Method for operating continuous catalytic reformer unit

Abstract

A method for operating a continuous catalytic reformer unit can include passing a hydrocarbon reactant stream to a continuous catalytic reformer unit comprising at least one stream preheater, at least one catalytic reactor, and at least one separation unit to form one or more product effluent streams; a hydrocarbon reformer process control system is implemented, the hydrocarbon reformer process control system including a hydrocarbon reformer variable data memory including processor executable instructions, a hydrocarbon reformer output conversion module, and one or more predictive hydrocarbon reformer modeling processors. The predictive hydrocarbon reformer modeling processor is configured to execute processor executable instructions and cause a process control system to: receive one or more signals from one or more state variable actuator hardware, the signals indicating one or more current state variables, wherein the current state variable is a process variable that cannot be directly set in the continuous catalytic reformer unit; and receiving one or more signals from the one or more control variable actuator hardware, the signals indicating one or more current control variables of the continuous catalytic reformer unit, where the current control variables are process variables that can be set directly in the continuous catalytic reformer unit; and generating an improved control variable by utilizing a machine learning model, the improved control variable increasing a selected performance variable based on an input of one or both of the one or more current state variables or the one or more current control variables, wherein the improved control variables in the machine learning model are trained by using inputs of at least historical state variable data, historical control variable data, and historical performance variable data; and adjusting one or more current control variables of the continuous catalytic reformer unit based on the improved control variables determined by the machine learning model.

Description

Cross Reference to Related Applications

[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 509,618 filed June 22, 2023, and U.S. Non-Provisional Patent Application No. 18 / 462,068 filed September 6, 2023, the entire contents of both of which are hereby incorporated herein in their entirety. TECHNICAL FIELD

[0002] Embodiments of the present disclosure generally relate to the field of chemical processing, and more specifically to methods and systems for reforming hydrocarbon materials. BACKGROUND

[0003] Catalytic reforming is a chemical process that can be used to convert refined hydrocarbons, such as naphtha streams distilled from crude oil, into higher octane products, which are sometimes referred to as reformates. Such reformates can be used as gasoline. Due to the industry’s large demand for reformate-produced gasoline and other products, improved methods for operating reformer units have become an urgent need in the industry. SUMMARY

[0004] According to one or more embodiments, a method for operating a continuous catalytic reformer unit can include: passing a hydrocarbon reactant stream to a continuous catalytic reformer unit to form one or more product effluent streams, the continuous catalytic reformer unit comprising at least one stream preheater, at least one catalytic reactor, and at least one separation unit; implementing a hydrocarbon reformer process control system, the hydrocarbon reformer process control system comprising a hydrocarbon reformer variable data store comprising processor-executable instructions, one or more predictive hydrocarbon reformer modeling processors configured to execute the processor-executable instructions and cause the process control system to: receive one or more signals from one or more state variable actuator hardware, the signals indicative of one or more current state variables, wherein the current state variables are process variables that cannot be directly set in the continuous catalytic reformer unit; and receive one or more signals from one or more control variable actuator hardware, the signals indicative of one or more current control variables of the continuous catalytic reformer unit, wherein the current control variables are process variables that can be directly set in the continuous catalytic reformer unit; and generate an improved control variable by utilizing a machine learning model, the improved control variable increasing a selected performance variable based on input of one or both of the one or more current state variables or the one or more current control variables, wherein the improved control variable in the machine learning model is trained utilizing input of at least historical state variable data, historical control variable data, and historical performance variable data; and adjust the one or more current control variables of the continuous catalytic reformer unit based on the improved control variable determined by the machine learning model.

[0005] Other features and advantages of the described embodiments will be set forth in the detailed description which follows, and in part will be readily apparent to those skilled in the art who practice the described embodiments, including realizing the features and advantages described by reading the description below, including the detailed description, claims, and drawings. BRIEF DESCRIPTION OF DRAWINGS

[0006] The following detailed description of specific embodiments of the present disclosure can best be understood when read in conjunction with the following drawings, in which like reference numerals indicate identical structures throughout the several views, in which: Figure 1 is a generalized schematic of a continuous catalytic reformer unit in accordance with one or more embodiments described in the present disclosure; Figure 2 is schematically depicted a machine learning model used in the operation of a continuous catalytic reformer unit in accordance with one or more embodiments described in the present disclosure; Figure 3 is shown an example model generated using a machine learning model used in the operation of a continuous catalytic reformer unit in accordance with one or more embodiments described in the present disclosure.

[0007] For the purpose of describing and simplifying the schematic diagrams and related illustrations of the relevant chemical processing systems, numerous valves, temperature sensors, electronic controllers, and other elements that can be employed, which are well known to those of ordinary skill in the art of chemical processing operations, are not included in the description. In addition, supporting components that are typically included in a typical chemical processing operation, such as air supply systems, catalyst hoppers, and flue gas treatment systems, are not shown in the figures. Supporting components in a hydrocracking unit, such as effluent streams, spent catalyst discharge subsystems, and catalyst replacement subsystems, are also not shown. It will be understood that these components are within the spirit and scope of the embodiments of the present disclosure. However, operational components, such as those described in the present disclosure, can be added to the embodiments described in the present disclosure.

[0008] It should also be noted that the relevant arrows in the figures can represent process streams. However, the arrows can also equivalently represent transfer lines that can be used to transfer process streams between two or more system components. Further, arrows connected to a system component define an inlet or outlet for each given system component. The direction of the arrow generally corresponds with the predominant flow direction of the material contained within the physical transfer line represented by the arrow. Further, arrows not connecting two or more system components represent product streams exiting the illustrated system, or system inlet streams entering the illustrated system. Product streams can be further processed in a companion chemical processing system, or can be commercialized as an end product. System inlet streams can be streams transferred from a companion chemical processing system, or can be streams of unprocessed feedstock. Some arrows can represent recycle streams, i.e., streams of effluent from a system component that are recycled back into the system. However, it should be understood that in some embodiments, any represented recycle stream can be replaced by a system inlet stream of the same material, and a portion of a recycle stream can exit the system as a system product.

[0009] Further, the arrows in the illustrations can schematically depict the process steps of transporting a stream from one system component to another. For example, an arrow pointing from one system component to another can represent the "passing" of an effluent of a system component to another system component, which can include the contents of the process stream "leaving" or being "removed" from one system component, and the contents of the product stream being "introduced" to another system component.

[0010] It should be understood that, depending on the embodiment presented in the relevant illustration, an arrow between two system components can represent that the stream is not processed between the two system components. In other embodiments, the stream represented by the arrow can have substantially the same composition throughout the transport between the two system components. Further, it should be understood that, in one or more embodiments, an arrow can represent that at least 75 wt%, at least 90 wt%, at least 95 wt%, at least 99 wt%, at least 99.9 wt%, or even 100 wt% of the stream is transported between system components. Thus, in some embodiments, less than the entirety of the stream represented by an arrow can be transported between system components, such as when there is a slip stream.

[0011] It should be understood that when two or more lines in a schematic flow diagram in the relevant illustration cross one another, it is meant that the two or more process streams are "mixed" or "combined." Mixing or combining can also include the direct introduction of two streams into the same reactor, separation unit, or other system component for mixing. For example, it should be understood that when two streams are depicted as being directly combined prior to entering a separation unit or reactor, in certain embodiments, the streams can equivalently be introduced into the separation unit or reactor and mixed within the reactor.

[0012] Various embodiments will now be described in greater detail, some of which are illustrated in the accompanying drawings. Wherever possible, the same reference numbers will be used in all drawings to refer to the same or like parts. DETAILED DESCRIPTION

[0013] Described herein are methods of operating a continuous catalytic reformer unit, sometimes referred to herein as a “CCR” unit, for processing a hydrocarbon reactant stream, such as those that convert naphtha to a product effluent stream (such as reformate). As described in detail herein, such continuous catalytic reformer units can utilize a hydrocarbon reformer process control system that can predict system properties of the continuous catalytic reformer unit through use of a machine learning model. Specifically, the continuous catalytic reformer units described herein can operate to generate improved control variables based on a machine learning model developed from at least historical control variable data and historical state variable data. Based on the model, the continuous catalytic reformer unit can be adjusted such that current control variables are adjusted to the generated improved control variables.

[0014] As described herein, continuous catalytic reformer units generally process a hydrocarbon reactant stream into one or more product streams that can be further processed or collected for use as a product. For example, a continuous catalytic reformer unit can utilize a naphtha stream or straight run gasoline as a reactant stream, and can generally increase the octane rating of such a reactant stream. Reforming, the process conducted in a continuous catalytic reformer unit, can generally be defined as the overall effect of molecular changes (i.e., hydrocarbon reactions) such as cyclohexane dehydrogenation, alkylcyclopentane dehydrogenation isomerization, and paraffin and olefin dehydrogenation cyclization to produce aromatics; n-paraffin isomerization; alkylcycloparaffin isomerization to cyclohexane; isomerization of substituted aromatics; and / or hydrogenation cracking of gas-producing paraffins and coke. Catalytic reforming utilizes a catalyst. Some such catalysts can include a metal hydrogenation-dehydrogenation (hydrogen transfer) component, or a component that is substantially atomically dispersed on the surface of a porous inorganic oxide support, such as alumina, sometimes platinum. The support generally contains a halide, particularly a chloride, which can provide an acidic function that contributes to isomerization, cyclization, and hydrogenation cracking reactions.

[0015] As used in the present disclosure, a "catalyst" refers to any substance that increases the rate of a particular chemical reaction. The catalysts described in the present disclosure can be used to facilitate various reactions, such as, but not limited to, hydrogenation, dehydroisomerization, isomerization, and hydrocracking. As used in the present disclosure, "cracking" generally refers to a chemical reaction in which carbon-carbon bonds are broken. For example, a molecule having carbon-carbon bonds is broken into more than one molecule by breaking one or more carbon-carbon chains, or is converted from a compound containing alkyl or cyclic moieties (such as alkanes, cycloalkanes, naphthalenes, arenes, etc.) to an olefin compound and / or a compound that does not contain a cyclic moiety or contains fewer cyclic moieties than before cracking.

[0016] Referring now to the drawings Figure 1 , a continuous catalytic reformer unit 100 is shown, which generally includes a chemical processing section 104 (including all chemical processing sections of the continuous catalytic reformer unit 100) and a hydrocarbon reformer process control unit 200. It should be understood that other continuous catalytic reformer units can be employed in the presently disclosed process, while Figure 1 The continuous catalytic reformer unit 100 of FIG. 1 is an example of a continuous catalytic reformer unit.

[0017] The chemical processing section 104 of the continuous catalytic reformer unit 100 can generally include a stream preheater 110, a catalytic reactor 130, and a separation unit 150. Figure 1 The continuous catalytic reformer unit 100 of FIG. 1 includes two stream preheaters 110 and two catalytic reactors 130, which are arranged in an alternating series. Specifically, the continuous catalytic reformer unit 100 includes a first grouping 180 of a stream preheater 110 and a catalytic reactor 130, followed by a second grouping 182 of another stream preheater 110 and a catalytic reactor 130. It is contemplated that additional groupings of stream preheaters 110 and catalytic reactors 130 can be included in the continuous catalytic reformer unit 110, such that three, four, five, or even more stream preheaters 110 and catalytic reactors 130 are present (but Figure 1 only two groupings are shown).

[0018] As used in the present disclosure, a "reactor," such as a catalytic reactor 100, refers to a vessel in which one or more chemical reactions can occur between one or more reactants, optionally in the presence of one or more catalysts. For example, a reactor can include a tank or tube reactor configured to operate as a batch reactor, a continuous stirred tank reactor (CSTR), or a plug flow reactor. Example reactors include packed bed reactors, such as fixed bed reactors and fluidized bed reactors. One or more "reaction zones" can be provided in a reactor. As used in the present disclosure, a "reaction zone" refers to a region of a reactor in which a particular reaction occurs. For example, a packed bed reactor having multiple catalyst beds can have multiple reaction zones, where each reaction zone is defined by the region of each catalyst bed.

[0019] As used in the present disclosure, a "separation unit", such as separation unit 150, refers to any separation device or system of separation devices that at least partially separates one or more chemical species mixed in a process stream. For example, a separation unit can selectively separate materials of different chemical species, phase, or size from one another, forming one or more chemical components. Examples of separation units include, but are not limited to, distillation columns, flash tanks, knock-out drums, knock-out pots, centrifuges, cyclones, filtration devices, traps, scrubbing columns, expansion devices, membranes, solvent extraction devices, and the like. It should be understood that the separation processes described in the present disclosure can not completely separate all of one chemical component from all of another chemical component. It should be understood that the separation processes described in the present disclosure "at least partially" separate different chemical components from one another, and even if not explicitly stated, it should be understood that the separation can include only a partial separation. As used in the present disclosure, one or more chemical components can be "separated" from a process stream to form a new process stream. Typically, a process stream can enter a separation unit and be split or separated into two or more process streams having a desired composition. Further, in certain separation processes, a "lower boiling point fraction" (sometimes referred to as a "light fraction") and a "higher boiling point fraction" (sometimes referred to as a "heavy fraction") can exit a separation unit, on average, with the contents of the lower boiling point fraction having a lower boiling point than the contents of the higher boiling point fraction. Other streams can be intermediate between the lower boiling point fraction and the higher boiling point fraction, such as an "intermediate boiling point fraction".

[0020] According to embodiments of the present disclosure, Figure 1 The hydrocarbon reactant stream 102 can be passed to the continuous catalytic reformer unit 100. Hydrogen 190 can be mixed with the hydrocarbon reactant stream 102 prior to being passed to the continuous catalytic reformer unit 100. Specifically, the hydrocarbon reactant stream 102 mixed with hydrogen 190 can be passed to the stream preheater 110 of the first grouping 180. The stream preheater 110 can generally heat the contents of the hydrocarbon reactant stream 102. The heating in the stream preheater 110 can be achieved by various methods, such as the combustion of a hydrocarbon fuel, or by heat exchange with other media. For example, a fuel gas can be combusted in the stream preheater 110 to heat the hydrocarbon reactant stream 102.

[0021] The heated hydrocarbon reactant stream 102 can exit the stream preheater 110 as a heated reactant stream 112 and be passed to the catalytic reactor 130. In the catalytic reactor 130, the heated reactant stream 112 is contacted with a catalyst and the heated reactant stream 112 undergoes one or more reactions to form an intermediate catalytic reactor effluent stream 132. As used in the present disclosure, the intermediate catalytic reactor effluent stream 132 can be "separated" into one or more chemical components to form a new process stream. Typically, the intermediate catalytic reactor effluent stream 132 can enter a separation unit and be split or separated into two or more process streams having a desired composition. Further, in certain separation processes, a "lower boiling point fraction" (sometimes referred to as a "light fraction") and a "higher boiling point fraction" (sometimes referred to as a "heavy fraction") can exit a separation unit, on average, with the contents of the lower boiling point fraction having a lower boiling point than the contents of the higher boiling point fraction. Other streams can be intermediate between the lower boiling point fraction and the higher boiling point fraction, such as an "intermediate boiling point fraction". Figure 1As shown, the intermediate catalytic reactor effluent 132 can be transferred to another stream preheater 110, and subsequently to another catalytic reactor 130 (shown as the stream preheater 110 and catalytic reactor 130 of the second group 182). That is, the intermediate catalytic reactor effluent 132 can be heated in the stream preheater 110 to form a heated catalytic reactor effluent 114, and the heated catalytic reactor effluent 114 can be transferred to the catalytic reactor 130 and react with the catalyst in the catalytic reactor 130 of the second group 182 to form the final catalytic reactor effluent 134.

[0022] Still refer to Figure 1 The catalyst is typically passed through one or more catalytic reactors 130, where it comes into contact with feed materials (such as heated reaction stream 112 or heated catalytic reactor effluent 114). In the embodiments described herein, the catalytic reforming unit is a “continuous” catalytic reforming unit 100, meaning that one or more catalytic reactors 130 are moving bed reactors rather than fixed bed reactors, and the catalyst can be continuously added and removed. Figure 1 As shown, the catalyst can be circulated between the catalytic reactor 130 and the catalyst regenerator 170 via catalyst streams 172, 174, and 176. The catalyst regenerator 170 can remove coke that may form during the reaction in the catalytic reactor 130 from the catalyst. Furthermore, the catalyst can be treated with chloride via stream 192 to rechlorinate it, whereby the catalyst may lose chloride over time. Although Figure 1 The scheme depicted is one in which the catalyst is circulated between two catalytic reactors 130 before returning to the catalyst regenerator 170, but other configurations are also considered, such as the catalyst being circulated through only a single catalytic reactor 130 before returning to the catalyst regenerator 170.

[0023] According to embodiments, reactions occurring in one or more catalytic reactors 130 can include dehydrogenation of naphthenes, isomerization of naphthenes, and / or dehydrocyclization of paraffins. Dehydrogenation of naphthenes can generally refer to a fast and very endothermic reaction. The reaction can be facilitated by the function of metal catalysts and is more likely to occur under high temperature and low pressure conditions. Naphthenes are generally the most desirable feedstock components because they not only facilitate the reaction, but also generate aromatics and produce hydrogen as a byproduct. Isomerization of naphthenes generally includes isomerization involving rearrangement of the ring, and the possibility of ring opening to form paraffins is relatively high. Paraffin dehydrocyclization can generally refer to a very endothermic reaction, and because of its relatively low rate, to facilitate the reaction, operating conditions tend to be more severe, resulting in increased coke make. As the molecular weight of paraffins increases, the paraffin cyclization step can become easier. Dehydrocyclization reactions can be more likely to occur under low pressure and high temperature conditions, and generally require the function of both metal catalysts and acid catalysts to facilitate the reaction.

[0024] According to one or more embodiments, the final catalytic reactor effluent stream 134 can be passed to a separation unit 150, where the contents of the final catalytic reactor effluent stream 134 are separated into a plurality of product effluent streams 160. The product effluent streams 160 can include offgas 152, a hydrogen-rich stream 154, an LPG stream 156, and a reformate stream 158. The various product effluent streams 160 can be passed to downstream processing units and further processed, or can be pooled as final products for sale and distribution. For example, offgas 152 can be passed to a gas network as fuel gas for heat input, hydrogen-rich stream 154 can be used for pressure swing adsorption (“PSA”) processing, isomerization processing, and / or naphtha hydroprocessing (“NHT”), LNG 156 can be passed to a liquefied natural gas (“LNG”) storage, and / or reformate stream 158 can be passed to a gasoline blending pool.

[0025] Having described the chemical processing section 104 in detail, it should be understood that, during operation, the continuous catalytic reformer unit 100 can include a number of "state variables" and a number of "control variables." As described herein, a control variable refers to a variable that can be directly set in the continuous catalytic reformer unit 100. On the other hand, a state variable refers to a variable that cannot be directly set in the continuous catalytic reformer unit 100. For example, an operator of the continuous catalytic reformer unit 100 can directly set a control variable, such as the amount of catalyst passed to a particular catalytic reactor 130. However, an operator of the continuous catalytic reformer unit 100 cannot directly set a variable such as the pressure within a particular catalytic reactor 130, and thus this variable is a state variable because the system cannot control the pressure within the catalytic reactor 130 by direct input. In other words, control variables can be understood as system inputs that can be directly selected by a plant operator or hydrocarbon reformer process control unit 200; whereas state variables are functions of the various control variables of the continuous catalytic reformer unit 100 and are indirectly regulated by adjusting the control variables.

[0026] Without limitation, state variables considered in the continuous catalytic reformer unit 100 include: catalyst activity, average pressure of catalytic reactors, reactor down time, product yield of reformate, product yield of hydrogen, product yield of off-gas, product yield of LPG, consumption of fuel gas, consumption of cooling water, consumption of electricity, temperature of separation units, or heat exchanger duty of combined feed and effluent.

[0027] Further, without limitation, control variables considered in the continuous catalytic reformer unit 100 include: weighted average inlet temperature of reactant streams, reactant stream feed rates, mass of reactant feed streams (e.g., N+2A feed value, initial boiling point, and final boiling point), chloride injection rate, average carbon burn rate in the catalyst regenerator 170, catalyst circulation rate, hydrogen / hydrocarbon ratio, and energy consumption in the stream preheater 110.

[0028] According to embodiments, both state variables and control variables can be monitored by one or more control variable monitoring hardware 232 and / or one or more state variable monitoring hardware 234, respectively. Further, control variables can be controlled by one or more control variable actuator hardware 236. Figure 1 The exemplary locations of the control variable monitoring hardware 232, the state variable monitoring hardware 234, and the control variable actuator hardware 236 are shown by way of example only, and these various sensors and actuators can be located in other locations of the continuous catalytic reformer unit 100, consistent with the description herein. None of the specifically depicted control variable monitoring hardware 232, state variable monitoring hardware 234, and control variable actuator hardware 236 should be deemed critical to the continuous catalytic reformer unit 100.

[0029] According to particular embodiments, generally, the control variable monitoring hardware 232 and / or the one or more state variable monitoring hardware 234 can be sensors specific to the type of variable being monitored (i.e., measured). For example, a control variable monitoring hardware 232 and / or one or more state variable monitoring hardware 234 related to pressure can be a pressure gauge, and a control variable monitoring hardware 232 or one or more state variable monitoring hardware 234 related to temperature can be a thermometer. Likewise, the control variable actuator hardware can be any device specific to the type of variable being controlled. For example, a control variable monitoring hardware 236 that controls flow rate can be an actuator on a valve that controls flow rate. In some embodiments, the control variable monitoring hardware 232 can be integrated with the control variable monitoring hardware 236, such as for the control variable of flow rate, a single integrated valve can both monitor and control flow rate, and simultaneously act as both the control variable monitoring hardware 232 and the control variable actuator hardware 236. The particular equipment suitable for the control variable monitoring hardware 232, the state variable monitoring hardware 234, and / or the control variable monitoring hardware 236 can be identified and selected by one of skill in the art according to the particular control variable or state variable or its operation.

[0030] Still referring to Figure 1 , according to embodiments described herein, the continuous catalytic reformer unit 100 further comprises a hydrocarbon reformer process control system 200. The hydrocarbon reformer process control system 200 can generally interact with the chemical processing section 104, and in certain embodiments can also directly control the chemical processing section 104. In one or more embodiments, the hydrocarbon reformer process control system 200 can comprise a hydrocarbon reformer variable data store 210, which can comprise processor-executable instructions, and a predictive hydrocarbon reformer modeling processor 220, which can be configured to execute the processor-executable instructions. The hydrocarbon reformer process control system 200 can also include a hydrocarbon reformer output conversion module 230. The hydrocarbon reformer variable data store 210, the predictive hydrocarbon reformer modeling processor 220, and the hydrocarbon reformer output conversion module 230 can remain in communication as shown in Figure 1 , where at least the hydrocarbon reformer variable data store 210 is in direct communication with the hydrocarbon reformer process control system 200, and the predictive hydrocarbon reformer modeling processor 220 is in direct communication with the hydrocarbon reformer output conversion module 230.

[0031] As described herein, a “current” control variable or “current” state variable refers to a variable that is currently present in the continuous catalytic reformer unit 100, as opposed to “historical” state or control variable data, i.e., past data related to control variables and / or state variables. Generally, historical state or control variable data can be collected over a significant period of time, such that the data is suitable for use in the machine learning models described herein.

[0032] According to the implementation scheme, the control variable monitoring hardware 232 can monitor any control variable of the chemical processing section 104, and the data collected from the control variable monitoring hardware 232 can be transmitted to the process control unit 200, thereby enabling the process control unit 200 to receive one or more signals indicating one or more current state variables in real time or near real time. In these implementation schemes, such as Figure 1 As shown, one or more control variable monitoring hardware 232 can maintain communication with the hydrocarbon reformer variable data storage 210 via a transmission module, for example, via a receiver module 272, which maintains communication with the hydrocarbon reformer variable data storage 210. Similarly, state variable monitoring hardware 234 can monitor any control variables of the chemical processing section 104, and this data collected from the control variable monitoring hardware 232 can be transmitted to the process control unit 200, thereby allowing the process control unit 200 to receive one or more signals indicating one or more current control variables in real time or near real time. In these embodiments, such as Figure 1 As shown, one or more state variable monitoring hardware 234 may communicate with the hydrocarbon reformer variable data storage 210 via a transmission module, for example, via a receiver module 274, which communicates with the hydrocarbon reformer variable data storage 210. As described herein, for example, control variable monitoring hardware 232 and / or state variable monitoring hardware 234 may be thermometers, pressure gauges, valves, etc.

[0033] According to one or more embodiments, the hydrocarbon reformer variable data storage 210 stores data related to one or more of the machine learning model and / or the current state variable and the current control variable. The hydrocarbon reformer variable data storage 210 can be configured with any conventional or undeveloped structure for storing sensor data, such as random access memory (RAM), read-only memory (ROM), data registers, databases, and / or other hardware for storing sensor data. The predictive hydrocarbon reformer modeling processor 220 represents hardware and software suitable for performing operations on data obtained at least from the hydrocarbon reformer variable data storage 210 of the process control unit 200. More specifically, in one or more embodiments, the predictive hydrocarbon reformer modeling processor 220 may include specific software-based logic modules, such as data acquisition logic, modeling logic, etc., for generating one or more improved control variables in real time or near real time based on inputs of the current state variable, the current control variable, and selected performance variables. For example, the predictive hydrocarbon reformer modeling processor 220 can receive one or more current state variables and one or more current control variables from the hydrocarbon reformer variable data storage 210, and receive selected performance variables from the hydrocarbon reformer output conversion module 230, and can calculate a model to determine improved control variables, as described below.

[0034] Still referring to Figure 2 The hydrocarbon reformer process control system 200 can include a hydrocarbon reformer output conversion module 230. The hydrocarbon reformer output conversion module 230 can comprise any hardware configured to convert the output of the predictive hydrocarbon reformer modeling processor 220 into a form that can be used to control technical operations associated with the chemical processing section 104, for example, it can comprise a hardware driver or controller, a control data transmitter, a document printer, a data display, or any other hardware capable of producing an operational output and that can be used in the chemical processing section 104 to change, enhance, or otherwise control technical operations or produce technical effects within the continuous catalytic reformer unit 100.

[0035] In embodiments, the hydrocarbon reformer output conversion module 230 can be in communication with the predictive hydrocarbon reformer modeling processor 220 to receive at least the improved control variable. Further, the hydrocarbon reformer output conversion module 230 can be in communication with one or more control variable actuator hardware 236 of the continuous catalytic reformer unit 100. The control variable actuator hardware 236 can be operable to change the current control variable. For example, the control variable actuator hardware 236 can comprise a controller or valve actuator capable of changing a control variable in the continuous catalytic reformer unit 100. For example, in some embodiments, the hydrocarbon reformer output conversion module 230 can send a message to the one or more control variable actuator hardware 236 to make an adjustment to the continuous catalytic reformer unit 100. The hydrocarbon reformer output conversion module 230 can also provide an interface for a plant operator to input a selected control variable to generate an improved control variable. The hydrocarbon reformer output conversion module 230 can also provide an interface for a plant operator to view information about the suggested improved control variable, machine learning model, etc., and then instruct the hydrocarbon reformer output conversion module 230 to send a message to the one or more control variable actuator hardware 236 to adjust the control variable of the continuous catalytic reformer unit 100.

[0036] According to embodiments, the method for operating the continuous catalytic reformer unit 100 can also include adjusting one or more current control variables of the continuous catalytic reformer unit according to the one or more improved control variables determined by the machine learning model. In some embodiments, the adjustment can be performed automatically by the process control unit 200 by causing the hydrocarbon reformer output conversion module 230 to send a signal to the control variable actuator hardware 236 to adjust the control variable. For example, the process control unit 200 can automatically instruct the control variable actuator hardware 236 to change the current control variable. In other embodiments, the hydrocarbon reformer output conversion module 230 can display the improved control variable to a plant operator, and the plant operator can adjust one or more current control variable units based on the improved performance variable after considering the change.

[0037] Now, the hydrocarbon reformer process control system 200 operating with a machine learning model is described in more detail, according to one or more embodiments. In some cases, this description can be limited to selected control variables, selected state variables, and selected performance variables. However, this description should be relevant to the expanded understanding of the presently described subject matter, e.g., embodiments involving different selected control variables, selected state variables, and selected performance variables.

[0038] In some embodiments, the machine learning model is trained by utilizing inputs of at least historical state variable data, historical control variable data, and historical performance variable data. As described herein, the historical performance variable data can include any existing data related to performance variables. As described herein, performance variables refer to variables of the chemical processing section 104 related to system performance. For example, one or more performance variables can be selected from product yield of reformate, product yield of hydrogen, product yield of off-gas, product yield of LPG, reactant stream input rate, chloride injection rate, or time of action of catalyst in the system. It should be understood that performance variables can be state variables, but typically, the machine learning model can use performance variables different from state variables as inputs and different from state variables as outputs. The machine learning model can be trained from past data collected from the chemical processing section 104 over a previous period of time. Such data can be collected by the control variable monitoring hardware 232 and / or the state variable monitoring hardware 234.

[0039] Now referring to Figure 2 Without limitation, the machine learning model can utilize a neural network model. Another machine learning model that can be considered for adoption is a random forest classifier model. As Figure 2 In the neural network model, as shown in FIG. 3, state variable data and control variable data can be used as inputs, and performance variable data can be used as outputs. Historical values of these data can be used to train the neural network model, assigning each node an improved value consistent with data regression to achieve the highest accuracy of the model for predictive use based on historical data.

[0040] The neural network model generally includes one or more layers with one or more nodes connected by node connections. The one or more layers can include an input layer 405, one or more hidden layers, and an output layer. The neural network model can be a deep neural network, a convolutional neural network, or other type of neural network. The neural network model can include one or more convolutional layers and one or more fully connected layers. The input layer represents raw information input into the neural network model. In embodiments, state variable data and control variable data can be input into the neural network model at the input layer, and the output layer can be a performance variable. For example, during a learning process, historical state variable data and historical control variable data can be input into the neural network model at the input layer, while historical performance variable data is at the output layer. In a training mode, these neural network models can employ one or more feedback or backpropagation techniques to train the neural network pathways.

[0041] The neural network model processes the raw information received at the input layer through the nodes and node connections. The one or more hidden layers perform computational activities based on the input from the input layer and the weights on the node connections. In other words, the hidden layers perform computations through their associated nodes and node connections and pass information from the input layer to the output layer.

[0042] Generally, when a neural network model is learning, the neural network model is identifying and determining patterns in the raw information received at the input layer (i.e., historical control variable data and historical state variable data). In response, one or more parameters, such as weights associated with the node connections between nodes, can be adjusted through a process known as backpropagation. It should be appreciated that learning can occur in various processes, however, two prevalent learning processes include associative mapping and regularity detection. Associative mapping refers to a learning process in which the neural network model learns to produce a particular pattern on an input set whenever another particular pattern is applied on the input set. Regularity detection refers to a learning process in which the neural network learns to respond to a particular attribute of an input pattern. In associative mapping, the neural network stores relationships between patterns, while in regularity detection, the response of each unit has a particular “meaning.” Such learning mechanisms can be used for feature discovery and knowledge representation.

[0043] The neural network possesses knowledge contained in the values of the node connection weights. Modifying the knowledge stored in the network according to experience means that a set of learning rules for changing the values of the weights is needed. Information is stored in the weight matrix of the neural network. Learning is the process of determining the weights.

[0044] To train a neural network model to perform a task, the weights are adjusted in a way that reduces the error between the desired output and the actual output. This process can require the neural network model to calculate the error derivative of the weights. In other words, it must calculate how the error changes when each weight is increased or decreased slightly. The backpropagation algorithm is one method used to determine the weights.

[0045] Data related to the trained machine learning model (e.g., node weights) can be stored in the hydrocarbon reformer variable data store 210 and utilized by the predictive hydrocarbon reformer modeling processor 220 to generate improved control variable outputs. As described herein, an improved control variable refers to a control variable that is improved compared to the current control variable based on information generated by the machine learning model described herein. Such improved control variable outputs can be used to adjust the settings of the chemical processing section 104. It should be understood that the improved control variable need not be completely optimized, but only need to be improved compared to the current control variable.

[0046] According to embodiments, current state variables (received from the control variable monitoring hardware 232 to the hydrocarbon reformer variable data store 210) and / or current control variables (received from the state variable monitoring hardware 234 to the hydrocarbon reformer variable data store 210) can be transmitted to the predictive hydrocarbon reformer modeling processor 220. In some embodiments, the predictive hydrocarbon reformer modeling processor 220 can utilize the trained machine learning model to generate a performance variable as a function of the current state and current control variables, as shown in Figure 3 In other embodiments, the predictive hydrocarbon reformer modeling processor 220 can retrieve data for the machine learning model for specific control and state variables from the hydrocarbon reformer variable data store 210 (e.g., rather than computing the model in real-time, relying on past computations of such model).

[0047] Referring now to Figure 3 , a simplified schematic diagram showing one example of a model formed by utilizing a machine learning model is shown. As shown, a performance variable can be calculated based on inputs of current state variables and current control variables. The performance variable model (e.g., a hemisphere as shown in Figure 3 ) is a function of the current control variables and the current state variables. Based on the results of the model, the process control unit 200 can determine and generate an improved control variable based at least on the current state variables and current control variables of the continuous catalytic reformer unit 100, as shown in Figure 3 For example, input to the machine learning model is the current control variables and the current state variables, which are used to calculate a model of a performance variable. The model can then be utilized to determine and generate an improved control variable. For example, in Figure 3In some embodiments, c* can depict a current control variable, and the model can be used to generate an improved control variable along the dashed line that improves performance. The improved control variable need not be optimal, but only need to improve performance by any degree as predicted by the machine learning model.

[0048] As described herein, the machine learning model can be trained by utilizing inputs containing historical state variables and historical control variables, while the predictive modeling of the process control unit 200 can generate an improved control variable based on both current control variables and current state variables. It has been found that, in at least some embodiments, utilizing a machine learning model that uses at least state variables and control variables can enable more accurate modeling of an improved control variable than using only control variables or only state variables.

[0049] In some embodiments, the control operations of the process control unit 200 on the chemical processing section 104 can be in real-time or near real-time. That is, for example, the process control system receives one or more signals indicative of one or more current state variables in real-time or near real-time, the process control system receives one or more signals indicative of one or more current control variables in real-time or near real-time, and / or the process control system generates one or more improved control variables in real-time or near real-time. As described herein, real-time refers to the calculations and / or execution operations performed by the process control unit 200 on the chemical processing section 104 when the process control unit 200 receives inputs such as current control variable data and current state variable data. Near real-time can refer to a longer delay, such as a delay of seconds, minutes, or even hours, between the calculations and / or execution operations performed by the process control unit 200 on the chemical processing section 104. In some embodiments, the control and adjustment of the continuous catalytic reformer unit 100 by the one or more control variable actuator hardware 236 can be performed in real-time. In other embodiments, the control and / or adjustment of the control variable actuator hardware 236 can be implemented by a plant operator who has reviewed data displayed on the hydrocarbon reformer output conversion module 230 by input.

[0050] In some embodiments, multiple machine learning models utilizing different state variables and control variables are combined to determine an improved control variable. In other embodiments, multiple control variables and / or multiple state variables are used in the same machine learning model.

[0051] Many non-limiting technical aspects are presently described herein, listed below as aspects 1-15.

[0052] Aspect 1, a method for operating a continuous catalytic reformer unit, the method comprising: passing a hydrocarbon reactant stream to a continuous catalytic reformer unit to form one or more product effluent streams, the continuous catalytic reformer unit comprising at least one stream preheater, at least one catalytic reactor, and at least one separation unit; implementing a hydrocarbon reformer process control system, the hydrocarbon reformer process control system comprising a hydrocarbon reformer variable data store, a hydrocarbon reformer output conversion module, and one or more predictive hydrocarbon reformer modeling processors, the hydrocarbon reformer variable data store comprising processor-executable instructions, the predictive hydrocarbon reformer modeling processors configured to execute the processor-executable instructions and cause the process control system to: receive: one or more signals from one or more state variable actuator hardware, the signals indicative of one or more current state variables, wherein a current state variable is a process variable that cannot be directly set in the continuous catalytic reformer unit; and one or more signals from one or more control variable actuator hardware, the signals indicative of one or more current control variables of the continuous catalytic reformer unit, wherein a current control variable is a process variable that can be directly set in the continuous catalytic reformer unit; and generate an improved control variable by utilizing a machine learning model, the improved control variable increasing a selected performance variable based on input of one or both of the one or more current state variables or the one or more current control variables, wherein the improved control variable in the machine learning model is trained with input of at least historical state variable data, historical control variable data, and historical performance variable data; and adjust the one or more current control variables of the continuous catalytic reformer unit based on the improved control variable determined by the machine learning model.

[0053] Aspect 2, the method of aspect 1, wherein the hydrocarbon reformer output conversion module sends one or more signals to the one or more control variable actuator hardware to adjust the one or more current control variables.

[0054] Aspect 3, the method of any of the preceding aspects, wherein the process control system automatically adjusts the one or more current control variables in real-time based on the one or more improved control variables determined by the machine learning model.

[0055] Aspect 4, the method of any of the preceding aspects, wherein the process control system receives the one or more signals indicative of the one or more current state variables in real-time or near real-time; the process control system receives the one or more signals indicative of the one or more current control variables in real-time or near real-time.

[0056] Aspect 5, the method of any of the preceding aspects, wherein the process control system generates the one or more improved control variables in real-time or near real-time.

[0057] Aspect 6. The method of any preceding aspect, wherein the one or more state variables are selected from the group consisting of reactant stream input rate, catalyst activity, average pressure of the catalytic reactor, reactor down time, product yield of reformate, product yield of hydrogen, product yield of off-gas, product yield of LPG, consumption of fuel gas, consumption of cooling water, consumption of electricity, temperature of separation unit, heat exchanger load of combined feed and effluent, or catalyst on-stream time in the system.

[0058] Aspect 7. The method of any preceding aspect, wherein the one or more control variables are selected from the group consisting of weighted average inlet temperature of the reactant stream, reactant stream feed rate, mass of reactant feed stream, chloride injection rate, average carbon burn rate in the catalyst regenerator, catalyst circulation rate, hydrogen / hydrocarbon ratio, and energy consumption in the stream preheater.

[0059] Aspect 8. The method of any preceding aspect, wherein the selected performance variable is selected from the group consisting of product yield of reformate, product yield of hydrogen, product yield of off-gas, product yield of LPG, reactant stream input rate, chloride injection rate, or catalyst on-stream time in the system.

[0060] Aspect 9. The method of aspect 8, wherein the selected performance variable is product yield of reformate.

[0061] Aspect 10. The method of aspect 8, wherein the selected performance variable is product yield of hydrogen.

[0062] Aspect 11. The method of aspect 8, wherein the selected performance variable is selected from the group consisting of product yield of off-gas or product yield of LPG.

[0063] Aspect 12. The method of aspect 8, wherein the selected performance variable is selected from the group consisting of reactant stream input rate, chloride injection rate, or catalyst on-stream time in the system.

[0064] Aspect 13. The method of any preceding aspect, wherein the reactant stream comprises refinery naphtha. Aspect 14. The method of any preceding aspect, wherein the machine learning model comprises a neural network model. Aspect 15. A continuous catalytic reformer unit operated using the method of any preceding aspect.

[0065] The subject matter of the present disclosure has been described with specificity to draw clear and distinct lines among the claims and the subject matter disclosed herein. It should be understood that any detailed description of components or features that is expressly included in one aspect or embodiment is equally applicable to any other aspect or embodiment, even if it is not repeated in the detailed description of the other aspects or embodiments. Furthermore, it should be understood that the various embodiments described herein can be combined in any way.

Claims

1. A method for operating a continuous catalytic reformer unit, the method comprising: The hydrocarbon reaction stream is passed to a continuous catalytic reformer unit to form one or more product effluent streams, the continuous catalytic reformer unit comprising at least one stream preheater, at least one catalytic reactor and at least one separation unit. A hydrocarbon reformer process control system is implemented, comprising a hydrocarbon reformer variable data storage, a hydrocarbon reformer output conversion module, and one or more predictive hydrocarbon reformer modeling processors. The hydrocarbon reformer variable data storage contains processor-executable instructions, and the predictive hydrocarbon reformer modeling processors are configured to execute the processor-executable instructions and cause the process control system to: take over: One or more signals from one or more state variable actuator hardware, the signals indicating one or more current state variables, wherein the current state variables are process variables that cannot be directly set in the continuous catalytic reformer unit; and One or more signals from one or more control variable actuator hardware, the signals indicating one or more current control variables of the continuous catalytic reformer unit, wherein the current control variables are process variables that can be directly set in the continuous catalytic reformer unit; as well as Improved control variables are generated by utilizing a machine learning model, which are based on inputs to one or more current state variables or one or two of the current control variables, and add selected performance variables, wherein the improved control variables in the machine learning model are trained using inputs of at least historical state variable data, historical control variable data, and historical performance variable data. as well as Based on the improved control variables determined by the machine learning model, one or more current control variables of the continuous catalytic reformer unit are adjusted.

2. The method according to claim 1, wherein the hydrocarbon reformer output conversion module sends one or more signals to one or more control variable actuator hardware to adjust the one or more current control variables.

3. The method according to any one of the preceding claims, wherein the process control system automatically adjusts the one or more current control variables in real time based on the one or more improved control variables determined by the machine learning model.

4. The method according to any one of the preceding claims, wherein: The process control system receives, in real-time or near real-time, the one or more signals indicating one or more current state variables; and The process control system receives, in real time or near real time, one or more signals indicating one or more current control variables.

5. The method according to any one of the preceding claims, wherein the process control system generates the one or more improved control variables in real time or near real time.

6. The method according to any one of the preceding claims, wherein the one or more state variables are selected from the following: reactant input rate, catalyst activity, average pressure of the catalytic reactor, reactor downtime, product yield of reformate, product yield of hydrogen, product yield of waste gas, product yield of LPG, fuel consumption, cooling water consumption, power consumption, temperature of the separation unit, heat exchanger load of combined feed and effluent, or catalyst action time in the system.

7. The method according to any one of the preceding claims, wherein the one or more control variables are selected from the weighted average inlet temperature of the reaction stream, the reaction stream feed rate, the mass of the reactant feed stream, the chloride injection rate, the average carbon combustion rate in the catalyst regenerator, the catalyst circulation rate, the hydrogen / hydrocarbon ratio, and the energy consumption in the stream preheater.

8. The method according to any one of the preceding claims, wherein the selected performance variable is selected from the product yield of the reformate, the product yield of hydrogen, the product yield of the exhaust gas, the product yield of LPG, the reaction stream input rate, the chloride injection rate, or the reaction time of the catalyst in the system.

9. The method of claim 8, wherein the selected performance variable is the product yield of the reformate.

10. The method of claim 8, wherein the selected performance variable is the product yield of the hydrogen.

11. The method of claim 8, wherein the selected performance variable is selected from the product yield of exhaust gas or the product yield of LPG.

12. The method of claim 8, wherein the selected performance variable is selected from the reactant stream input rate, chloride injection rate, or the time the catalyst has been in the system.

13. The method according to any one of the preceding claims, wherein the reaction stream comprises refinery naphtha.

14. The method according to any one of the preceding claims, wherein the machine learning model comprises a neural network model.

15. A continuous catalytic reformer unit, which operates using the method described in any one of the preceding claims.

Images