Method and device for monitoring operating characteristics of an industrial gas equipment complex

Abstract

A method for monitoring the operating characteristics of an industrial gas equipment complex including a plurality of industrial gas equipment is provided. The method is executed by at least one hardware processor and includes: assigning a machine learning model to each of the industrial gas equipment forming the industrial gas equipment complex; training a corresponding machine learning model for each industrial gas equipment based on the received historical time-related operating characteristic data of the corresponding industrial gas equipment; executing the trained machine learning model for each industrial gas equipment to predict the operating characteristics of each corresponding industrial gas equipment within a predetermined future time period; and comparing the predicted operating characteristic data of each corresponding industrial gas equipment within the predetermined future time period with the measured operating characteristic data within the corresponding time period to identify deviations in the performance of the industrial gas equipment.

Description

Technical Field

[0001] The present invention relates to a method and system for monitoring the operating characteristics of an industrial gas equipment complex including a plurality of industrial gas equipment. Background Art

[0002] An industrial gas equipment complex may include one or more process equipment that produce gases or are involved in the production of gases. In non-limiting examples, these gases may include: industrial gases, commercial gases, medical gases, inorganic gases, organic gases, fuel gases, and green fuel gases in gaseous, liquefied, or compressed form.

[0003] An exemplary industrial gas is ammonia. Ammonia is produced using hydrogen from water electrolysis and nitrogen separated from air. These gases are then fed into the Haber-Bosch process, where hydrogen and nitrogen react together at high temperature and high pressure to produce ammonia.

[0004] Generally, regardless of the type or configuration of the industrial gas equipment complex, it is desirable for such a complex to operate in a substantially stable state. An unexpected failure, unplanned maintenance, or efficiency loss in any equipment that forms part of the equipment complex can lead to undesirable downtime, significantly reduced production rates, and poor power utilization. Summary of the Invention

[0005] Some concepts are introduced below in simplified form to provide a basic understanding of some aspects of the present disclosure. The following is not an extensive overview of the present disclosure and is not intended to identify key or important elements of the present disclosure or to depict the scope of the present disclosure. The following only outlines some concepts of the present disclosure as a prelude to the more detailed description provided hereafter.

[0006] According to a first aspect, there is provided a method for determining and utilizing predicted available power resources from one or more renewable power sources for one or more industrial gas equipment including one or more storage resources, the method being executed by at least one hardware processor, the method comprising: obtaining historical time-related environmental data associated with one or more renewable power sources; obtaining historical time-related operating characteristic data associated with the one or more renewable power sources; training a machine learning model based on the historical time-related environmental data and the historical time-related operating characteristic data; executing the trained machine learning model to predict the available power resources of the one or more industrial gas equipment within a predetermined future time period; and controlling the one or more industrial gas equipment in response to the predicted available power resources for the predetermined future time period.

[0007] In an embodiment, controlling the one or more industrial gas equipment includes maximizing the use of the predicted available power resources for the predetermined future time period.

[0008] In an embodiment, the storage resources include one or more industrial gas storage containers and / or one or more energy storage resources.

[0009] In an embodiment, one or more energy storage resources include one or more of the following: a battery energy storage system; a compressed air energy storage; a liquid air energy storage; or a pumped hydro energy storage.

[0010] In an embodiment, maximizing the use of the predicted power resources further includes controlling the utilization rate of the industrial gas storage containers and / or one or more energy storage resources in response to the predicted available power resources.

[0011] In an embodiment, controlling the utilization rate includes using an algorithm to select one or more storage resources from a group of storage resources for a given pattern of predicted power availability as a function of time.

[0012] In an embodiment, the selection of the storage resources is based on the physical characteristics of the storage resources.

[0013] In an embodiment, one or more renewable power sources include one or more of the following: a solar power source; a wind power source; a tidal power source; a hydro power source; or a geothermal power source.

[0014] In an embodiment, the environmental data is selected from one or more of the following: wind speed; cloud cover; precipitation; humidity; air temperature; atmospheric pressure; solar intensity; and tidal time.

[0015] In an embodiment, the operating characteristic data includes the power output from one or more renewable power sources.

[0016] In an embodiment, the step of training the machine learning model is periodically performed at a predetermined training time.

[0017] In an embodiment, at the training time, the machine learning model is trained based on historical time-related environmental data and historical time-related operating characteristic data obtained within one or more predetermined historical time windows.

[0018] In an embodiment, the method further includes comparing the value of the predicted power resources for a predetermined future time period with the actual power resources at the end of the prediction time period to generate a prediction error value.

[0019] In an embodiment, when the prediction error value exceeds a predetermined threshold, a predetermined training time is selected.

[0020] In an embodiment, the predetermined training time is selected based on a predetermined empirical interval, unless the prediction error value exceeds the predetermined threshold within the predetermined empirical interval.

[0021] In an embodiment, one or more industrial gas devices include a hydrogen production device having at least one electrolytic cell.

[0022] In an embodiment, one or more industrial gas devices include an ammonia production device complex that includes a hydrogen production device.

[0023] In an embodiment, the machine learning model includes one or more of the following: a gradient boosting algorithm; a long short-term memory (LSTM) algorithm; a support vector machine (SVM) algorithm; or a random decision forest algorithm.

[0024] According to a second aspect, there is provided a system for determining and utilizing predicted available power resources from one or more renewable power sources for one or more industrial gas devices including one or more storage resources, the system including: at least one hardware processor operable to perform: obtaining historical time-related environmental data associated with one or more renewable power sources; obtaining historical time-related operating characteristic data associated with the one or more renewable power sources; training a machine learning model based on the historical time-related environmental data and the historical time-related operating characteristic data; executing the trained machine learning model to predict the available power resources of the one or more industrial gas devices within a predetermined future time period; and controlling the one or more industrial gas devices in response to the predicted available power resources for the predetermined future time period.

[0025] According to a third aspect, there is provided a computer-readable storage medium storing a program of instructions executable by a machine to perform a method for determining and utilizing predicted available power resources from one or more renewable power sources for one or more industrial gas devices including one or more storage resources, the method including: obtaining historical time-related environmental data associated with one or more renewable power sources; obtaining historical time-related operating characteristic data associated with the one or more renewable power sources; training a machine learning model based on the historical time-related environmental data and the historical time-related operating characteristic data; executing the trained machine learning model to predict the available power resources of the industrial gas devices within a predetermined future time period; and controlling the one or more industrial gas devices in response to the predicted available power resources for the predetermined future time period.

[0026] According to a fourth aspect, a method for monitoring operating characteristics of an industrial gas equipment complex including a plurality of industrial gas equipment is provided. The method is executed by at least one hardware processor and includes: assigning a machine learning model to each industrial gas equipment forming the industrial gas equipment complex; training the corresponding machine learning model for each industrial gas equipment based on the received historical time-related operating characteristic data of the corresponding industrial gas equipment; executing the trained machine learning model for each industrial gas equipment to predict the operating characteristics of each corresponding industrial gas equipment within a predetermined future time period; and comparing the predicted operating characteristic data of each corresponding industrial gas equipment within the predetermined future time period with the measured operating characteristic data within the corresponding time period to identify deviations in the performance of the industrial gas equipment.

[0027] In an embodiment, the comparing step is performed at the end of the predetermined future time period of the predicted operating characteristic data or at a timestamp therein.

[0028] In an embodiment, the comparing step includes comparing the predicted operating characteristic data predicted for a predetermined time window with the actually measured operating characteristic data for the same time window.

[0029] In an embodiment, the received historical time-related operating characteristic data of the corresponding industrial gas equipment includes data obtained from direct measurements of processes or parameters of the corresponding industrial gas equipment.

[0030] In an embodiment, the received historical time-related operating characteristic data of the corresponding industrial gas equipment includes data obtained from a physics-based model representing the operating characteristics of the corresponding industrial gas equipment.

[0031] In an embodiment, the measurement data related to the processes or parameters of the corresponding industrial gas equipment is input into the corresponding physics-based model.

[0032] In an embodiment, the predicted operating characteristics of each industrial gas equipment are used to determine predicted future resources, future failures, and / or predicted future maintenance.

[0033] In an embodiment, one or more industrial gas equipment include hydrogen processing equipment having a plurality of electrolyzer modules.

[0034] In an embodiment, a machine learning model is assigned to each electrolyzer module.

[0035] In an embodiment, the predicted operating characteristics of each corresponding industrial gas equipment are used in an additional model to generate an operating performance metric for the industrial gas equipment complex.

[0036] In an embodiment, the operating performance metric includes an efficiency value of the industrial gas equipment complex.

[0037] In an embodiment, the industrial gas equipment complex includes an ammonia equipment complex, and the determined efficiency value enables a predicted determination of ammonia produced for a given level of energy input.

[0038] According to a fifth aspect, there is provided a system for monitoring operating characteristics of an industrial gas equipment complex including a plurality of industrial gas equipment, the system including at least one hardware processor operable to perform assigning a machine learning model to each industrial gas equipment forming the industrial gas equipment complex; training a corresponding machine learning model for each industrial gas equipment based on received historical time-related operating characteristic data of the corresponding industrial gas equipment; executing the trained machine learning model for each industrial gas equipment to predict operating characteristics of each corresponding industrial gas equipment within a predetermined future time period; and comparing the predicted operating characteristic data of each corresponding industrial gas equipment within the predetermined future time period with measured operating characteristic data within the corresponding time period to identify deviations in industrial gas equipment performance.

[0039] In an embodiment, the comparing step is performed at the end of the predetermined future time period of the predicted operating characteristic data or at a timestamp therein.

[0040] In an embodiment, the comparing step includes comparing predicted operating characteristic data predicted for a predetermined time window with actual measured operating characteristic data for the same time window.

[0041] In an embodiment, the received historical time-related operating characteristic data of the corresponding industrial gas equipment includes data obtained from direct measurement of processes or parameters of the corresponding industrial gas equipment.

[0042] In an embodiment, the received historical time-related operating characteristic data of the corresponding industrial gas equipment includes data obtained from a physics-based model representing operating characteristics of the corresponding industrial gas equipment.

[0043] In an embodiment, the predicted operating characteristics of each industrial gas equipment are utilized to determine predicted future resources, future faults, and / or predicted future maintenance.

[0044] In an embodiment, the predicted operating characteristics of each corresponding industrial gas equipment are utilized in an additional model to generate an operating performance metric of the industrial gas equipment complex.

[0045] According to a sixth aspect, there is provided a computer-readable storage medium storing a program of instructions executable by a machine to perform a method of monitoring operating characteristics of an industrial gas equipment complex including a plurality of industrial gas devices, the method being executed by at least one hardware processor, the method including: allocating a machine learning model to each industrial gas device forming the industrial gas equipment complex; training a corresponding machine learning model for each industrial gas device based on received historical time-related operating characteristic data of the corresponding industrial gas device; executing the trained machine learning model for each industrial gas device to predict the operating characteristics of each corresponding industrial gas device within a predetermined future time period; and comparing the predicted operating characteristic data of each corresponding industrial gas device within the predetermined future time period with the measured operating characteristic data within the corresponding time period to identify deviations in the performance of the industrial gas device.

[0046] According to a seventh aspect, there is provided a method of controlling an industrial gas equipment complex including a plurality of industrial gas devices powered by one or more renewable power sources, the method being executed by at least one hardware processor, the method including: receiving time-related predicted power data for a predetermined future time period from the one or more renewable power sources; receiving time-related predicted operating characteristic data for each industrial gas device; using the predicted power data and the predicted characteristic data in an optimization model to generate a set of state variables for the plurality of industrial gas devices; using the generated state variables to generate a set of control setpoints for the plurality of industrial gas devices; and sending the control setpoints to a control system to control the industrial gas equipment complex by adjusting one or more control setpoints of the industrial gas devices.

[0047] In an embodiment, the optimization model defines the predicted power data and the predicted characteristic data as a set of non-linear equations.

[0048] In an embodiment, the state variables are generated by solving the set of non-linear equations.

[0049] In an embodiment, the time-related predicted power data is generated from a trained machine learning model.

[0050] In an embodiment, the time-related predicted power data is obtained by: obtaining historical time-related environmental data associated with the one or more renewable power sources; obtaining historical time-related operating characteristic data associated with the one or more renewable power sources; training a machine learning model based on the historical time-related environmental data and the historical time-related operating characteristic data; and executing the trained machine learning model to predict the available power resources for the one or more industrial gas devices within a predetermined future time period.

[0051] In an embodiment, time - related predicted operation characteristic data is generated from a trained machine - learning model of each industrial gas device.

[0052] In an embodiment, time - related predicted operation characteristic data for each industrial device is obtained by: assigning a machine - learning model to each industrial gas device forming an industrial gas device complex; training a corresponding machine - learning model for each industrial gas device based on the received historical time - related operation characteristic data of the corresponding industrial gas device; and executing the trained machine - learning model for each industrial gas device to predict the operation characteristics of each corresponding industrial gas device within a predetermined future time period.

[0053] In an embodiment, the industrial gas device complex includes storage resources, and the storage resources include one or more industrial gas storage containers and / or one or more energy storage resources.

[0054] In an embodiment, one or more energy storage resources include one or more of the following: a battery energy storage system; a compressed air energy storage; a liquid air energy storage; or a pumped - hydro energy storage.

[0055] In an embodiment, the predicted power data further includes data representing operation parameters of the storage resources.

[0056] In an embodiment, the data representing operation parameters of the storage resources includes one or more of the following: resource memory availability; fill level; and utilization rate.

[0057] According to an eighth aspect, there is provided a system for controlling an industrial gas device complex including a plurality of industrial gas devices powered by one or more renewable power sources, the system including: at least one hardware processor operable to execute: receiving time - related predicted power data for a predetermined future time period from the one or more renewable power sources; receiving time - related predicted operation characteristic data of each industrial gas device; using the predicted power data and the predicted characteristic data in an optimization model to generate a set of state variables of the plurality of industrial gas devices; using the generated state variables to generate a set of control set - points for the plurality of industrial gas devices; and sending the control set - points to a control system to control the industrial gas device complex by adjusting one or more control set - points of the industrial gas devices.

[0058] In an embodiment, the optimization model defines the predicted power data and the predicted characteristic data as a set of non - linear equations.

[0059] In an embodiment, the state variables are generated by solving the set of non - linear equations.

[0060] In an embodiment, time - related predicted power data is generated from a trained machine - learning model.

[0061] In an embodiment, time - related predicted operating characteristic data is generated from a trained machine - learning model for each industrial gas device.

[0062] According to a ninth aspect, there is provided a computer - readable storage medium storing a program of instructions, the program of instructions being executable by a machine to perform a method of controlling an industrial gas device complex that includes a plurality of industrial gas devices powered by one or more renewable power sources, the method being executed by at least one hardware processor, the method including: receiving time - related predicted power data for a predetermined future time period from the one or more renewable power sources; receiving time - related predicted operating characteristic data for each industrial gas device; using the predicted power data and the predicted characteristic data in an optimization model to generate a set of state variables for the plurality of industrial gas devices; using the generated state variables to generate a set of control set - points for the plurality of industrial gas devices; and sending the control set - points to a control system to control the industrial gas device complex by adjusting one or more control set - points of the industrial gas devices.

[0063] In an embodiment, the optimization model defines the predicted power data and the predicted characteristic data as a set of non - linear equations.

[0064] In an embodiment, the state variables are generated by solving the set of non - linear equations.

[0065] In an embodiment, the time - related predicted power data is generated from a trained machine - learning model, and / or the time - related predicted operating characteristic data is generated from a trained machine - learning model for each industrial gas device. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Embodiments of the present invention will now be described by way of example only and with reference to the accompanying drawings, in which:

[0067] Figure 1 is a schematic diagram of an industrial gas device complex and a control system;

[0068] Figure 2 is Figure 1 a detailed schematic diagram of the control system of

[0069] Figure 3 is a graph showing measured and predicted wind power values using a model according to an embodiment of the present invention;

[0070] Figure 4 is a graph showing measured and predicted solar power values using a model according to an embodiment of the present invention;

[0071] Figure 5 is a flowchart of a method according to an embodiment;

[0072] Figure 6 is a flowchart of a method according to an embodiment;

[0073] Figure 7 is a diagram showing an efficiency model of a process variable;

[0074] Figure 8 is a graph of the optimal ammonia plant rate as a function of time over a predicted period of 48 hours; and

[0075] Figure 9 is a flowchart of a method according to an embodiment.

[0076] Embodiments of the present disclosure and their advantages can be best understood by referring to the following detailed description. It should be understood that the same reference numerals are used to identify the same elements shown in one or more of the figures, wherein the illustration is for the purpose of illustrating embodiments of the present disclosure and not for the purpose of limiting the embodiments of the present disclosure. Detailed Description

[0077] Various examples and embodiments of the present disclosure will now be described. The following description provides specific details for a thorough understanding and to be able to describe these examples. However, one or more of the embodiments described herein can be practiced without many of these details by those of ordinary skill in the relevant art. Similarly, one or more of the embodiments of the present disclosure may include other features and / or functions not described in detail herein by those of skill in the relevant art. Additionally, some well-known structures or functions may not be shown or described in detail below to avoid unnecessarily obscuring the relevant description.

[0078] Figure 1 shows a schematic diagram of an industrial gas plant complex 10 and a control system 100.

[0079] The industrial gas plant complex 10 includes a hydrogen production plant 12, a hydrogen storage unit 14, an air separation unit (ASU) 16, an ammonia synthesis plant 18, and an ammonia storage unit 20. The ammonia storage unit 20 is connected to an external supply chain 22 for forward distribution of ammonia.

[0080] The power for driving the industrial gas plant complex 10 is at least partially generated by renewable energy sources (such as wind 24 and / or solar 26), although other sources such as diesel, gasoline, or hydrogen-powered generators (not shown) or the national grid (not shown) may optionally be utilized.

[0081] To address the intermittency of power supply from renewable sources, an energy storage resource 28 is provided. The energy storage resource 28 can include one or more resource storage devices or energy storage devices. For example, one or more resource storage devices can include a hydrogen storage unit 14. Producing hydrogen through electrolysis requires a large amount of power, and using the stored hydrogen as a hydrogen source for ammonia production can significantly reduce the power consumption of the industrial gas equipment complex 10 during periods of low renewable power supply. Additionally, a liquid nitrogen storage 16a can also be provided as part of the energy storage resource 28, as Figure 1 shown.

[0082] Additionally or alternatively, in a non-exhaustive arrangement, the energy storage device can include one or more of the following: a battery energy storage system (BESS) 28a, a compressed / liquid air energy system (CAES or LAES) 28b, or a pumped hydro storage system (PHSS) 28c.

[0083] The BESS 28a utilizes electrochemical technology and may include one or more of the following: lithium-ion batteries, lead-acid batteries, zinc bromide, sodium sulfur, or redox flow batteries. Electrochemical devices such as batteries have the advantage in terms of fast charging rates and fast (almost instantaneous) ramp rates to supply power in response to sudden drops in energy supply. However, their power capacity tends to be more limited than other systems. Therefore, they may be more suitable for use in situations where, for example, power shortages from renewable sources are expected to be temporary or of short duration.

[0084] The CAES 28b compresses air and stores the air at a high pressure of about 70 bar. It is typically stored in underground caverns. When power is needed, the compressed air is heated and expanded in an expansion turbine to drive a power generator.

[0085] The LAES 28b includes an air liquefier to draw air from the environment and compress and cool the air to achieve liquefaction. The liquefied air is then stored in an insulated tank until power is needed. To convert the liquefied air into usable energy, the liquid air is pumped to a high pressure and heated through a heat exchanger. The resulting high-pressure gas is used to drive a turbine to generate electricity.

[0086] CAES and LAES are capable of storing significantly more energy than most BESS 28a systems. However, CAES and LAES have a slower ramp rate than electrochemical storage devices and require more time to store larger amounts of energy. For example, the compression stage operating at full load may take about 5 - 10 minutes, and generating full power on demand takes 10 - 20 minutes. Therefore, such storage devices are more suitable for long-term storage and are more suitable for supplying power during long-term shortages of renewable energy.

[0087] The PHSS28c stores energy in the form of gravitational potential energy by pumping water from a lower elevation reservoir to a higher elevation reservoir. When power is needed, the water is released to drive a turbine. Some PHSS devices utilize reversible pump - turbine units.

[0088] Given the large storage capacity of PHSS configurations, they are generally suitable for long - term storage. Additionally, especially for reversible pump turbines, time scales of about 5 - 10 minutes from shutdown to full - load power generation, about 5 - 30 minutes from shutdown to pumping, and about 10 - 40 minutes from pumping to load generation, or vice versa, are common. Thus, this storage seems more suitable for long - term power shortages.

[0089] Although all these elements are shown in Figure 1 this is for illustrative purposes only. The energy storage resource 28 need not include every described element and may include only one or more of the described elements. Additionally, the energy storage resource 28 may include additional elements.

[0090] The components of the industrial gas equipment complex 10 will now be described in detail.

[0091] Hydrogen production equipment 12

[0092] The hydrogen production equipment 12 is operable to electrolyze water to form hydrogen and oxygen. Any suitable water source can be used. However, in embodiments where seawater is used to produce water for electrolysis, the equipment will further include at least one desalination and demineralization device for treating the seawater.

[0093] The hydrogen production equipment 12 includes a plurality of electrolysis units 12a, 12b... 12n or electrolytic cells. Each unit or cell may be referred to as an "electrolyzer" 12a, 12b... 12n.

[0094] The electrolyzers can give the hydrogen production equipment 12 a total capacity of at least 1 GW. However, the ultimate capacity of the hydrogen production equipment 12 is limited only by practical considerations such as power supply.

[0095] Any suitable type of electrolyzer can be used. In embodiments, multiple electrolyzers typically consist of multiple individual cells combined into "modules" that also contain process equipment such as pumps, coolers, and / or separators. Hundreds of cells can be used and grouped in different buildings. Each module typically has a maximum capacity of greater than 10 MW, although this is not intended to be limiting.

[0096] Any suitable type of electrolyzer can be used in the present invention. Generally, three conventional types of electrolyzers are used - alkaline electrolyzers; PEM electrolyzers; and solid oxide electrolyzers. Any of these types can be used in the present invention.

[0097] Alkaline electrolyzers transport hydroxide ions (OH - ) from the cathode to the anode through an electrolyte, where hydrogen gas is generated on the cathode side. Typically, a liquid alkaline solution of sodium hydroxide or potassium hydroxide is used as the electrolyte.

[0098] PEM electrolyzers utilize a solid plastic material as the electrolyte, and water reacts at the anode to form oxygen and positively charged hydrogen ions. Electrons flow through an external circuit, and the hydrogen ions selectively move through the PEM to the cathode. At the cathode, the hydrogen ions combine with electrons from the external circuit to form hydrogen gas.

[0099] Solid oxide electrolyzers use a solid ceramic material as the electrolyte, which selectively conducts negatively charged oxygen ions (O 2- ) at high temperatures. Water at the cathode combines with electrons from the external circuit to form hydrogen gas and negatively charged oxygen ions. The oxygen ions pass through the solid ceramic membrane and react at the anode to form oxygen gas and generate electrons for the external circuit.

[0100] The electrolyzers can be arranged in any suitable group. For example, they can be arranged in parallel.

[0101] Hydrogen gas is produced by the hydrogen production device 12 at approximately atmospheric pressure. The hydrogen gas stream thus generated is removed from the electrolyzer at a slightly elevated pressure and can be transported through a pipeline to the ammonia synthesis device 18.

[0102] Alternatively, any hydrogen gas in excess of demand can be stored in the hydrogen storage unit 14. The hydrogen storage unit 14 includes multiple short-term and long-term storage options with different sizes, filling / discharge rates, and round-trip efficiencies. A typical storage system can include pressure vessels and / or pipeline sections connected to a common inlet / outlet manifold. For example, the pressure vessels can be spheres with a diameter of approximately 25 m, or "projectiles", which are horizontal vessels with a large L / D ratio (usually up to approximately 12:1) and a diameter of up to approximately 12 m. In some regions, underground caverns are included as storage systems to eliminate seasonal variations associated with renewable power.

[0103] Preferably, the hydrogen gas is compressed by a compressor and stored in the hydrogen storage unit 14 under pressure to reduce the volume requirement. At this point, it can be used commercially (e.g., sold for automotive use), or it can be used as a reservoir for the ammonia synthesis device 18 through the pipeline 30.

[0104] Optionally, a purification system may be implemented to purify or dry the hydrogen before continued use. For example, the hydrogen gas may be dried in an adsorption unit, such as a temperature swing adsorption (TSA) unit for downstream processes.

[0105] Air separation unit 16

[0106] In a non-limiting embodiment, nitrogen required for ammonia production is produced by cryogenic distillation of air in an air separation unit (ASU) 16. Typically, the ASU 16 operates at a pressure of about 10 bar. The pressure is then reduced to provide a nitrogen gas stream in one or more pipes arranged to deliver the nitrogen to the ammonia synthesis apparatus 18. However, if desired, other nitrogen sources may be used, such as a liquid nitrogen gas storage 16a.

[0107] A liquid nitrogen gas storage 16a may also be provided, which may be used as a resource storage as described below. The liquid nitrogen gas storage 16a may include, together with the hydrogen storage unit 14, a plurality of short-term and long-term storage options having different sizes, filling / discharge rates, and round-trip efficiencies.

[0108] Typical storage systems for nitrogen may include a plurality of pressure vessels and / or pipe sections connected to a common inlet / outlet manifold. For example, the pressure vessels may be spheres with a diameter of about 25 m, or "projectiles", which are horizontal vessels with a large L / D ratio (usually up to about 12:1) and a diameter of up to about 12 m. In some regions, underground caverns may be used to eliminate seasonal variations associated with renewable power.

[0109] Preferably, the nitrogen is compressed by a compressor and stored under pressure in the liquid nitrogen gas storage 16a to reduce the volume requirement. It may be used as a reservoir for the ammonia synthesis apparatus 18, which may be fed through a connecting pipe.

[0110] Ammonia synthesis apparatus 18

[0111] The ammonia synthesis apparatus 18 operates by the Haber-Bosch process and includes an ammonia loop. The ammonia loop is a single-unit equilibrium reaction system that processes synthesis gas of nitrogen and hydrogen to produce ammonia.

[0112] Nitrogen is provided by one or more pipes from the ASU 16, and in an embodiment, the air separation unit may operate continuously to provide nitrogen. Hydrogen is provided by one or more pipes from the hydrogen production apparatus 12 (if it operates based on the availability of renewable power in a given case). Otherwise, hydrogen is supplied from the hydrogen storage unit 14.

[0113] The stoichiometric composition of the synthesis gas is processed by the synthesis gas compressor system, and the resulting ammonia product is refrigerated by another set of compressors and sent to the storage. The performance of the ammonia loop is determined by the equilibrium conversion rate of the exothermic reaction. The parameters used for this will be discussed below.

[0114] Power generation and management system

[0115] The power of the industrial gas equipment complex 10 as a whole can be generated by any suitable energy source (including renewable or non-renewable energy sources). As Figure 1 shown, the power is generated by at least one renewable energy source among wind energy 24 (via a suitable wind power generation field including multiple wind turbines) and / or solar energy 26 (via a solar power generation field including multiple solar cells). In addition, other renewable energy sources can be used, such as a hydro power source (not shown) and / or a tidal power source (not shown).

[0116] In addition, the power or resources of the industrial gas equipment complex 10 as a whole or of the sub-equipment of the industrial gas equipment complex 10 can be extracted from the energy storage resource 28. As referred to Figure 1 above, the energy storage resource 28 can include one or more storage resources. For example, one or more storage resources can include a hydrogen storage unit 14 and a liquid nitrogen storage 16a.

[0117] In addition or alternatively, in a non-exhaustive arrangement, the energy storage device can include one or more of the following: a battery energy storage system (BESS) 28a, a compressed / liquid air energy system (CAES or LAES) 28b, or a pumped hydro storage system (PHSS) 28c.

[0118] When the power supply from renewable sources is high or is predicted to be high, these elements are optimally used to store additional resources and / or energy, and then these resources and / or energy are utilized when the renewable power resources are predicted to be low.

[0119] The prediction and control of these facilities will be described below. It is very important to select these facilities under optimal conditions so that, for example, the correct energy source is selected for a specific predicted power shortage period.

[0120] Control system 100

[0121] Figure 1 The control system 100 of Figure 2 is shown in detail in the schematic diagram of

[0122] The control system 100 includes three main categories: the equipment complex control system 110, the renewable energy control system 120, and the optimization system 150. These are non-limiting terms and do not necessarily imply any interconnection or grouping among the components of systems 110, 120, 150, and are shown in a common grouping only for the purpose of clarity.

[0123] The equipment complex control system 110 includes a hydrogen production equipment control system 112, a hydrogen storage control system 114, an ASU control system 116, and an ammonia synthesis equipment control system 118.

[0124] As an example, the hydrogen production equipment control system 112 can be configured to monitor the amount and rate of hydrogen generated from electrolysis by measurement. Such measurement can be derived from sensor measurements such as direct flow measurement, or alternatively inferred by indirect measurements such as electrolyzer current or power demand.

[0125] As another example, for the hydrogen storage control system 114, the pressure and flow rate of compressed hydrogen from the electrolyzer and compression system to the storage system, as well as the pressure and flow rate of compressed hydrogen to the ammonia synthesis equipment 18, can be monitored.

[0126] In each case, the control system 110 is operable to control the parameters of the corresponding industrial gas equipment and is capable of outputting usage and process data from each industrial gas equipment. This will be described in detail below.

[0127] The renewable energy control system 120 in the described embodiment includes a wind control system 124 and a solar control system 126. These control systems control and monitor the process parameters of renewable energy, such as energy generation, storage, and load. They are also configured to send usage, power, and process data to external systems as needed. Similar control systems will apply if hydroelectric or tidal renewable power sources are used.

[0128] The optimization system 150 includes a computer system, and the computer system includes three modules: a power prediction module 152, an equipment operation module 154, and a real-time optimization module 156.

[0129] The power prediction module (PPM) 152 receives usage and power generation data from the renewable energy control system 120 and the weather and forecast database 160, and the database includes information related to past (known and historical) environmental and weather data as well as future (predicted and forecast) environmental and weather data. The power prediction module 152 includes machine learning algorithms implemented on a computing system as will be described below, and is used to generate a model related to future power generation.

[0130] The Plant Operations Module (POM) 154 is operable to receive plant operation data from the plant complex control system 110 and generate a model of the plant operation. The plant operations module 154 includes machine learning algorithms implemented on a computing system as will be described below.

[0131] The Real-Time Optimization Module (RTOM) 156 is arranged to receive inputs from the power prediction module 152 and the plant operations module 154, and derive a plant operation strategy that includes setpoint operating parameters. These are then fed to the plant complex control system 110 to control the associated processes thus controlled.

[0132] Details and operations of each component will now be described.

[0133] Power Prediction Module (PPM) 152

[0134] The power prediction module 152 includes machine learning algorithms implemented on a computing system, and is operable to generate a model to predict future power generation. In an embodiment, one aspect of the power prediction module 152 is the ability to predict future power generation from variable and / or intermittent sources such as renewable power sources, such that one or more industrial gas devices (which typically require a constant power load) can be controlled without the risk of power shortages in the devices.

[0135] In an embodiment, a further aspect of the power prediction module 152 is to use the predicted power generation data to control the industrial gas plant complex 10 or aspects thereof. This will be discussed in more detail below.

[0136] The model for predicting future power is based on a machine learning framework. Any suitable machine learning algorithm can be used. For example, the model can utilize techniques such as gradient boosting (using e.g., XGboost), long short-term memory (LSTM), support vector machines (SVM), or random decision forests can be used in such models.

[0137] Gradient boosting is a machine learning technique for regression and classification problems. A strong prediction model is formed that includes an ensemble of weak prediction models such as decision trees. A stepwise process can be used to generate the model by minimizing steepest descent (and other methods).

[0138] LSTM is an artificial recurrent neural network architecture with both feedback and feedforward connections. A common LSTM cell consists of a cell, an input gate, an output gate, and a forget gate. The cell is operable to remember values over any time interval where the flow of information into and out of the cell is regulated by the gates.

[0139] A support vector machine uses a set of training instances, each of which is included in one of two classes, and generates a model that assigns new instances to a specific class. Thus, SVMs include non-probabilistic binary linear classifiers.

[0140] Random decision forests include an ensemble machine learning method that operates by constructing a large number of decision trees during the training process and outputting the class of the mode as the class of the individual trees (classification) or the medium / average prediction (regression).

[0141] Regardless of the machine learning algorithm used, the model utilizes a two-step operation process, where a training phase is required before the prediction phase. Both phases are implemented as computer programs on one or more computer systems.

[0142] Professional hardware can also be used. For example, the training phase may involve using the central processing unit (CPU) and graphics processing unit (GPU) components of a computer system. In addition, other professional hardware such as field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), or other stream processor technologies can also be used.

[0143] Training phase

[0144] The model is trained periodically (e.g., on a daily basis) or on demand (if, for example, the accuracy of the model requires a training process) to establish relationships between the predictor variables of the renewable energy power device, thereby determining the predicted energy availability within a predetermined future period.

[0145] PPM 152 is designed to determine the following predictor variables:

[0146] Wind power WPi

[0147] Solar power SPi

[0148] where the index i represents the time from period n to n + k (in intervals of a fixed duration). In a non-limiting embodiment, the interval can include 15 minutes or 1 hour.

[0149] The predictor variables include the following time-related operating characteristic data:

[0150] Wind power WPi

[0151] Solar power SPi

[0152] Power load Li from period n - m to n - 1.

[0153] The predictor variables also include measured environmental and meteorological signals, which can include but are not limited to time-related environmental data, and the time-related environmental data includes:

[0154] Air temperature Ti,

[0155] Atmospheric pressure Pi

[0156] Wind speed WSi,

[0157] Cloud cover CCi,

[0158] Precipitation Pi

[0159] Humidity Hi

[0160] where the index i represents the time from period n - m to n + k.

[0161] During the training phase, various machine learning algorithms are used to create the mathematical relationships between the predicted variables and the predictor variables. These relationships are stored in the computer in the form of a series of equations that can be accessed and used to make future predictions.

[0162] These models are generated during the training process, which is carried out regularly at a predetermined training time. At each training time, the machine learning model is trained based on the historical time - related environmental data and historical time - related operating characteristic data obtained within one or more predetermined historical time windows. In other words, a moving historical window of time - related data (e.g., the previous 2 days, the previous 2 weeks, etc.) is used to train the machine learning model.

[0163] As described above, the meteorological measurements are from the weather and forecast database 160, which may include weather data services or other Internet - connected resources. The load, solar power, and wind power are measured by the operator or fed through the renewable energy control system 120 as part of an automated data collection system.

[0164] Prediction phase

[0165] During the prediction phase, the equations generated and stored during the training phase are used to make predictions about future behavior.

[0166] The predicted variables include the prediction of the wind power WPi and solar power SPi of i from c + 1 to c + p, where c represents the current time, and p represents the prediction horizon, which can be any suitable time scale; for example, non - limiting instances can be 24 hours, 48 hours, or longer.

[0167] The predictor variables are:

[0168] Wind power WPi

[0169] Solar power SPi

[0170] Power load Li

[0171] Measure the meteorological signal from c-r to c (according to the training phase), and predict the meteorological signal from c+1 to c+p.

[0172] Regarding the training phase, in an embodiment, the environmental and meteorological measurements are from the weather and forecast database 160, which may include weather data services or other Internet-connected resources. The load, solar power, and wind power are measured by the operator or automatically fed through the renewable energy control system 120.

[0173] Figure 3 and 4 Shows the prediction of wind power and solar power, where the predicted variables are generated from the above model prediction phase.

[0174] In an embodiment, the accuracy of the machine learning model is evaluated by comparing the values of the predicted power resources within a predetermined future time period with the actual power resources available for the industrial gas equipment at the end of the predicted future time period. This enables the determination of the prediction error value that provides a measure of the accuracy of the model.

[0175] If the model is not accurate enough, it may need to be trained at an additional training time. The predetermined training time can be selected based on the prediction error value, which in an embodiment can be when the prediction error value exceeds a predetermined threshold.

[0176] In an embodiment, the training time can be selected empirically on a periodic basis; for example, every 24 hours or every 48 hours as appropriate. In an embodiment, this periodic basis can be the default training time strategy.

[0177] However, in an embodiment, when the prediction error value exceeds a predetermined threshold within a predetermined empirical interval, this can be interrupted, and in this case, the training time is scheduled based on the prediction error value.

[0178] Control Based on Predicted Data

[0179] PPM 152 can utilize the predicted data to control one or more industrial gas devices in response to the predicted available power resources for a predetermined future time period. For example, in addition to RTM 156, PPM 152 can control the operational aspects of one or more industrial gas devices.

[0180] In an embodiment, the forecast power data can additionally or alternatively control the use of the energy storage resource 28. As described above, the energy storage resource 28 can include resource storage such as stored hydrogen and / or nitrogen, as well as energy storage through BESS28a, CAES / LAES28b, and PHSS28c, etc.

[0181] Thus, the PPM 152 may include control and optimization algorithms that utilize predicted power resources to optimize the use of the available energy storage resources 28 to best operate the industrial gas equipment complex 10 and to plan for future power availability and power usage.

[0182] The PPM 152 system uses the predicted powers WPi and SPi as inputs in a real-time optimization problem and applies an optimization algorithm to propose an optimal rate for operating the industrial gas equipment complex 10 to maximize the utilization of the available power and the stored resources in the energy storage resources 28.

[0183] For example, by determining the optimal use of the short-term (e.g., BESS 28a) and long-term (e.g., CAES / LAES 28b and PHSS 28c) storage systems in the energy storage resources 28, the expected daily and seasonal variability in the predicted power generation can be addressed by the PPM 152.

[0184] In an embodiment, the PPM 152 may determine the optimal on and off times of the energy storage resources 28 to maximize the utilization of the available power. Examples in this regard will be described below.

[0185] In a non-limiting example, the PPM 152 may determine that there is sufficient power available within the next 24 hours to operate the electrolyzers 12a...n of the hydrogen production equipment 12 at a greater load to generate more hydrogen than is required to produce the optimal ammonia production rate over a predetermined period of time. The additional hydrogen can then be stored in the hydrogen storage unit 14 for use during periods of lower power availability. This also applies to the ASU 16 and the liquid nitrogen storage 16a.

[0186] The PPM 152 may receive the operating characteristic data of the hydrogen storage unit 14 and the liquid nitrogen storage 16a. This may include fill levels and other operating data (such as fill pressure, fill volume, density, etc.). This operating data can be used by an algorithm that forms part of the PPM 152 to determine the optimal storage requirements, thus addressing the predicted future power availability profile as a function of time.

[0187] In an embodiment, the PPM 152 may also determine that there is sufficient power available within a predicted time window to store additional power in the energy memories 28a, 28b, 28c. The PPM 152 may utilize an optimization algorithm to select the appropriate energy memories 28a, 28b, 28c based on the prediction of power availability relative to time.

[0188] In a non - limiting example, PPM 152 can determine that power availability exceeding the demand of the industrial gas equipment complex 10 is available within a relatively short time period (e.g., 1 - 2 hours). Given the short ramp rate, fast charging time, and lower capacity of the BESS28a solution, it can then be determined that BESS28a represents the most suitable energy storage solution for this time period.

[0189] In an alternative non - limiting example, PPM 152 can determine that significant power availability exceeding the demand of the industrial gas equipment complex 10 is available within a longer time period (e.g., 5 - 10 hours). In this case, given the slower ramp rate and higher storage capacity of such an energy storage solution, it can then be determined that CAES / LAES 28b and PHSS28c may be more suitable for storing the available power.

[0190] The PPM 152 system solves the optimization problem for the next p time periods and applies the available power prediction to a real - time optimization model to maximize the utilization of power. The model can involve the generation of power utilization metrics, and the optimization seeks to optimize the value of the power generation metric.

[0191] PPM 152 can also include a tracking system to calculate the predicted power utilization metric as a function of the complementarity of the renewable power resources and adjust the generation level / utilization to maximize the long - term utilization of the renewable resources.

[0192] The PPM 152 system is implemented on a computer and receives various inputs from other computer systems. An example of the PPM 152 system can utilize a mixed - integer nonlinear program (MINLP) because some decisions require some devices to operate in one of multiple possible modes that result in integer variables. Figure 5 A method according to one embodiment is shown. It should be noted that the following steps do not need to be performed in the order described below, and some steps can be performed concurrently with other steps.

[0193] In an embodiment, a method for predicting available power resources from one or more renewable power sources for one or more industrial gas devices is provided. The method is executed by at least one hardware processor.

[0194] In step 200, historical time - related environmental data associated with one or more renewable power sources 24, 26 is obtained. Historical refers to past environmental data. This can be collected in any suitable time window and can contain data within the window extending up to but not including the current time.

[0195] In step 210, historical time-related operating characteristic data associated with one or more renewable power sources. Historical refers to past operating characteristic data, such as power output. This can be collected over any suitable time window and can include data within a window extending up to but not including the current time.

[0196] In step 220, train a machine learning model based on the historical time-related environmental data and historical time-related operating characteristic data captured in steps 200 and 210. The model can be trained any suitable number of times and at predetermined intervals or as needed.

[0197] In step 230, execute the trained machine learning model to predict the available power resources of one or more industrial gas devices over a predetermined future time period.

[0198] In step 240, the predicted data can optionally be used to control one or more industrial gas devices in response to the predicted available power resources for the predetermined future time period. For example, one or more operating set points of one or more industrial gas devices can be set based on the predicted available power resources.

[0199] In addition, as described above, an optimization algorithm can be utilized based on the predicted available power resources for a predetermined future period to determine how to best manage the energy storage resources 28, which include resource memories such as the hydrogen storage unit 14 and the liquid nitrogen memory 16a, as well as the energy storage resources 28a, 28b, 28c. The optimization algorithm can select the best schedule and / or select the best resources for a given pattern of predicted power availability as a function of time.

[0200] The selection and scheduling can be based on the operating characteristics of the storage resources; for example, the fill levels, pressures, and other characteristics of the hydrogen storage unit 14 and the liquid nitrogen memory 16a, or the ramp-up / ramp-down, time dependence, maximum generation power / maximum stored energy operating profiles of the energy storage resources 28a, 28b, 28c.

[0201] In step 250, determine whether an additional training process is needed. This can be based on empirical metrics, such as a predetermined time period. Alternatively, it can be based on an assessment of the accuracy of the machine learning model by comparing the value of the predicted power resources for a predetermined future time period with the actual power resources available to the industrial gas devices at the end of the predicted future time period. This enables determination of a prediction error value that provides a measure of the accuracy of the model.

[0202] If it is determined that the model needs to be retrained, a training time can be scheduled, and the method can return to the training in step 220. It should be noted that this can occur after or during either step 230 or 240.

[0203] Plant Operation Module (POM) 154

[0204] Optimization of plant process control is crucial for achieving efficiency. Considering that renewable power sources almost always result in variations in available power, the industrial gas plant complex 10 may frequently operate in a dynamic mode. This requires real-time performance models to design robust operation strategies. POM 154 includes machine learning-based and physics-based models to model plant operations. The physics-based models can be used to generate predictor variables indicative of the operating characteristics of the corresponding industrial gas plants. In an embodiment, these can be used, together with other operating plant data (e.g., power input, power output, gas output), as time-correlated data inputs to the corresponding machine learning models.

[0205] As described above, the industrial gas plant complex 10 includes a hydrogen production plant 12, a hydrogen storage unit 14, an air separation unit (ASU) 16, an ammonia synthesis plant 18, and an ammonia storage unit 20, some of which are controlled by corresponding control systems: a hydrogen production plant control system 112, a hydrogen storage control system 114, an ASU control system 116, and an ammonia synthesis plant control system 118.

[0206] In an embodiment, at each time step, the power availability forecast from PPM 152 is used to define the operating set points of the different industrial gas plants 112, 114, 116, 118.

[0207] These decisions are actionable in an embodiment to achieve high process efficiency. This requires an accurate quantitative understanding of the different process units in terms of their real-time performance, system availability information, and resource availability and upcoming maintenance issues.

[0208] POM 154 is configured to capture the time-varying properties of each subsystem (i.e., the corresponding industrial gas plant) and predict the production efficiency of the overall process based on the ammonia produced for a given energy consumption level.

[0209] The high-fidelity models for such predictions are based on a real-time machine learning framework that uses time-correlated historical data provided by a distributed control system (DCS) in the form of time series for various process tags. Any suitable machine learning algorithm can be used to build an integrated model for an individual subsystem. For example, the model can utilize techniques such as gradient boosting (using e.g., XGboost), long short-term memory (LSTM), support vector machine (SVM), or random decision forest.

[0210] POM154 - Modeling of Hydrogen Production Plant 12

[0211] Water electrolysis is an energy-intensive process and a key process step in the production of green hydrogen. Each of the electrolyzer modules 12a, 12b... 12n in the hydrogen production device 12 consists of hundreds of electrolytic cells that work together to convert renewable power into hydrogen molecules controlled by a time-dependent efficiency η.

[0212] Each electrolyzer module 12[k] in the hydrogen production device 12 is independently modeled based on its historical performance data. The predicted operating characteristic variables used are:

[0213] Electrolyzer power consumed [EP(i,k)]

[0214] Electrolyzer hydrogen produced [EH(i,k)]

[0215] Based on the number of predictor variables, such predictor variables as:

[0216] Demineralized water flow rate [ED(i,k)]

[0217] Average cell temperature [ECT(i,k)],

[0218] Average cell pressure [ECP(i,k)],

[0219] Current flowing through the electrode [I(i,k)]

[0220] And other key process indicators.

[0221] The historical time series data of the predictor variables and the response variables are sampled at an appropriate frequency [s] over several months and used to develop a model of the actual module efficiency. An equation derived from the predictor variables is used to establish the model.

[0222] In addition, an accumulator variable is used to track the functional age of the module, which is one of the predictor variables in the model. Reliability and maintenance event information from the asset management system.

[0223] The model will be trained regularly or on demand (to establish a relationship between the response variable and the predictor variables at each time instance starting from n - 1 to n - k within a fixed-duration interval such as 15 minutes or 1 hour at time n).

[0224] In the training system, various machine learning algorithms are used to create a mathematical relationship between the predicted variables and the predictor variables.

[0225] POM154 - Ammonia loop modeling

[0226] The ammonia circuit is a single-unit equilibrium reaction system that processes synthesis gas of nitrogen and hydrogen to produce ammonia. Nitrogen is provided by the ASU 16, which, in the embodiment, operates continuously to supply nitrogen.

[0227] If the hydrogen production device 12 operates based on the availability of renewable power in a given situation, hydrogen is provided from this hydrogen production device; otherwise, hydrogen is supplied from the hydrogen storage unit 14.

[0228] The stoichiometric composition of the synthesis gas is processed by the synthesis gas compressor system, and the product is refrigerated by another set of compressors and sent to the memory.

[0229] The performance of the ammonia circuit is controlled by the equilibrium conversion rate of the exothermic reaction and is monitored in real time according to a prediction model of AFi as a function of various predictor variables, including:

[0230] The power consumed by the ammonia circuit, APi,

[0231] The ammonia circuit pressure and temperature, ALPi, ALTi

[0232] The feed flow rates of the nitrogen and hydrogen streams, ANFi, AHFi

[0233] The synthesis gas compressor pressure of the ammonia equipment, ACPi.

[0234] Additional information regarding the health status of the catalyst bed and the time of maintenance events can be used in the model to obtain the most realistic picture of the conversion circuit efficiency.

[0235] Another aspect of the ammonia circuit is different operating modes. In the embodiment, there are two main modes: normal and standby. The normal mode involves ramping up and down in response to the available amount of hydrogen. This data is tracked in the DCS and is used to view any performance differences or diagnose any process deviations from the production plan. This time-related operating characteristic data can be used as input for a trained machine learning model to predict future operating behavior.

[0236] As described below, the model is regularly trained on a longer-range historical dataset (e.g., it can be from 6 months to one year) to capture all modes and different levels of ramp rates.

[0237] POM154 - Modeling of Hydrogen Storage Unit 14

[0238] In the embodiment, the hydrogen storage unit 14 includes a hydrogen purification chain, a memory, and a set of compressors that are operable in a dynamic manner to deliver hydrogen to the ammonia processing equipment and manage the hydrogen inventory to avoid downtime due to lack of available gas resources.

[0239] The overall performance of the system is measured by reaching a specified set point of the header pressure in a reliable and energy-efficient manner. The compressors are modeled based on isentropic efficiency based on operating temperature and ambient conditions. Real-time status monitoring of all compressors is based on adaptive multivariable (principal component analysis (PCA) and partial least squares (PLS) models built on a moving 3-month window of historical data of key process tags.

[0240] Real-time tracking of the hydrogen storage can be based on a first principal thermodynamic model that is based on using:

[0241] Storage system pressure and temperature, SPi, STi

[0242] Hydrogen compressor pressure and flow, HCPi, HCFi.

[0243] The efficiency of the compressor system is tracked using the consumed compressor power CPi. Additionally, the system can perform real-time monitoring of the hydrogen purification system in terms of efficiency and reliability.

[0244] POM154 - ASU 16 Modeling

[0245] To model the air separation unit 16, in an embodiment, multivariate partial least squares (PLS) and principal component analysis (PCA) models as well as engineering models are utilized. These models are operable to diagnose performance impacts and identify preferred operating modes. Data from these models can be used as operating characteristic data input into a trained machine learning model.

[0246] Several key performance indicators (KPIs) can be selected. In non-limiting examples, they can include specific power, N2 recovery rate, and temperature difference in heat exchangers that form part of the ASU. The KPIs are tracked in real-time for early detection and diagnosis of inefficient operation as well as emerging equipment health degradation. The generation of the KPIs is not important for the present invention, however these values can be used by a predictive machine learning model to build a predictive model of the performance of ASU 16.

[0247] Industrial Gas Equipment Model Component Training and Prediction

[0248] For each industrial gas equipment, a machine learning model is assigned and implemented as described above. For each equipment, the machine learning model is operable to generate equations using a training process that model the behavior of the corresponding industrial gas equipment. This is done using the predictor variables described above for each industrial gas equipment.

[0249] The training process can take operation characteristic data related to historical time as input to train a machine learning model. The operation characteristic data can include physical measurement data related to the corresponding industrial gas equipment. For example, in non-limiting embodiments, input power, power usage, gas output, measured efficiency, etc.

[0250] In addition, the operation characteristic data can include data generated by one or more physics-based models as described above. The physics-based models can adopt the measured specific industrial gas equipment characteristics (specific to each industrial gas equipment) and can generate one or more metrics indicating the performance of the industrial gas equipment. Then, these time-related metrics can be used as the input operation characteristic data to train the machine learning model assigned to the corresponding industrial gas equipment.

[0251] Once the training process at training time is completed, the model can be used to predict the behavior of the corresponding industrial gas equipment. In the prediction phase, the equations generated and stored during the training phase are used to make predictions about future behavior.

[0252] In an embodiment, the accuracy of the machine learning model is evaluated by comparing the predicted future behavior values of each industrial gas equipment for a predetermined future time period with the actual behavior of the industrial gas equipment at the end of the predicted future time period. This enables the determination of a prediction error value that provides a metric for the accuracy of the model.

[0253] If the model is not accurate enough, it may need to be trained at additional training times. The predetermined training times can be selected based on the prediction error value. In an embodiment, this can be when the prediction error value exceeds a predetermined threshold.

[0254] In an embodiment, the training times can be selected empirically on a periodic basis; for example, every 24 hours or every 48 hours as appropriate. In an embodiment, given that the historical data may span several months, training may not be required so frequently. In an embodiment, this periodic basis can be the default training time strategy.

[0255] However, in an embodiment, when the prediction error value exceeds a predetermined threshold within a predetermined empirical interval, this can be interrupted, and in this case, the training time is scheduled based on the prediction error value.

[0256] POM154 Summary

[0257] As described above, machine learning models are provided for each industrial gas equipment forming an industrial gas complex. These machine learning models are trained on operation characteristic data generated from historical time-related data measured related to the associated equipment and / or from time-related data generated by one or more physics-based models of the corresponding equipment.

[0258] These models each generate prediction data related to one or more of the following: performance; capacity; efficiency; maintenance status; and / or utilization of related equipment.

[0259] In Figure 7 an example of the generated efficiency curve of the process variable is shown. Figure 7 A curve showing the efficiency of the electrolyzer relative to the load on the electrolyzer is shown. This curve is generated from time - related operating characteristic data of the electrolyzer that forms part of the hydrogen - forming device 12, and this data is input into the corresponding machine - learning model to generate operating characteristic data.

[0260] In addition to the predictions of the individual machine - learning models assigned to each industrial gas device, additional models can be utilized. The additional models determine the overall performance of the device complex in a given situation based on the combined efficiency of the individual modules from each of the above - mentioned models. An integrated machine - learning algorithm is employed to improve the quality of the prediction, and the best model is selected for the prediction system.

[0261] In other words, the time - varying operating characteristics of each industrial gas device forming part of the device complex are predicted by each trained machine - learning model and input into an additional model to predict the production efficiency of the entire process - equipment complex. Data from each model can be input into the collective model on a periodic basis; for example, in a non - limiting embodiment, this can be once every 15 minutes.

[0262] For the ammonia - production device in an exemplary embodiment, the efficiency determination enables the determination of the prediction of the ammonia produced for a given level of energy input.

[0263] All modeling performed in POM 154 is implemented as a computer program on one or more computers in the device. Regardless of the machine - learning algorithm used, the model utilizes a two - step operation process, where a training phase is required before the prediction phase. Both of these phases are implemented as computer programs on one or more computer systems.

[0264] Professional and non - professional hardware can also be used. For example, the training phase may involve using the central processing unit (CPU) and graphics processing unit (GPU) components of a computer system. In addition, other professional hardware, such as field - programmable gate arrays (FPGA), application - specific integrated circuits (ASIC), or other stream - processor technologies, can also be used.

[0265] The model execution computer will be connected to other computer database systems, where data from the devices and weather data services will be stored. When the devices operate, the performance model is used every 15 minutes to make predictions to obtain a production profile.

[0266] All performance models, along with some unit operation level data, are used in RTOM 156 at a predefined frequency to optimize operation efficiency and define set points for different models.

[0267] Method of operation

[0268] Figure 6 A method according to an embodiment is shown. It should be noted that the following steps do not need to be performed in the order described below, and some steps can be performed simultaneously with other steps.

[0269] In an embodiment, a method for predicting the operating characteristics of an industrial gas equipment complex including a plurality of industrial gas devices is provided. The method is executed by at least one hardware processor.

[0270] In step 300, a machine learning model is assigned to each industrial gas device that forms the industrial gas equipment complex. The model can take any suitable form as described above. It can utilize the historical time-related operating characteristics of the industrial gas device to generate equations for future predictions.

[0271] In step 310, the corresponding machine learning model of each industrial gas device is trained based on the received historical time-related operating characteristic data of the corresponding industrial gas device. The data can take any suitable form and can be specific to the particular type of industrial gas device as described above. Historical refers to past operating characteristic data. This can be collected in any suitable time window and can include data within a moving window that extends up to but does not include the current time. In an embodiment, the window can be six months to one year long.

[0272] The historical time-related operating characteristic data can include physical measurement data related to the corresponding industrial gas device, such as, in a non-limiting embodiment, input power, power usage, gas output, measured efficiency, etc.

[0273] In addition, the operating characteristic data can include data generated by one or more physics-based models as described above. The physics-based model can take the measured characteristics of the particular industrial gas device (specific to each industrial gas device) and can generate one or more metrics indicating the performance of the industrial gas device. Then, these time-related metrics can be used as input operating characteristic data to train the machine learning model assigned to the corresponding industrial gas device.

[0274] In step 320, a machine learning model trained for each industrial gas device is executed to predict the operating characteristics of each corresponding industrial gas device over a predetermined future time period. This prediction can optionally be used to control the behavior of the corresponding industrial gas device, predict possible usage, maintenance schedules, resource allocation, or identify process issues and potential problems.

[0275] For example, the predicted data can be used to infer other technical performance of the industrial gas device. These predictions can be used to determine resource planning, maintenance schedules, or repair requirements. Such maintenance schedules can be carried out in conjunction with the determination of power resources and the capacity of storage units. For example, the maintenance of gas generation components (e.g., electrolyzers, ASUs, ammonia production equipment) can be scheduled during a time period when the gas storage level is high and the predicted available renewable power is low, in order to minimize interruptions and maintain the continuity of service provision.

[0276] Furthermore, in step 330 below, the data can also be used to determine setpoint characteristics.

[0277] In step 330, the predicted operating characteristics determined by the model in step 320 for each corresponding industrial gas device are used in an additional collective model to generate an operating performance metric for the industrial gas device complex. In an ammonia production device in an exemplary embodiment, efficiency determination enables the determination of the predicted production of ammonia for a given level of energy input.

[0278] In step 340, the predicted data for a predetermined future time period is compared with the actual measured data at the end of that time period. This comparison is used to infer other technical performance of the industrial gas device. These predictions can be used to determine resource planning, maintenance schedules, or repair requirements.

[0279] In an embodiment, the prediction for a predetermined time period (e.g., a two-week, one-month, six-month time window starting from the generation of the predicted data or from the timestamp in the predicted data) is then compared with the actual data at the end of the time window or time period covered by the predicted data (e.g., two weeks / one month / six months after the generation of the predicted data). This enables the early identification of potential problems in the industrial gas device complex 10, as any deviation of an industrial gas device or storage system from the prediction model based on past actual behavior may indicate the occurrence of production or maintenance problems.

[0280] By using this method, potential future problems can be identified early, enabling remedial measures to be taken before any critical failures or unplanned downtimes that require equipment service for emergency maintenance or repair.

[0281] In addition, in an embodiment, maintenance scheduling can be done in conjunction with the determination of power resources and storage unit capacity. For example, maintenance of gas generation components (e.g., electrolyzers, ASUs, ammonia production equipment) can be scheduled during periods when gas storage levels are high and predicted available renewable power is low, in order to minimize disruptions and maintain continuity of service provision.

[0282] In an embodiment, the predicted data can optionally be used to control one or more industrial gas devices in response to predicted operating characteristics of the industrial gas devices over a predetermined future time period. For example, one or more operating set points of one or more industrial gas devices can be set based on the predicted behavior.

[0283] The control in step 340 can be done in conjunction with the methods described in steps 200 to 250, where the predicted power availability is utilized in conjunction with the predicted efficiency determination in step 330, so as to enable the determination of set points based on the predicted power availability and equipment efficiency.

[0284] In step 350, it is determined whether an additional training process is needed. This can be based on empirical metrics such as a predetermined time period. Alternatively, it can be based on an assessment of the accuracy of the machine learning model by comparing the values of the predicted operating characteristics over a predetermined future time period with the actual operating characteristics of the industrial gas device at the end of the predicted future time period. This enables the determination of a prediction error value that provides a metric for the accuracy of the model.

[0285] If it is determined that the model needs to be retrained, a training time can be scheduled and the method can move back to the training in step 310. It should be noted that this can occur after any one of steps 320, 330, or 340.

[0286] Real-Time Optimization Module (RTOM) 156

[0287] RTOM 156 includes a system for determining the rates at which various industrial gas devices should operate to optimally manage the production and storage of renewable hydrogen while maximizing ammonia production. In other words, RTOM 156 is capable of performing real-time optimization of an industrial gas device complex using renewable power. More specifically, in an embodiment,

[0288] In an embodiment, the system solves an optimization algorithm for a dynamic mathematical model applied to the industrial gas device complex. The RTOM 156 system uses the predicted power WPi and SPi and the states of the industrial gas devices, such as the efficiency inferred from device-specific factors (such as current, pressure, temperature, and flow rate measurements) discussed above with respect to POM154. These values are used as inputs, and an optimization algorithm is applied to propose the optimal operating rates of the ammonia devices over the time period from c + 1 to c + p.

[0289] The rates of other industrial gas equipment such as hydrogen production equipment 12, hydrogen storage unit 14, air separation unit 16, liquid nitrogen gas storage 16a, and water equipment are associated with the rate of ammonia and are controlled by the low-level controller as described above. Only the first value from the list of optimal values for time c + 1 is implemented, and when new data is available, the calculation is repeated at time c + 1 with the new data.

[0290] RTOM 156 can also utilize data related to the resource storage device or energy storage device of the energy storage resource 28. As described above, the energy storage device can include one or more of the following: battery energy storage system (BESS) 28a, compressed / liquid air energy system (CAES or LAES) 28b, or pumped hydro storage system (PHSS) 28c. The state, operating characteristics, availability, resource storage level, and power availability ease of each unit of the energy storage resource 28 can be taken into account in the optimization problem.

[0291] The RTOM 156 system is implemented on a computer and receives various inputs from other computer systems. The predicted power comes from PPM 152. The state of the industrial gas equipment is represented by equations that relate the power consumption of various units (such as electrolyzers (ECi), hydrogen compressors (CPi), ammonia equipment (APi), and ASU (NPi)) to process variables associated with these units.

[0292] The RTOM 156 system solves an optimization problem that seeks to maximize ammonia production subject to constraints on the total predicted power available over the next p time periods, the amount of hydrogen available in storage, and physical or data-based equations that describe the operation of the process equipment.

[0293] Typically, such problems are mixed integer nonlinear programming (MINLP) because the process equipment equations are nonlinear and some decisions require some equipment to operate in one of several possible modes that result in integer variables. The optimization generates setpoints from time c + 1 to c + p to balance the predicted generated power and consumed power such that an appropriate amount of hydrogen is produced and the ammonia equipment operates at the correct rate to maximize ammonia production. The RTOM 156 system also takes into account hydrogen storage, and based on future power predictions, hydrogen can be stored or consumed from storage.

[0294] In addition, the RTOM 156 system can select to recommend that equipment enter standby or shutdown mode based on power availability and equipment availability information. The output of the RTOM 156 system is the recommended ammonia production rate, which is automatically transmitted to an advanced control system that controls the ammonia equipment, ASU, electrolyzers, water equipment, and hydrogen compression and storage.

[0295] Examples of this are Figure 8 , which shows the optimum ammonia plant rate as a function of time over a 48 hour forecast period.

[0296] Method of operation

[0297] In an embodiment, a method of controlling an industrial gas plant complex including a plurality of industrial gas plants powered by one or more renewable power sources is provided. The method is performed by at least one hardware processor.

[0298] At step 410, time-dependent predicted power data for a predetermined future time period is received from one or more renewable power sources. In a non-limiting embodiment, this may be determined by the PPM 152 according to the process and method discussed in steps 200 to 250. Alternatively, the predicted power data may be obtained according to any other suitable process.

[0299] At step 420, time-dependent predicted operating characteristic data for each industrial gas plant is received. In this context, "industrial gas plant" may also include industrial gas storage, for example, hydrogen, nitrogen and / or ammonia storage as described above in the present embodiment.

[0300] In an embodiment, the operational characteristic data may be based on information about the POM 156 and Figure 6 The data may be generated by the protocol described in method steps 300 to 350. However, the data may also be determined according to any other suitable process.

[0301] At step 430, the predicted power data and the predicted characteristic data are utilized in an optimization model to generate a set of state variables (which may be optimized state variables) for a plurality of industrial gas devices. In an embodiment, this may be accomplished by solving an optimization problem. For example, in a non-limiting embodiment, the optimization model may define the predicted power data and the predicted characteristic data as a set of nonlinear equations. Storage resource data may optionally be included in the power data. The state variables are then generated by solving the set of nonlinear equations.

[0302] At step 440, a set of control set points for a plurality of industrial gas plants are generated using the generated state variables (which may be optimized state variables). The set points may be defined to achieve any particular goal. For example, the set points may be defined at a particular time to ensure that the industrial gas plants operate efficiently and effectively given available power and storage resources. The set points may alternatively or additionally be used to maximize the production of industrial gas given a predicted power availability.

[0303] As another example, the predicted power availability, predicted efficiency, and operating characteristics of individual industrial gas equipment and / or an industrial gas equipment complex as a whole can be used to prevent power shortages in the individual equipment and / or equipment complex, or to increase production when power availability is sufficient at the current time but a future shortage is predicted.

[0304] In step 450, control setpoints are sent to the control system to control the industrial gas equipment complex by adjusting one or more control setpoints of the industrial gas equipment.

[0305] In summary, in an exemplary embodiment, a control and optimization system for an industrial gas equipment complex is provided. The industrial gas equipment complex includes a plurality of industrial gas equipment, including hydrogen production equipment using water electrolysis, air separation equipment for producing nitrogen, ammonia synthesis equipment for producing ammonia, a hydrogen storage system, and an ammonia storage and transportation system.

[0306] In an exemplary embodiment, given uncertain future power inputs to a system, the system seeks to maximize ammonia production. The system also includes a machine learning-based program to predict future power inputs based on weather forecasts. This program continuously learns from the environment, weather patterns, and generated power to make future power generation predictions. The system also includes another machine learning program to learn operating data from the industrial gas equipment and modify the mathematical model for optimization.

[0307] Although the invention has been described with reference to the preferred embodiments shown in the drawings, it should be understood that various modifications are possible within the spirit or scope of the invention as defined by the appended claims.

[0308] In the specification and claims, the term "industrial gas equipment" is intended to mean process equipment that produces or participates in the production of industrial gases, commercial gases, medical gases, inorganic gases, organic gases, fuel gases, and green fuel gases in gaseous, liquefied, or compressed form.

[0309] For example, the term "industrial gas equipment" may include process equipment for manufacturing gases such as those described in NACE class 20.11 and including but not limited to: elemental gases; liquid or compressed air; refrigeration gases; mixed industrial gases; inert gases such as carbon dioxide; and shielding gases. Additionally, the term "industrial gas equipment" may also include process equipment for manufacturing industrial gases (such as ammonia) in NACE class 20.15, for extracting and / or manufacturing methane, ethane, butane or propane (NACE classes 06.20 and 19.20) and for manufacturing gaseous fuels as defined by NACE class 35.21. The above has been described in terms of the European NACE system but is intended to cover equivalent classes under the North American classifications SIC and NAICS. Additionally, the above list is non-limiting and non-exhaustive.

[0310] In some instances, a hydrogen storage system is shown and in some cases a purification unit is shown. However, it should be understood that the present invention may be implemented without using a hydrogen storage system or a purification unit and the hydrogen storage system or purification unit is shown herein only for completeness.

[0311] In this specification, unless otherwise expressly stated, the word "or" is used in the sense of an operator that returns a true value when either or both of the stated conditions are satisfied, rather than in the sense of the exclusive "or" operator which requires only one of the conditions to be satisfied. The word "comprising" is used in the sense of "including" rather than "consisting of".

[0312] In the discussion of embodiments of the present invention, the pressures given are absolute pressures unless otherwise stated.

[0313] All of the above prior teachings are hereby incorporated herein by reference. The acknowledgement of any previously published document herein should not be taken as an admission or indication that its teachings were common general knowledge in Australia or elsewhere at the time.

[0314] Where applicable, the various embodiments provided by the present disclosure may be implemented using hardware, software, or a combination of hardware and software. Additionally, where applicable, the various hardware components and / or software components set forth herein may be combined into composite components including software, hardware, and / or both without departing from the spirit of the present disclosure. Where applicable, the various hardware components and / or software components set forth herein may be divided into sub-components including software, hardware, or both without departing from the scope of the present disclosure. Additionally, where applicable, it is contemplated that software components may be implemented as hardware components and vice versa.

[0315] Software in accordance with the present disclosure, such as program code and / or data, may be stored on one or more computer-readable media. It is also contemplated that the software identified herein may be implemented using one or more general or special purpose computers and / or computer systems that are networked and / or otherwise. Where applicable, the order of the various steps described herein may be altered, combined into composite steps, and / or divided into sub-steps to provide the features described herein.

[0316] Although various operations have been described herein in terms of "modules", "units", or "components", it should be noted that these terms are not limited to a single unit or function. Additionally, the functions attributed to some of the modules or components described herein may be combined and attributed to fewer modules or components. Further, although the invention has been described with reference to specific examples, these examples are merely illustrative and are not intended to limit the invention. It will be apparent to those of ordinary skill in the art that changes, additions, or deletions may be made to the disclosed embodiments without departing from the spirit and scope of the invention. For example, one or more portions of the above-described methods may be performed in a different order (or simultaneously) and still obtain the desired results.

Claims

1. A method for monitoring the operating characteristics of an industrial gas equipment complex including a plurality of industrial gas equipment, the method being executed by at least one hardware processor, the method comprising: Assigning a machine learning model to each of the industrial gas equipment forming the industrial gas equipment complex, the industrial gas equipment including a hydrogen production equipment for producing hydrogen, an air separation unit (ASU) for producing nitrogen, and an ammonia production equipment for producing ammonia; Training a corresponding machine learning model for each industrial gas equipment based on the received historical time-related operating characteristic data of the corresponding industrial gas equipment; Executing the trained machine learning model for each industrial gas equipment to predict the operating characteristics of each corresponding industrial gas equipment within a predetermined future time period, so as to maximize ammonia production and produce a sufficient amount of hydrogen to support ammonia production under the constraints of the total predicted power available within the given predetermined future time period and the amount of hydrogen in storage; Generating setpoints for the industrial gas equipment based on the operating characteristics predicted by executing the trained machine learning model for use in the predetermined future time period, such that the industrial gas equipment maximizes ammonia production when operating during the predetermined future time period, thereby producing as much ammonia as possible within the predetermined future time period based on the total predicted power available within the predetermined future time period; The industrial gas equipment uses the generated setpoints to produce ammonia during the predetermined future time period, thereby producing as much ammonia as possible within the predetermined future time period; Comparing the predicted operating characteristic data of each corresponding industrial gas equipment within the predetermined future time period with the measured operating characteristic data within the corresponding time period to identify deviations in the performance of the industrial gas equipment for the production and / or storage of nitrogen, ammonia, and hydrogen, so as to identify future problems in the industrial gas equipment complex; and Arranging at least one remedial measure based on the identified future problems to avoid unplanned downtime of the industrial gas equipment complex, the at least one remedial measure including arranging maintenance based on the capacity of the storage unit to maintain the continuity of the services provided by the industrial gas equipment complex during maintenance.

2. The method according to claim 1, wherein the comparing step is performed at the end of the predetermined future time period of the predicted operating characteristic data or at a timestamp therein.

3. The method according to claim 1, wherein the comparing step includes comparing the predicted operating characteristic data predicted within a predetermined time window with the actually measured operating characteristic data within the same time window.

4. The method according to claim 1, wherein the received historical time-related operating characteristic data of the corresponding industrial gas equipment includes data obtained from direct measurement of the processes or parameters of the corresponding industrial gas equipment.

5. The method according to claim 1, wherein the received historical time-related operating characteristic data regarding the respective industrial gas equipment includes data obtained from a physics-based model representing the operating characteristics of the respective industrial gas equipment.

6. The method according to claim 5, wherein measurement data related to the process or parameters of the respective industrial gas equipment is input into the respective physics-based model.

7. The method according to claim 1, wherein predicted operating characteristics regarding each industrial gas equipment are utilized to determine predicted future resources, future faults, and / or predicted future maintenance.

8. The method according to claim 1, wherein the hydrogen production equipment includes a plurality of electrolyzer modules.

9. The method according to claim 8, wherein a machine learning model is assigned to each of the electrolyzer modules.

10. The method according to claim 1, wherein predicted operating characteristics regarding each respective industrial gas equipment are utilized in an additional model to generate an operating performance metric for the industrial gas equipment complex.

11. The method according to claim 10, wherein the operating performance metric includes an efficiency value of the industrial gas equipment complex.

12. The method according to claim 11, wherein the determined efficiency value enables the determination of a prediction of the ammonia produced by the ammonia production equipment for a given level of energy input.

13. A system for monitoring the operating characteristics of an industrial gas equipment complex including a plurality of industrial gas equipment, the system including at least one hardware processor operable to perform: assigning a machine learning model to each of the industrial gas equipment forming the industrial gas equipment complex, the industrial gas equipment including a hydrogen production equipment for producing hydrogen, an air separation unit (ASU) for producing nitrogen, and an ammonia production equipment for producing ammonia; training a respective machine learning model for each industrial gas equipment based on received historical time-related operating characteristic data regarding the respective industrial gas equipment; executing the trained machine learning model for each industrial gas equipment to predict the operating characteristics regarding each respective industrial gas equipment within a predetermined future time period, to maximize ammonia production and produce a sufficient amount of hydrogen to support ammonia production subject to constraints on the total amount of predicted power available and the amount of hydrogen in storage within the predetermined future time period; generating a set point for the industrial gas equipment based on the operating characteristics predicted by executing the trained machine learning model for use in the predetermined future time period, such that the industrial gas equipment maximizes ammonia production when operating during the predetermined future time period, thereby producing as much ammonia as possible within the predetermined future time period based on the total amount of predicted power available during the predetermined future time period; providing the generated set point to a control system of the industrial gas equipment to control the industrial gas equipment during the predetermined future time period, thereby producing as much ammonia as possible within the predetermined future time period; Compare the predicted operating characteristic data for each respective industrial gas device over the said predetermined future time period with the measured operating characteristic data over the corresponding time period to identify deviations in the performance of the industrial gas devices for the production and / or storage of nitrogen, ammonia, and hydrogen, so as to identify future problems in the industrial gas device complex; and Arrange at least one remedial measure based on the identified future problems to avoid unplanned downtime of the industrial gas device complex, the at least one remedial measure including arranging maintenance based on the capacity of the storage unit to maintain the continuity of the services provided by the industrial gas device complex during maintenance.

14. The system according to claim 13, wherein the comparison step is performed at the end of the predetermined future time period of the predicted operating characteristic data or at a timestamp therein.

15. The system according to claim 13, wherein the comparison step includes comparing the predicted operating characteristic data predicted within a predetermined time window with the actually measured operating characteristic data within the same time window.

16. The system according to claim 13, wherein the received historical time-related operating characteristic data regarding the respective industrial gas devices includes data obtained from direct measurements of the processes or parameters of the respective industrial gas devices.

17. The system according to claim 13, wherein the received historical time-related operating characteristic data regarding the respective industrial gas devices includes data obtained from a physics-based model representing the operating characteristics of the respective industrial gas devices.

18. The system according to claim 13, wherein the predicted future resources, future failures, and / or predicted future maintenance are determined using the predicted operating characteristics of each industrial gas device.

19. The system according to claim 13, wherein the predicted operating characteristics of each respective industrial gas device are used in an additional model to generate an operating performance metric for the industrial gas device complex.

20. A computer-readable storage medium storing a program of instructions, the program of instructions being executable by a machine to perform a method for monitoring the operating characteristics of an industrial gas device complex including a plurality of industrial gas devices, the method being executed by at least one hardware processor, the method comprising: Assign a machine learning model to each of the industrial gas devices forming the industrial gas device complex, the industrial gas devices including a hydrogen production device for producing hydrogen, an air separation unit (ASU) for producing nitrogen, and an ammonia production device for producing ammonia; Train the respective machine learning models for each industrial gas device based on the received historical time-related operating characteristic data regarding the respective industrial gas devices; A machine learning model that performs training on each industrial gas device to predict the operating characteristics of each corresponding industrial gas device within a predetermined future time period, to maximize ammonia production and produce a sufficient amount of hydrogen to support ammonia production under the constraints of the total predicted power available within the said predetermined future time period and the amount of hydrogen in storage; Based on the operating characteristics predicted by performing the trained machine learning model, generate set points for the industrial gas device for the predetermined future time period, such that the industrial gas device maximizes ammonia production when operating during the predetermined future time period, thereby producing as much ammonia as possible within the predetermined future time period based on the total predicted power available within the predetermined future time period; Provide the generated set points to the control system of the industrial gas device to control the industrial gas device during the predetermined future time period, thereby producing as much ammonia as possible within the predetermined future time period; Compare the predicted operating characteristic data of each corresponding industrial gas device within the predetermined future time period with the measured operating characteristic data within the corresponding time period to identify deviations in the performance of the industrial gas device for the production and / or storage of nitrogen, ammonia, and hydrogen, in order to identify future problems in the industrial gas device complex; and Arrange at least one remedial measure based on the identified future problems to avoid unplanned shutdowns of the industrial gas device complex, the at least one remedial measure including arranging maintenance based on the capacity of the storage unit to maintain the continuity of the services provided by the industrial gas device complex during maintenance.

Images