unit, and thus of the current operational state of the chemical plant 500. The plant control system comprises a system 300, or an arrangement, for monitoring and / or controlling a configuration and / or operation of the chemical plant 500, which will be discussed with more detail with reference to FIG. 5 below.

[0087] The plant control system 520 may comprise a system 400, or arrangement, for generating a trained data-driven model for monitoring and / or controlling a configuration and / or operation of the chemical plant 500. The system 400 may be integrated within plant control system. Alternatively, or additionally, the plant control system 520 may be signally connected to an external system 400b for generating the trained data-driven model for monitoring and / or con-240629

[0088] 11

[0089] trolling a configuration and / or operation of the chemical plant 500. The alternative system 400b may be a cloud system or a system located outside of the chemical plant 500 and connected to the plant control system via a suitable communication network 530.

[0090] FIG. 2 shows a schematic flowsheet simulation 501 of a chemical plant 500 implemented as a gas treatment plant.

[0091] Gas treatment plants are typically used in large-scale plants like oil and gas facilities, gas cleaning plants, carbon dioxide capture facilities, liquefied natural gas (LNG) plants, oil refineries, petrochemical facilities or chemical facilities. In such large-scale plants, fluid streams occur which contain acid gases such as CO2, H2S, SO2, CS2, HCN, COS or mercaptans. These fluid streams can be, for example, gas streams (such as natural gas, synthesis gas or heavy oil or heavy residues, coke-oven off-gases, refinery gas or reaction gases formed in the partial oxidation of organic materials, for example coal or mineral oil) or liquid or liquefied hydrocarbon streams, such as LPG (liquefied petroleum gas) or NGL (natural gas liquids). The removal of the acid gases from these fluid streams is desirable for various reasons.

[0092] Before these fluids can be transported or further processed, the acid gas content of the fluids must be markedly reduced. CO2, for example, must be removed from natural gas, since a high concentration of CO2 reduces the calorific value of the gas. Furthermore, CO2, in combination with the water frequently entrained in fluid streams can lead to corrosion on pipes and fittings.

[0093] The removal of sulfur compounds from these fluid streams is important for various reasons. For example, the sulfur compound content of natural gas must be reduced by suitable treatment measures immediately at the natural gas source, since the sulfur compounds, together with the water frequently entrained by natural gas, also form acids which act corrosively.

[0094] For the transport of the natural gas in a pipeline, therefore preset limit values of the sulfurous impurities must not be exceeded. Furthermore, numerous sulfur compounds, even at low concentrations, are foul-smelling, and, in particular hydrogen sulfide, toxic.

[0095] The exemplary chemical plant 500 of FIG. 2 is a gas treatment plant that includes two unit operations, namely an absorber 502 and a regenerator 504. The absorber 502 includes two inputs 11, 12, for receiving a fluid containing one or more acid gases and an absorbent that is depleted of acid gases. The absorber 502 is configured to provide contact between the fluid containing one or more acid gases and the absorbent that is depleted of acid gases. The provision of contact between the fluid and the absorbent in the absorber 502 results in a fluid with a reduced concentration of acid gas and an absorbent at least partially loaded with acid gases, e.g. part of the acid gases are transferred from the fluid to the absorbent such that the concentration of acid gases in the fluid is reduced and the concentration of acid gases in the absorbent is increased. The regenerator 504 is connected to the absorber via a stream 508a and it240629

[0096] 12

[0097] is arranged and configured to regenerate the absorbent at least partially loaded with acid gases in order to obtain a fluid stream comprising acid gases and an at least partially regenerated absorbent.

[0098] The operation of the absorber 502 in the chemical production plant is specified by a plurality of unit operation parameters 502a, 502b. The unit operation parameters 502a, 502b associated with the absorber 502 are the flow rate, pressure temperate and composition of the fluid. The absorber is further specified, in terms of unit operation parameters 502a, 502b, by the absorber pressure, the type of mass transfer internals, the temperature, the acid gas loading in the absorbent depleted of acid gases and

[0099] I) at least by one of the following parameters of the total mass transfer height or a composition specifying a proportion of one or more depleted gas components in the treated outlet stream 508a,

[0100] ii) at least by one of the following parameters of the diameter or an acceptable hydraulic load indicating an acceptable hydraulic operational regime in the absorber; and

[0101] ill) at least by the flow rate of the absorbent depleted of acid gases or a loading factor indicating a distance to an equilibrium capture capacity of the treatment solution in the absorber.

[0102] In terms of unit operation parameters 504a, 504b, the regenerator 504 is specified by the regenerator pressure, the type of mass transfer internals, the total mass transfer height, and

[0103] I) at least by one of the following parameters of the reboiler 505 duty or fraction quality of the regenerated treatment solution or a strip steam ratio OR a loading factor indicating the distance to the equilibrium capture capacity of the regenerated treatment solution at the absorber top or a composition specifying a proportion of one or more depleted gas component(s) in the treated outlet stream or the vapor outlet temperature at the regenerator top or an acid gas removal rate in the absorber; and

[0104] II) at least by one of the following parameters of the diameter or an acceptable hydraulic load indicating an acceptable hydraulic operational regime in the regenerator.

[0105] The values of the unit operation parameters 502a, 502b, 504a, 504b of an existing chemical plant can be determined using suitable sensors or based on design-data, for example in the case of unit operation parameters associated with dimensions of the units. The determined values can then be provided to the plant control system 502 as it is shown with reference to FIG. 1, and the values can be used for monitoring and / or controlling the operation of the plant.

[0106] For cases where the process of the chemical production plant has to be optimized or where a new chemical production plant has to be designed a flowsheet simulator may be used.240629

[0107] 13

[0108] In process engineering flowsheet simulators are applied to design new chemical plants (design case) and to optimize existing operating plants (rating case). The flowsheet is defined by single unit operations (like mixer, heater / cooler, flash stage, equilibrium stage column and rate based column) that are connected by streams 508. For process simulation, the unit operations 502, 504 have to be specified according to the given problem. Basically, there are two different approaches:

[0109] 1) In the design case, the specification of the unit operations is done according to given guidelines, whereas the dimensions of the equipment items and operational process parameters are coming as a result of the simulation.

[0110] 2) In the rating case, the dimensions of the equipment items and operational process parameters are available and thus used for specification.

[0111] The main characteristic of a flowsheet simulator is the method used to obtain solutions. There are sequential-modular solvers, which solve unit operations in sequence and iterate to meet recycle conditions. There are also equation-oriented solvers, which consider the entire flowsheet as a matrix of variables and equations. Both approaches have advantages and disadvantages, including flexibility of specification, convergence behavior, and algorithm robustness. However, the equation-oriented approach is generally considered superior to the sequential-modular approach.

[0112] One of the main challenges in solving equation-oriented models is finding a suitable starting point. Without a proper starting point, the simulation runs tend to fail to converge or run in to a "wrong” secondary solution. The closer the starting point is to the solution, the better the chances of a convergent run. The usual approach is selecting a known value / profile from a database. However, this approach has two main problems: firstly, with larger databases, the runtime to find the best fitting profile increases significantly; secondly, if the database lacks a profile close enough to the desired solution, it becomes difficult to find a suitable starting point.

[0113] For the simulation, the starting point, provided in the form of user input data, corresponds to pre-specified values of a set of unit operation parameters, and the result of solution of the simulation typically comprises values for a complete set of the unit operation parameters, or are least of unit operation parameters that are independent.

[0114] FIG. 3 shows a schematic flow diagram of an exemplary method 100 for controlling and / or monitoring configuration and / or operation of a chemical plant in accordance with an embodiment of the present disclosure, linked to a flow chart of an exemplary method 200 for training a data-driven model.

[0115] The method 100 is suitable for monitoring and / or controlling a configuration and / or operation of a chemical plant (see, for example,, the chemical plant 500 of FIG. 1) associated with a plurality of unit operations 502, 504, in particular connected by streams 508, as shown in, and discussed with, reference to FIG. 2. Also as discussed above with reference to FIG. 2, the unit operations 502, 504 are associated with respective unit operation parameters 502a, 502b, 504a, 504b.240629

[0116] 14

[0117] The method 100 comprises, in a step 102, obtaining (e.g. receiving) input data ID indicative of, or otherwise associated with at least a subset of unit operation parameters 502a, 502, 504a, 504b, that are associated with target unit operations 502, 504 of the chemical plant to be configured and / or operated. The method 100 also comprises, in a step 103, providing the obtained user data to a data-driven model 150.

[0118] The method 100 also include, in a step 104, generating output profile data associated with output unit operation parameters, in particular a complete set of unit operation parameters that specify the chemical plant. The data-driven model 150 is trained, for example as indicated by method 200 of FIG. 3, to relate profile data PD associated with sets of unit operation parameters, in particular complete sets thereof, for configuring and / or operating the chemical plant, and corresponding input parameter data IPD associated with subsets of unit operation parameters 502a, 502b, 504a, 504b and / or is trained based on historical profile data and corresponding subsets of unit operation parameters.

[0119] The method thus generates the output profile data in step 104, and provides the output profile data for monitoring and / or controlling the configuration and / or operation of the chemical plant in step 106. Thus, the method 100 is based on the use of the data-driven model to interpolate between available profiles or complete sets of unit operation parameters as a function of a given subset provided as input data to generate the output profile data.

[0120] The unit operation parameters associated with target unit operation may be those unit operation parameters whose values are available, or known, or fixed, and based on which the remaining unit operation parameters have to be determined using the data driven model. For instance, if one of the unit operations of a chemical plant has to be replaced by another one having different unit operation parameters, such as different dimensions, different working parameters in terms of temperature, pressure, flow, etc., these unit operation parameters are given as input data and the output profile data determined using the data-driven model may be indicative of the remaining unit operation parameters of the unit operations that were not provided in the input data. The method 100 can also be used for designing a chemical plant, by providing, as input data, the values of unit operation parameters associated with predetermined fixed requirements of the chemical plant being designed, for instance in terms of desired throughput, of input materials, of unit operations already available to be used in the plant, etc.

[0121] The provided output profile data can be compared, in a step 108a, with target operation profile data that is associated with a current chemical plant, and indicative, for instance, of a current working point of the chemical plant for monitoring, in a step 110a, the operation of the chemical plant. For instance, input data associated with a subset of unit operation parameters that are known can be provided to the trained data-driven model. The resulting output profile data can be compared to the current values of the unit operation parameters, for example as determined by respective sensors associated to the respective unit operations, for determining whether the determined real values of the unit operation parameters correspond to the values provided in the output profile data. These values can also be used, in step 108a, to control the operation of the chemical plant, by causing a change in the unit operations such that the values of the operation unit parameters are driven towards the values provided in the output profile data.240629

[0122] 15

[0123] Alternatively, or additionally, the provided output profile data can be compared, in a step 108b, with target configuration profile data associated with an expected chemical plant, e.g. a chemical plant to be designed. Here the target configuration profile is associated with, or otherwise indicative of, predetermined requirements that have to be fulfilled by the chemical plant to be designed. The comparison can be then used, in a step 110b, to monitor or control the configuration or design of the chemical plant.

[0124] The user data obtained in step 102 may be user input data provided by a user, in particular provided via a user interface. The user may therefore provide, as input data, user input data associated with the subset of unit operation parameters that the users considers to be relevant, or the values based on which the remaining unit operation parameters are to be determined. The user interface may be a graphical interface including fields for selecting the corresponding unit operation parameter and another field for introducing the value of the selected unit operation parameter.

[0125] Additionally, or alternatively, the input data may comprise system input data that is obtained, in particular received, from a database and / or using a mathematical model, wherein the mathematical model comprises one or more equations for relating user input data with the system input data. This is particularly advantageous when the unit operation parameter value specified by the user is directly correlated with a value of another unit operation parameter, for instance due to a univocal interdependency of both unit operation parameters. The univocal interdependency may be given for example by the one or more equations.

[0126] FIG. 3 also shows a flow chart of an exemplary method 200 for generating the trained data-driven model for monitoring and / or controlling a configuration and / or operation of a chemical plant. The method comprises, in a step 202, obtaining (e.g., receiving) one or more sets of input parameter data, and obtaining, in a step 204, and for each set of input parameter data, corresponding profile data associated with a complete set of parameters for the unit operations of the chemical plant and training, in a step 206, a data driven model using the plurality the sets of input parameter data and the corresponding profile data associated with the corresponding set of input parameter data for generating the trained data-driven model, that is then used in the method 100 described above.

[0127] FIG. 4 shows a schematic flow diagram of another exemplary method 200 for generating a trained data driven model for monitoring and / or controlling a configuration and / or operation of a chemical plant in accordance with an embodiment of the present disclosure. Here, the profile data provided in step 204 (see FIG. 3) is obtained by providing the set of input parameter data to a predetermined mathematical model associated with the chemical plant, in particular to a topology of the chemical plant. The mathematical model may be an equation-oriented model and / or a sequential modular solver.

[0128] The method 200 discussed with reference to FIG. 4 may also comprise, in a step 208, generating, using the set of input parameter data, a plurality of subsets of data items of the set of input parameter data, wherein the data items of240629

[0129] 16

[0130] the subsets are different. Here, the step 206 of training the data driven model is performed further using the plurality of generated subsets of data items and the respective profile data associated to the set of input parameter data.

[0131] In particular, the data-driven model can be trained to relate profile data and input parameter data associated with a plurality of different chemical compositions produced and / or processed by the one or more unit operations of the chemical plant. Additionally, or alternatively, the data-driven model can be trained based on historical profile data and corresponding subsets of unit operation parameters associated with a plurality of different chemical compositions produced and / or processed by the one or more unit operations of the chemical plant.

[0132] In particular, the method 200 shown in FIG. 4 may be used when a starting point needs to be found for only a portion of the flowsheet equipment items, e.g. the operation units. In this case, the relevant operation units, or equipment items of the chemical plant, are divided into contextual groups, and a data-driven model is trained to deliver partial starting point solutions for each group. The starting point, e.g. the output profile data delivered by the trained data driven model, can then be obtained as a combination of these partial starting points.

[0133] As an example, this embodiment in particularly advantageous when there is a need to exchange compositions or mixtures, which using the known techniques results in difficult procedures that require many runs with minimal changes in concentration, in particular due to the way the program utilizes starting values and converges. However, using the trained data-driven model, a convergent solution can be directly determined, as the data-driven model can be trained for example across solvents / compositions and thus can compute a process solution for a particular solvent A with configuration B and one for another solvent C with another configuration D, as well as for a mixture like solvent A with configuration D. Thus, it is both possible to train at least one data-driven model per group (as a specialized and more realistic data-driven model) or at least one model for all groups.

[0134] FIG. 5 shows a schematic block diagram of an exemplary system 300 for monitoring and / or controlling a configuration and / or operation of a chemical plant 500. In particular, the system 300 is adapted and arranged for monitoring and / or controlling a configuration and / or operation of a chemical plant associated with a plurality of unit operations connected by streams, wherein the unit operations are associated with respective unit operation parameters, as shown for example with reference to FIG. 2.

[0135] The system 300 comprises a user input interface 302 configured to obtain (e.g. receive) input data ID indicative of, or otherwise associated with, at least a subset of unit operation parameters associated with unit operations of a chemical plant to be configured and / or operated. The input data can be data provided by the user via a input interface, and referred to as user input data. Additionally or alternatively, the input data can be system input data, as described above with reference to FIG. 3.240629

[0136] 17

[0137] The system also comprises a processing unit 150 configured to implement a data-driven model 150 that has been trained, for instance as discussed with reference to FIG. 3 and FIG. 4, to generate output profile data OPD associated with output unit operation parameters, in particular a complete set of unit operation parameters that specify the chemical plant 500. The system further comprises an output interface 304 configured to provide the output profile data for monitoring and / or controlling the configuration and / or operation of the chemical plant.

[0138] FIG. 6 shows a schematic block diagram of an exemplary system 400 for generating a trained data-driven model 150 for monitoring and / or controlling a configuration and / or operation of a chemical plant. The system 400 comprises a data input interface 402 that is configured to obtain (e.g., receive) one or more sets of input parameter data IPDa, IPDb, and, for each set of input parameter data IPDa, IPDb, corresponding profile data PDa, PDb associated with a complete set of parameters for the unit operations of the chemical plant. In particular, the profile data based on the respective set of input parameter data is determined in accordance with a predetermined mathematical model. The system comprises or is configured to implement a data driven model DDM that is configured to be trained using the provided plurality of sets of input parameter data and the corresponding profile data associated with the corresponding set of input parameter data to generate the trained data-driven model 150. For instance, by providing existing profiles, e.g. complete sets of parameters for the unit operations (unit operation parameters) as profile data PDa, PDb, any subset of parameter values may be regarded as a set of input parameter data IPDa, IPDb. Additionally, or alternatively, pairs of input parameter data, which using for instance the predetermined mathematical model have resulted in convergent and applicable profile data, can be provided as training data to the system.

[0139] The input datasets IPDa, IPDb provided for training the data-driven model DDM typically comprises of values of the operation unit parameters (e.g., composition of feed gas, height / diameter of columns, pressure, temperature, flow rates, and dimensionless specifications such as safety factors or strip steam ratio) and information about the chemical plant (which operation units are available, which solvent is used, which packings are in the column, etc.). From each of these datasets, some real user inputs can be extracted. This is because in each calculation, the user can only specify a subset of these parameters, which depend on each other. For example, either the height of the absorber or the purity of the treated gas stream can be specified, as they are dependent on each other. Similarly, the safety factor, diameter, and other values also depend on each other. In addition, there is a corresponding profile data PDa, PDb (e.g. in the form of a CPR file) for each input dataset. The CPR file has a specific structure but can be displayed like a .txt file. It contains all the information and values the unit operation parameters, e.g., flows, concentrations, dimensions, etc., and thus represents the exact solution of the otherwise iteratively solvable system of equations. This way, many training pairs consisting of a valid input IPDa, IPDb (i.e., a specification-complete subset of the dataset) and the corresponding profile data PDa, PDb (e.g., CPR) file can be obtained from each stored dataset.

[0140] The data-driven model, e.g., a deep neural network, can then be trained to learn to map the valid input to the values stored in the profile data (e.g., CPR file).FIG. 7 shows a schematic representation of a data driven model 150 implemented as a deep neural network DNN. The deep neural network DNN comprises an input later 151 including a plurality of input neurons 151.1 - 151. n, each input neuron associated with a respective data item of the input parameter data IPDa. The deep neural network DNN also comprises an output layer 152 including a plurality of output neurons 152.1 - 152. n, each output neuron associated with a respective data item of the output profile data OPD. Arranged between the input layer 151 and the output layer 152, the deep neural network DNN comprise one or more hidden layers 153.1 - 153.k that configured to map dependencies between the respective data items.

[0141] In particular, the input layer 151 thus consists of a set of n neurons, where n corresponds to the number of stored values in the dataset provided (note that in each valid input, a subset of the neurons is then assigned a value of 0). These are connected to each other in a first hidden layer 153.1 in such a way that the dependence / redundancy of the different specifications is mapped. By means of a deep neural structure, which can be based, for instance on a variational autoencoder (VAE), a generative adversarial network (GAN), a convolutional neural network (CNN), a recurrent neural network (RNN) such as a long short-term memory (LSTM), the mapping is performed on an output layer 152 consisting of m neurons, where m corresponds to the number of all output values that occur in the profile data (e.g., the CPR files). The output profile data OPD may be read out and assembled and stored according to the write / read logic of the CPR files.

[0142] In summary, the disclosure is directed to a method for monitoring operation of a chemical plant with unit operations connected by streams and associated with respective unit operation parameters. The method comprises obtaining input data related to a subset of unit operation parameters of the chemical plant and providing it to a data-driven model for generating output profile data associated with output unit operation parameters. The data-driven model is trained to relate profile data associated with sets of unit operation parameters for operating the chemical plant, and corresponding input parameter data and / or is trained based on historical profile data and corresponding subsets of unit operation parameters. The generated output profile data is provided for controlling the operation of the chemical plant.

[0143] The present disclosure has been described in conjunction with preferred embodiments and examples as well. However, other variations can be understood and effected by those persons skilled in the art and practicing the claimed disclosure, from the studies of the drawings, this disclosure and the claims.

[0144] Any steps presented herein can be performed in any order. The methods disclosed herein are not limited to a specific order of these steps. It is also not required that the different steps are performed at a certain place or in a certain computing node of a distributed system, i.e. each of the steps may be performed at different computing nodes using different equipment / data processing.

[0145] As used herein ..determining" also includes ..initiating or causing to determine", "generating" also includes ..initiating and / or causing to generate" and "providing” also includes "initiating or causing to determine, generate, select, send19

[0146] and / or receive”. "Initiating or causing to perform an action” includes any processing signal that triggers a computing node or device to perform the respective action.

[0147] In the claims as well as in the description the word "comprising” or "including” or similar wording does not exclude other elements or steps and shall not be construed limiting to the elements or steps lined out. The indefinite article "a” or "an” does not exclude a plurality. A single element or other unit may fulfill the functions of several entities or items recited in the claims. The mere fact that certain measures are recited in the mutual different dependent claims does not indicate that a combination of these measures cannot be used in an advantageous implementation or further elements may be included.

[0148] Providing in the scope of this disclosure may include any interface configured to provide data. This may include an application programming interface, a human-machine interface such as a display and / or a software module interface. Providing may include communication of data or submission of data to the interface, in particular display to a user or use of the data by the receiving node, entity or interface.

[0149] Various units, circuits, entities, nodes or other computing components may be described as "configured to” perform a task or tasks. Configured to shall recite structure meaning "having circuitry that” performs the task or tasks on operation. The units, circuits, entities, nodes or other computing components can be configured to perform the task even when the unit / circuit / component is not operating. The units, circuits, entities, nodes or other computing components that form the structure corresponding to "configured to” may include hardware circuits and / or memory storing program instructions executable to implement the operation. The units, circuits, entities, nodes or other computing components may be de-scribed as performing a task or tasks, for convenience in the description. Such descriptions shall be interpreted as including the phrase "configured to.” Any recitation of "configured to” is expressly intended not to invoke 35 U.S.C. § 112(f) interpretation.

[0150] In general, the methods, apparatuses, systems, computer elements, nodes or other computing components described herein may include memory, software components and hard-ware components. The memory can include volatile memory such as static or dynamic random-access memory and / or nonvolatile memory such as optical or magnetic disk storage, flash memory, programmable read-only memories, etc. The hardware components may include any combination of combinatorial logic circuitry, clocked storage devices such as flops, registers, latches, etc., finite state machines, memory such as static random-access memory or embedded dynamic random-access memory, custom designed circuitry, programmable logic arrays, etc.

[0151] Moreover, any of the methods, method steps, processes and actions described or illustrated herein may be implemented using executable instructions in a general-purpose or special-purpose processor and stored thereon. A processor may be a processor of any suitable type, and is preferably a processor configured for parallel processing of at least a hundred or a at least a thousand threads in parallel, e.g. a graphical processing unit (GPU). For instance, the processor comprises at least a hundred or at least a thousand parallel processing cores. In particular, the processor20

[0152] may comprise at least one (preferably at least a thousand) compute unified device architecture (CUDA) core(s), which may allow for using a graphical processing unit as the processor, which may increase computational efficiency. For instance, the processor may comprise at least one (e.g. at least a hundred) streaming multiprocessor cores, which may allow for increasing the data throughput. As a further ex-ample, the processor may comprise one or more (e.g. at least a hundred) tensor core(s). A tensor core may be specifically adapted to perform matrix operations and may allow to accelerate large matrix operations. A tensor core may be configured to perform mixed-precision matrix multiply and accumulate calculations in a single operation. For instance, a tensor core may perform mixed-precision floating-point matrix arithmetic, specifically utilizing FP16 (half-precision) inputs to produce either full-precision (FP32) or half-precision (FP16) outputs. In the case of FP16 output, a tensor core may provide a performance boost by storing the intermediate accumulation results in FP32 format, thereby maintaining the precision necessary for accurate results. For example, a processor may comprise several thousand tensor cores, each capable of performing 64 floating point FMA (Fused Mui tiply-Add) operations per clock cycle. With these capabilities, such a GPU may allow for hundreds of TFLOPs (Tera Floating-Point Operations per Second) of performance in mixed-precision computations. Furthermore, a tensor core may support a variety of numerical formats, including IEEE standard half-precision, single-precision, and double-precision floating-point formats, as well as a range of integer formats.

Claims

24062921CLAIMS1. Method (100) for monitoring and / or controlling a configuration and / or operation of a chemical plant (500) associated with a plurality of unit operations (502, 504), wherein the unit operations are associated with respective unit operation parameters (502a, 502b, 504a, 504b), the method comprising:obtaining, in particular receiving, (102) input data (ID) indicative of at least a subset of unit operation parameters associated with target unit operations (502, 504) of the chemical plantproviding (103) the input data (ID) to a data-driven model (150) for generating (104) output profile data (OPD) associated with output unit operation parameters,wherein the data-driven model is trained to relate profile data (PD) associated with sets of unit operation parameters for configuring and / or operating the chemical plant, and corresponding input parameter data (IPD) associated with subsets of unit operation parameters (502a, 502b, 504a, 504b) and / or is trained based on historical profile data and corresponding subsets of unit operation parameters; andproviding (106) the output profile data for monitoring and / or controlling the configuration and / or operation of the chemical plant.

2. The method (100) of claim 1, wherein the unit operation parameters (502a, 502b, 504a, 504b) are associated with operating parameters associated with an operation of the unit operation (502, 504) and / or dimensioning parameters associated with a geometry of the unit operation (502, 504).

3. The method (100) of claim 1 or 2, further comprising:comparing (108a) the output profile data with target operation profile data associated with a current chemical plant; andmonitoring and / or controlling (110a) the operation of the chemical plant based on the comparison.

4. The method of any of the preceding claims 1-3, further comprisingcomparing (108b) the output profile data with target configuration profile data associated with an expected chemical plant; andmonitoring (110b) and / or controlling a configuration of a chemical plant based on the comparison.240629225. The method of any of the preceding claims, wherein the input data comprises user input data, in particular received via a user interface.

6. The method of any of the preceding claims, wherein the input data comprises system input data obtained, in particular received, from a database and / or obtained using a mathematical model, wherein the mathematical model comprises one or more equations for relating user input data with the system input data.

7. The method of any of the preceding claims wherein the unit operation parameters include one or more of:a pressure, a type of mass transfer internals, a temperature and acid gas loading in an absorbent depleted of acid gases in a unit operation, in particular in an absorber, a total mass transfer height or a composition specifying a proportion of one or more component processed by the chemical plant, a diameter of a unit operation, in particular an absorber, of the chemical plant or an acceptable hydraulic load of a unit operation, in particular an absorber, of the chemical plant, a flow rate associated with a unit operation, in particular an absorber, of the chemical plant or a loading factor indicating a distance to an equilibrium capture capacity of a treatment solution associated with a unit operation, in particular an absorber, of the chemical plant; and / or or one more ofa pressure, a type of mass transfer internals, a total mass transfer height of a unit operation, in particular a regenerator, a fraction quality of a part of the material processed by the chemical plant or a strip steam ratio or a loading factor indicating the distance to the equilibrium capture capacity of a regenerated treatment solution at an absorber top of the chemical plant or a composition specifying a proportion of one or more component processed by the chemical plant or the vapor outlet temperature at a regenerator top or an acid gas removal rate in the absorber.

8. Method (200) for generating a trained data-driven model for monitoring and / or controlling a configuration and / or operation of a chemical plant (500) that is associated with a plurality of unit operations associated with respective unit operation parameters, the method comprising:obtaining (202), in particular receiving, one or more sets of input parameter data;obtaining (204), in particular receiving or determining, for each set of input parameter data, corresponding profile data associated with a complete set of parameters for the unit operations of the chemical plant;training (206) a data driven model using the plurality the sets of input parameter data and the corresponding profile data associated with the corresponding set of input parameter data for generating the trained data-driven model.

9. The method (200) of claim 8, wherein the profile data is obtained by providing the set of input parameter data to a predetermined mathematical model associated with the chemical plant.

10. The method of any of the preceding claims,wherein the data-driven model is trained to relate profile data and input parameter data associated with a plurality of different chemical compositions produced and / or processed by the one or more unit operations of the chemical plant; and / orwherein the data-driven model is trained based on historical profile data and corresponding subsets of unit operation parameters associated with a plurality of different chemical compositions produced and / or processed by the one or more unit operations of the chemical plant.

11. Apparatus (300) for monitoring and / or controlling a configuration and / or operation of a chemical plant associated with a plurality of unit operations connected by streams, wherein the unit operations are associated with respective unit operation parameters, the system comprising:an input interface (302) configured to obtain, in particular receive, input data indicative of at least a subset of unit operation parameters associated with unit operations of a chemical plant to be configured and / or operated; anda processing unit (150) configured to implement a data-driven model trained according to any of the claims 5 to 7 to generate output profile data (OPD) associated with output unit operation parameters; andan output interface (304) configured to provide the output profile data for monitoring and / or controlling the configuration and / or operation of the chemical plant.

12. Apparatus (400, 400b) for generating a trained data-driven model for monitoring and / or controlling a configuration and / or operation of a chemical plant that comprises a plurality of unit operations associated to parameters, the apparatus comprising:a data input interface (402) configured to obtain, in particular receive, one or more sets of input parameter data, and, for each set of input parameter data, corresponding profile data associated with a complete set of parameters for the unit operations of the chemical plant, the profile data based on the respective set of input parameter data in accordance with a predetermined mathematical model; and24a data driven model (DDM) configured to be trained using the provided plurality of sets of input parameter data and the corresponding profile data associated with the corresponding set of input parameter data to generate the trained data-driven model.

13. Use of a trained data driven model for generating output profile data indicative of a complete set of unit operation parameters associated with obtained input data for monitoring and / or controlling the configuration and / or operation of the chemical plant as generated according to the method of any of the preceding claims 1 to 7.

14. Use of output profile data generated according to the method of any of the preceding claims 1 to 7 for monitoring and / or controlling the configuration and / or operation of the chemical plant.

15. Chemical plant (500) associated with a plurality of unit operations (502, 504, wherein the unit operations are associated with respective unit operation parameters (502a, 502b, 504a, 504b), the chemical plant comprising an apparatus (300) for monitoring and / or controlling a configuration and / or operation of the chemical plant according to claim 11.